A meta-analysis of the association between RGS4 gene polymorphisms and schizophrenia

Background: Schizophrenia is a complex brain disorder, the pathogenesis of which remains unclear. Regulator of G-protein signaling 4 ( RGS4 ) is regarded as a candidate gene for schizophrenia risk. The association between the RGS4 gene and the risk of schizophrenia is complicated and controversial, thus, an updated meta-analysis is needed. Methods: A search strategy using Medical Subject Headings was developed in English (PubMed, SZGene) and Chinese (CNKI, Wanfang and Weipu) databases. Inclusion and exclusion criteria were used to screen for eligible studies. Parameters, such as P -value of Hardy − Weinberg equilibrium ( P HWE ), odds ratios (ORs), 95% confidence intervals (CIs), P values of association ( P z ,), heterogeneity ( P h ), and publication bias ( P e ), were analyzed by the Stata software using a random effects model. Subgroup analyses were performed to detect heterogeneity. Results: There were 15 articles regarding rs10917670 (8,046 cases and 8,837 controls), 16 regarding rs951436 (8,990 cases and 10,568 controls), 15 regarding rs951439 (7,995 cases and 8,646 controls), 15 regarding rs2661319 (8,320 cases and 9,440 controls), and 4 regarding rs10759 (2,752 cases and 2,866 controls). The frequencies of rs10917670 and rs951439 were not significantly different between the case and control groups (p > 0.05). As shown by the East Asian and hospital-based subgroup analyses, the genotype TT of rs951436 might be related to the risk of schizophrenia. The genotypes CC+CT of rs2661319 and CC+CA of rs10759 were statistically different between the two groups, and the East Asian population contributed to these differences. Conclusion: The genotypes CC+CT of rs2661319 and CC+CA of rs10759 might be associated with the risk of schizophrenia.


Introduction
Schizophrenia is a complex brain disorder, the pathogenesis of which remains unclear [1].
It has been shown that schizophrenia is caused by both genetic and environmental factors [2], and genetic factors play an important role to the etiology of schizophrenia [3,4].
Regulator of G-protein signaling (RGS) proteins control the duration and timing of intracellular signaling of many G-protein coupled receptors (GPCRs). The major mechanism by which RGS proteins negatively regulate G proteins is via their GTPase accelerating activity [5]. RGS4 is known to play a fundamental role in neurotransmission and neuronal differentiation, in addition to axonogenesis during embryogenesis [6]. RGS4 regulation of G-protein activity, may inhibit the interaction between neurotransmitters and their receptors, leading to dysfunction of glutamatergic neurotransmission [7], which is classically related to the etiