Association of nitric oxide synthase 1 adaptor protein gene polymorphisms with schizophrenia in a Chinese Han population.

Introduction: Previous studies have analyzed the association between nitric oxide synthase 1 adaptor protein (NOS1AP) polymorphisms and schizophrenia; however, the results were inconsistent and there was a lack of evidence in a larger sample of Chinese Han population. Subjects and
Methods: We decided to determine the association between four NOS1AP single-nucleotide polymorphisms (i.e., rs1858232A/G, rs4531275C/T, rs4657178C/T, and rs6704393C/T) and schizophrenia in northern Chinese Han population (350 patients and 522 controls) using restriction fragment length polymorphism.
Results: Between schizophrenia group and healthy group, the genotype and allele frequencies for rs1858232A/G differed significantly (χ 2 = 6.256, 4.145; P = 0.044, 0.045), but neither genotype nor allele frequencies of rs4531275C/T differed significantly. The genotype frequencies for rs4657178C/T and rs6704393C/T differed significantly (χ 2 = 19.782, 12.683; P < 0.01, P = 0.002) between schizophrenia group and healthy group. In the gender-specific analysis, we found statistically significant difference in genotype frequencies between patients and controls in both subgroups for rs4657178C/T (χ 2 = 9.356, 9.585; P = 0.009, 0.008). There was also a significant difference in the genotype frequency between patients and controls in male subgroup for rs6704393C/T (χ 2 = 8.800, P = 0.012). In the haplotype analysis, only the TCT haplotype frequency of rs6704393C/T, rs4531275C/T, and rs4657178C/T differed significantly between patients and controls in total population (χ 2 = 5.215, P = 0.022). In conclusion: Individuals with G allele of rs1858232A/G and C allele of rs4657178C/T which may be risk factors for schizophrenia should be given more attention, and also to individuals with the TCT haplotype, who are more likely to have schizophrenia. These results provide novel evidence for an association between NOS1AP polymorphisms and schizophrenia. Copyright:

INTRODUCTION

Nitric oxide synthase 1 adaptor protein (NOS1AP) is a gene located on chromosome 1q22 that encodes the cytosolic NOS1AP. Also known as CAPON, the N-terminal phosphotyrosine binding domain of NOS1AP binds Dexras1 (a small monomeric G protein) and the C-terminal PDZ-binding domain interacts with neuronal nitric oxide synthase (nNOS) ([Homo sapiens (human)] Gene ID: 9722, updated on 3-Apr-2016). nNOS is a signaling molecule encoded by NOS1 that synthesizes nitric oxide (NO) from L-arginine in the central nervous system. NO is a highly reactive signaling molecule and NO signaling may be enhanced by NOS1AP that couples nNOS to specific target proteins by protein–protein interactions. For example, the brain-enriched Dexras1 has been identified to be a physiologic effect factor of NO and nNOS, also NOS1AP is necessary for the interaction of NO and nNOS with their targets.[1] Dexras1 activity can also be activated by N-methyl D-aspartate (NMDA) in a NO-dependent manner. NO is involved in storage, uptake, and release of mediators and neurotransmitters, and changes in these processes may result in neurodevelopmental disorders such as schizophrenia. Behavioral abnormalities were observed in animals with nitrinergic dysregulation, and schizophrenia and cognition is associated with reduced NO signaling in the prefrontal cortex.[2] NMDA receptor hypoactivity, one of the mechanisms underlying schizophrenia, might be exacerbated by NOS1AP.[3] Using linkage disequilibrium (LD) analysis, Brzustowicz et al. reported the involvement of NOS1AP in the NMDA receptor signaling pathways and the etiology of schizophrenia.[4] NOS1AP plays a role in the inhibition of glutamate neurotransmission by disrupting the binding of nNOS to postsynaptic density protein PSD-93 and PSD-95.[5] Xu et al. showed significant overexpression of NOS1AP in both schizophrenia and bipolar disorder patients.[6] The increase in NOS1AP expression is significantly associated with single-nucleotide polymorphisms (SNPs) previously found to be in LD with schizophrenia. Based on these studies, NOS1AP is recognized as one of the risk factors of schizophrenia.

There have been some researches on the association of NOS1AP with schizophrenia in different populations.[789] Although Fang et al. performed family-based association study on five NOS1AP SNPs in 319 Chinese Han trios, and found none of the SNPs associated with schizophrenia,[10] a lack of evidence was found in the previous report in a Chinese Han population, we decided to determine the association between NOS1AP polymorphisms in northern Chinese Han population. Four NOS1AP SNPs (i.e., rs1858232A/G, rs4531275C/T, rs4657178C/T, and rs6704393C/T) were selected based on previous studies.[11] All four SNPs lie in the intronic regions of NOS1AP, and the functional impact of these mutations on NOS1AP has not been studied.

SUBJECTS AND METHODS

Subjects

The case group was made up of participants according to the diagnostic criteria for schizophrenia,[1213] and were unrelated Chinese Han inpatients recruited randomly from six hospitals in Liaoning province, including 218 males (mean ± standard deviation [SD] age: 46.12 ± 9.56 years; range: 18–64 years) and 132 females (mean ± SD age: 45.55 ± 9.67 years; range: 18–62 years). Exclusion criteria included intellectual disability, cerebral organic disorders, serious physical illnesses, and comorbid psychiatric disorders such as substance abuse and depression.[1213] The control group was made up of healthy individuals without the presence of psychiatric disorders or serious physical illnesses in the same region of China through structured interviews and questionnaires, consisting of 249 males (mean ± SD age: 30.97 ± 7.95 years; range: 19–58 years) and 273 females (mean ± SD age: 29.34 ± 7.79 years; range: 19–56 years). The ages of participants between the two groups have no significant difference (P > 0.05). In this study, all the participants and their family members gave written informed consent based on programs of the Ethics Committee of China Medical University.

Genotyping

According to the methods as detailed in our previous studies,[121314] DNA purification, polymerase chain reaction (PCR) amplification, and restriction fragment length polymorphism (RFLP) analysis were manipulated. Genomic DNA amplification was carried out in a 25-ml PCR reaction admixture (TaqDNA polymerization 12.5 ml, DNA template 1.0 ml, ddH2O 10.5 ml, sense 0.5 ml, and antisense 0.5 ml) and the primers used are following (Takara, Dalian, China): The primers for rs1858232 were 5’-GAA TCG TCA AAA TCA GCA TTT CCT A-3’ (sense) and 5’- TGG ACC CTG GCA TGT GTT AT-3’ (antisense). The primers for rs4531275 were 5’-ACA AAC CTG CTG TAA TCC TGG TA-3’ (sense) and 5’- AAA GCT CCT TCC CTT CAC AGC-3’ (antisense). The primers for rs4657178 were 5’-TCA GCG TGT TAT TTG GCA GC-3’ (sense) and 5’- ACT CCA CTG TGA CCA TCC CT-3’ (antisense). The primers for rs6704393 were 5’-CAG AGG CCC AGA AAG AGG TA-3’ (sense) and 5’- CCT GGT CAG TCA GTA GCC AT-3’ (antisense). PCR reaction conditions included degeneration (94°C, 5 min), degeneration (94°C, 30 s), annealing (56.0°C–62°C, 1 min), elongation (72°C, 1 min) (30 circles), and elongation (72°C, 10 min).

The genotypes for rs1858232A/G, rs4531275C/T, rs4657178C/T, and rs6704393C/T were confirmed, respectively, by RFLP analysis using restriction endonucleases: AluI, HinfI, BstNI, and MluCI (New England Biolabs, Beijing, China). The reaction of the PCR products with the corresponding restriction endonuclease was done sufficiently in a 20-ml reaction admixture at 37°C for rs1858232A/G, rs4531275C/T, and rs6704393C/T, and 60°C for rs4657178C/T, and subsequently was analyzed by electrophoresis. The fragments for rs1858232A/G, rs4531275C/T, rs6704393C/T, and rs4657178C/T were 305 bp/92 bp [Figure 1], 285 bp/109 bp [Figure 2], 179 bp/83 bp [Figure 3], and 176 bp/120 bp [Figure 4], respectively.

Figure 1

Electrophoresis figure for rs1858232A/G

Figure 2

Electrophoresis figure for rs4531275C/T

Figure 3

Electrophoresis figure for rs4657178C/T

Figure 4

Electrophoresis figure for rs6704393C/T

Statistical analysis

The Chi-square test (or Fisher’s exact test) and odds ratio (ORs) calculation by version 22.0 of SPSS® software and Haploview version 4.2 were used in our study[121315] (SPSS Inc., IL, USA) (http://www.broad.mit.edu/mpg/haploview). All the data were showed as frequency, percentage, and mean ± SD. It was considered statistically significant when two-tailed P < 0.05.

RESULTS

Of the four NOS1AP SNPs, the genotype distributions of rs4657178C/T and rs6704393C/T showed significant deviations from the Hardy–Weinberg equilibrium (HWE) in patients (χ2 = 21.139, 12.162; P = 0.01, 0.002, respectively), whereas rs1858232A/G and rs4531275C/T showed no significant deviations (χ2 = 4.844, 0.824; P = 0.092, 0.669, respectively). Patients group genotype frequencies of rs4657178 C/T have no significant diversity from those of 319 Chinese Han individuals observed in an earlier study (χ2 = 0.536, P = 0.765).[10] The healthy controls were all in HWE and represented the northern and general Han Chinese population (χ2 = 1.549, 0.877, 0.766, 0.124; P = 0.471, 0.638, 0.692, 0.936, respectively).

The genotype and allele frequencies of the four NOS1AP SNPs in case group and control group are shown in Table 1. The genotype and allele frequencies of rs1858232A/G differed significantly between case group and healthy group (χ2 = 6.256a, 4.145a; P = 0.044, 0.045; odds ratio [OR] = 1.220, respectively). Dominant models of rs1858232A/G (AG + GG, AA) showed significant difference (χ2 = 6.199a; P = 0.014; OR = 1.518), but not in the recessive models. The genotype frequency and recessive models (TT, TC + CC) for rs4657178C/T differed significantly between patients and controls (χ2 = 19.782a, 16.356a; P < 0.01, <0.01; OR = 0.425, respectively), but not allele frequency or recessive models. The genotype frequency and dominant models (TC + TT, CC) for rs6704393C/T differed significantly between patients and controls (χ2 = 12.683a, 6.322a; P = 0.002, 0.013; OR = 1.419, respectively), but not allele frequency or dominant models. Finally, neither genotype nor allele frequencies of rs4531275C/T showed significant differences between patients and controls. Moreover, there was no significant difference in both recessive and dominant models.

Table 1

Genotype and allele distributions of the four analyzed SNPs in schizophrenic patients and healthy controls

SNP	Whole patients, n (%)	Whole controls, n (%	χ 2	P	OR (95% CI)	Male patients, n (%)	Male controls, n (%)	χ 2	P	OR (95% CI)	Female patients, n (%)	Female controls, n (%)	χ 2	P	OR (95% CI)
rs1858232
Genetypes
AA	67 (19.1)	138 (26.4)				45 (20.6)	70 (28.1)				22 (16.7)	68 (24.9)
GG	79 (22.6)	104 (19.9)				48 (22.0)	45 (18.1)				31 (23.5)	59 (21.6)
AG	204 (58.3)	280 (53.6)	6.256a	0.044		125 (57.3)	134 (53.8)	3.803a	0.151		79 (59.8)	146 (53.5)	3.510a	0.180
Alleles
A	338 (48.3)	556 (53.3)	4.145a	0.045	1.220(1.007-1.478)	215 (49.3)	274 (55.0)	3.037a	0.088	1.257(0.972-1.627)	123 (46.6)	282 (51.6)	1.821a	0.202	1.225(0.912-1.644)
G	362 (51.7)	488 (46.7)				221 (50.7)	224 (45.0)				141 (53.4)	264 (48.4)
Dominant models
AG + GG	283 (80.9)	384 (73.6)	6.199a	0.014	1.518(1.092-2.111)	173 (79.4)	179 (71.9)	3.495a	0.068	1.503(0.979-2.308)	110 (83.3)	205 (75.1)	3.497a	0.074	1.659(0.973-2.828)
AA	67 (19.1)	138 (26.4)				45 (20.6)	70 (28.1)				22 (16.7)	68 (24.9)
Recessive models
GG	271 (77.4)	418 (80.1)	0.886a	0.352	1.172(0.842-1.630)	170 (78.0)	204 (81.9)	1.135a	0.298	1.280(0.812-2.017)	101 (76.5)	216 (78.4)	0.181a	0.703	1.113(0.679-1.826)
AG+AA	79 (22.6)	104 (19.9)				48 (22.0)	45 (18.1)				31 (23.5)	59 (21.6)
rs4531275
Genetypes
TT	80 (22.9)	138 (26.4)				46 (21.1)	72 (28.9)				34 (25.8)	66 (24.2)
CC	83 (23.7)	109 (20.9)				51 (23.4)	57 (22.9)				32 (24.2)	52 (19.0)
TC	187 (53.4)	275 (52.7)	1.860a	0.390		121 (55.5)	120 (48.2)	4.026a	0.130		66 (50.0)	155 (56.8)	1.997a	0.371
Alleles
C	353 (50.4)	493 (47.2)	1.725a	0.204	1.137(0.939-1.377)	223 (51.1)	234 (47.0)	2.842a	0.178	1.201(0.930-1.552)	130 (49.2)	259 (47.4)	0.233a	0.653	0.930(0.693-1.248)
T	347 (49.6)	551 (52.8)				213 (48.9)	264 (53.0)				134 (50.8)	287 (52.6)
Dominant models
TC + CC	270 (77.1)	384 (73.6)	1.432a	0.264	1.213(0.884-1.664)	172 (78.9)	177 (71.1)	3.759a	0.055	1.521(0.994-2.327)	98 (74.2)	207 (75.8)	0.120a	0.806	1.088(0.674-1.756)
TT	80 (22.9)	138 (26.4)				46 (21.1)	72 (28.9)				34 (25.8)	66 (24.2)
Recessive models
CC	267 (76.3)	413 (79.1)	0.979a	0.359	1.178(0.852-1.629)	167 (76.6)	192 (77.1)	1.169a	0.778	1.080(0.708-1.649)	100 (75.8)	221 (81.0)	1.461a	0.241	0.735(0.446-1.212)
TC + TT	83 (23.7)	109 (20.9				51 (23.4)	57 (22.9)				32 (24.2)	52 (19.0)
rs4657178
Genetypes
CC	92 (26.3)	149 (28.5)				60 (27.5)	73 (29.3)				32 (24.2)	76 (27.8)
TT	32 (9.1)	100 (19.2)				19 (8.7)	44 (17.7)				13 (9.8)	56 (20.5)
CC	226 (64.6)	273 (52.3)	19.782a	<0.01		139 (63.8)	132 (53.0)	9.356a	0.009		87 (65.9)	141 (51.6)	9.585a	0.008
Alleles
C	410 (58.6)	571 (54.7)			0.854(0.704-1036)	259 (59.4)	278 (55.8)			0.864(0.666-1.120)	151 (57.2)	293 (53.7)			1.154(0.858-1.552)
T	290 (41.4)	473 (45.3)	2.561a	0.115		177 (40.6)	220 (44.2)	1.219a	0.289		113 (42.8)	253 (46.3)	0.897a	0.366
Recessive models
TT	318 (90.9)	42 (80.8)			0.425	199 (91.3)	205 (82.3)			0.445(0.251-0.788)	119 (90.2)	217 (79.5)			2.362(1.241-4.496)
TC + CC	32 (9.1)	100 (19.2)	16.356a	<0.01		19 (8.7)	44 (17.7)	7.987a	0.006		13 (9.8)	56 (20.5)	7.159a	0.011
Dominant models
TC + TT	258 (73.7)	373 (71.5)			1.120(0.826-1.519)	158 (72.5)	176 (70.7)			1.092(0.730-1.635)	100 (75.8)	197 (72.2)			0.829(0.514-1.338)
CC	92 (26.3)	149 (28.5)	0.534a	0.487		60 (27.5)	73 (29.3)	0.184a	0.682		32 (24.2)	76 (27.8)	0.589a	0.474
rs6704393 Genetypes
TT	16 (4.6)	40 (7.7)				9 (4.1)	23 (9.2)				7 (5.3)	17 (6.2)
CC	146 (41.7)	263 (50.4)				89 (40.8)	118 (47.4)				57 (43.2)	145 (53.1)
TC	188 (53.7)	219 (42.0)	12.683a	0.002		120 (55.0)	108 (43.4)	8.800a	0.012		68 (51.5)	111 (40.7)	4.260a	0.120
Alleles
T	220 (31.4)	299 (28.6)			1.142(0.927-1.407)	138 (31.7)	154 (30.9)			1.034(0.784-1.365)	82 (31.1)	145 (26.6)			1.246(0.902-1.720)
C	480 (68.6)	745 (71.4)	1.559a	0.219		298 (68.3)	344 (69.1)	0.057a	0.832		182 (68.9)	401 (73.4)	1.790a	0.183
Recessive models
TT	334 (95.4)	482 (92.3)			0.577(0.318-1.048)	209 (95.9)	226 (90.8)			0.423(0.191-0.935)	125 (94.7)	256 (93.8)			0.843(0.341-2.086)
TC+CC	16 (4.57)	40 (7.7)	3.332a	0.090		9 (4.1)	23 (9.2)	4.752a	0.042		7 (5.3)	17 (6.2)	0.136a	0.825
Dominant models
TC+TT	204 (58.3)	259 (49.6)			1.419	129 (59.2)	131 (52.6)			1.306(0.904-1.885)	75 (56.8)	128 (46.9)			1.491(0.981-2.265)
CC	146 (41.7)	263 (50.4)	6.322a	0.013		89 (40.8)	118 (47.4)	2.029a	0.162		57 (43.2)	145 (53.1)	3.511a	0.071

aIndicates statistical significance. SNP – Singlenucleotide polymorphisms; OR – Odds ratio; CI – Confidence interval

In gender-specific analyses, for rs1858232A/G and rs4531275C/T, no significant difference was observed between male patients and controls or between female patients and controls. There was statistically significant difference in genotype and recessive models (TT, TC + CC) of rs4657178C/T between male patients and controls (χ2 = 9.356a, 7.987a; P = 0.009, 0.006, OR = 0.445, respectively), as well as between female patients and controls (χ2 = 9.585a, 7.159a; P = 0.008, 0.011; OR = 2.362, respectively). For rs6704393C/T, there was a significant difference in the genotype frequency between male patients and controls (χ2 = 8.800a, P = 0.012), and the recessive models (TT, TC + CC) between male patients and controls showed statistically significant difference (χ2 = 4.752a, P = 0.042), and no significant difference in dominant models (TC + TT, CC). We found no significant difference for rs6704393C/T between female patients and controls.

In the case group, there was no significant difference in the age of the first onset among the different genotypes (AA/GG/AG, TT/CC/TC, CC/TT/TC, and TT/CC/TC) of four SNPs (i.e., rs1858232A/G, rs4531275C/T, rs4657178C/T, and rs6704393C/T) (χ2 = 2.449, 0.529, 0.090,0.113; P = 0.294, 0.768, 0.956, 0.945, respectively).

LDs were observed between following pairs of SNPs in total population by Haploview (The value is usually from 0 to 1; 0 indicates the complete linkage equilibrium, and 1 indicates complete LD): (1) rs6704393C/T and rs4531275C/T (D¢=0.783, r2 = 0.245), (2) rs6704393C/T and rs4657178C/T (D¢=0.448, r2 = 0.109), and (3) rs4531275C/T and rs4657178C/T (D¢=0.487, r2 = 0.174) [Figure 5]. In the male population, LDs were seen between (1) rs6704393C/T and rs4531275C/T (D¢=0.769, r2 = 0.257), (2) rs6704393C/T and rs4657178C/T (D¢=0.366, r2 = 0.083), and (3) rs4531275C/T and rs4657178C/T (D¢=0.454, r2 = 0.146) [Figure 6]. In the female population, LDs were found between (1) rs6704393C/T and rs4531275C/T (D¢=0.8043, r2 = 0.233), (2) between rs6704393C/T and rs4657178C/T (D¢=0.555, r2 = 0.145), and (3) rs4531275C/T and rs4657178C/T (D¢=0.522, r2 = 0.208) [Figure 7].

Figure 5

Linkage disequilibrium plots for the four analyzed SNPs of the whole schizophrenic patients and healthy controls

Figure 6

Linkage disequilibrium plots for the four analyzed SNPs of the male schizophrenic patients and healthy controls

Figure 7

Linkage disequilibrium plots for the four analyzed SNPs of the female schizophrenic patients and healthy controls

In the total population and the gender-specific population, the haplotype frequencies of the three SNPs (rs6704393C/T, rs4657178 C/T, and rs4531275 C/T) between the patients and controls were computed [Table 2]. When frequencies were <1% in all groups, the haplotype blocks were rejected. In the total population, the TCT haplotype had a higher frequency in patients compared to controls (χ2 = 5.215, P = 0.022). Other haplotype frequencies between patients and controls did not differ significantly in the total population as well as in the male- and female-specific populations.

Table 2

Distribution of haplotypes for the three analyzed single-nucleotide polymorphisms between schizophrenic patients and healthy controls

Haplotype	Whole frequency	Whole case-control ratio	χ 2	P	Male frequency	Male case-control ratio	χ 2	P	Female frequency	Female case-control ratio	χ 2	P
CCC	0.365	0.386:0.351	2.228	0.136	0.359	0.385:0.337	2.345	0.126	0.372	0.384:0.366	0.252	0.616
TTT	0.189	0.187:0.190	0.022	0.883	0.182	0.177:0.186	0.131	0.717	0.196	0.200:0.194	0.044	0.833
CTT	0.142	0.126:0.152	2.234	0.135	0.131	0.124:0.137	0.359	0.549	0.155	0.131:0.166	1.617	0.204
CTC	0.106	0.099:0.111	0.658	0.417	0.102	0.094:0.108	0.493	0.483	0.110	0.106:0.113	0.087	0.768
CCT	0.089	0.074:0.099	3.309	0.069	0.096	0.080:0.108	2.155	0.142	0.083	0.068:0.090	1.116	0.291
TTC	0.078	0.083:0.075	0.411	0.522	0.096	0.093:0.099	0.073	0.786	0.059	0.071:0.054	0.933	0.334
TCT	0.018	0.027:0.012	5.215	0.022	0.017	0.025:0.010	3.078	0.079	0.019	0.029:0.014	2.065	0.151
TCC	0.013	0.017:0.010	1.883	0.170	0.018	0.021:0.015	0.614	0.433	<0.01

DISCUSSION

Schizophrenia is a complex and strongly heritable mental disorder, including behavioral and cognitive syndromes. Disruption of brain development caused by genetic and environmental factors is linked to the development of schizophrenia.[1617] Brain imaging techniques have shown morphological brain changes, cortical atrophy, and ventricular enlargement in individuals with schizophrenia.[18] Excessive dopaminergic neurotransmission also contributes to the genesis of schizophrenia. This dysfunction likely causes disturbances of synaptic function that leads to abnormal neuronal connectivity.[1619] Studies have suggested that the mutation and overexpression of NOS1AP and the interaction between nNOS and NOS1AP can affect brain morphology and dendritic branching and alter filopodial outgrowth and dendritic spine development that are key neuropathological features characteristic of schizophrenia.[2021] Carrel et al. showed a critical role of NOS1AP in cortical patterning which may contribute to the potential pathophysiology seen in schizophrenia.[22] These evidences suggest that NOS1AP is associated with various symptoms of schizophrenia.

Several earlier studies have reported association between NOS1AP polymorphisms and schizophrenia. However, there is conflicting evidence for the association of NOS1AP with schizophrenia. Miranda et al. found that NOS1AP is associated with schizophrenia in a Colombian population.[7] The work by Kremeyer et al. supported this association in a South American population, especially in schizophrenia patients with negative symptoms.[8] Cheah et al. genotyped nine NOS1AP SNPs including four SNPs of our study in 235 schizophrenia subjects, and found eight SNPs (rs1415259, rs386231, rs4531275, rs1415263, rs4656355, rs4657178, rs6704393, and rs6683968) to be associated with schizophrenia depression-related phenotypes and rs1858232 was found to be related with the broad diagnosis of schizophrenia.[11] However, no evidence for allelic association between SNPs in NOS1AP and schizophrenia was found in a British population.[9] The diversity between studies may be attributed to genetic heterogeneity in different ethnic populations or methodological errors.

However, there were none that examined the relationship in northern Han Chinese population. Therefore, we conducted this investigation to specifically analyze the association between NOS1AP gene polymorphisms and schizophrenia using a case–control design in northern Han Chinese population. We compared the genotype, allele, and haplotype frequencies of four SNPs between schizophrenia case group and healthy group. We identified several potential genetic risk factors for schizophrenia in northern Han Chinese population. The C allele and CC and CT genotypes of rs4657178C/T had significantly higher frequencies in the case group and gender-specific subgroups. The G allele and GG and AG genotypes of rs1858232A/G also had significantly higher frequencies in the case group. Individuals with G allele of rs1858232A/G and C allele of rs4657178C/T which may be risk factors for schizophrenia should be given more attention, who are more likely to have schizophrenia. The TCT haplotype frequencies of rs6704393C/T, rs4531275C/T, and rs4657178C/T were also significantly higher in patients; thus, the TCT haplotype may also be risk factor for schizophrenia. Interestingly, the CC genotype frequency of rs6704393C/T was significantly higher in the overall control group and the male control subgroup, which may be a protective factor for schizophrenia.

Gender difference may be a factor to influence the association between NOS1AP polymorphisms and schizophrenia. Several studies have shown gender difference in the epidemiology and clinical expression of schizophrenia. Women tend to display more affective symptoms and experience more benign course and better outcomes, whereas men show a higher propensity to negative symptoms, onset at earlier age, lower social functioning and comorbid substance abuse. It is unknown whether these differences extend to clinical high-risk subjects for psychosis.[23] In our study, two NOS1AP polymorphisms (rs4657178C/T and rs6704393C/T) showed gender-specific association with schizophrenia. The frequency of CC genotype of rs6704393C/T was significantly higher in the male subgroup in control group than that in the previous Cheah et al.’s study.[11] If we can further determine the existence of the superposition of these genders, it may help us to pay more attention to the sex population with risk alleles earlier. At the same time, the effect of sex can be divided into other detail studies, which is the direction to improve the research in the future.

Age is another factor influencing the outset of schizophrenia. To be exact, the age of onset is more significant for genetic study than that of age at the time of admission. Previous studies showed that childhood adversity may be associated with the expression of schizophrenia, but whether genetic factors play any role remained unknown.[24] Regretfully, in this study, we found no significant difference in age at first onset among genotypes, including the GG, AG genotypes of rs1858232A/G, and the CC, TC genotypes of rs4657178C/T, which have been shown to be risk factors for schizophrenia, indicating that the locus genotypes maybe have no correlation and influence on the onset age of patients. However, larger sample with more accurate collection of age data and covering larger age range should be done in the future, and it may be more powerful to explain the onset age influence in detail.

In this work, we reported several potential risk factors for schizophrenia relating to NOS1AP polymorphism in northern Han Chinese, including the G allele of rs1858232A/G, C allele of rs4657178C/T, and the TCT haplotype. However, there were some limits in our study. For example, the genotype distributions of rs4657178C/T and rs6704393C/T significantly deviated from HWE. Population stratification and the size of sample were considered as possible factors for this deviation. Further association studies in larger sample of Han Chinese and other racial populations should be done to identify and know novel NOS1AP SNPs in the future, to get complete understanding of NOS1AP allelic and haplotypes variants and their effect on the pathomechanism and symptom development of schizophrenia. Comprehensive evaluation of the interplay between gender or other factors and NOS1AP on the occurrence and development and the physiological effects of schizophrenia is needed.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.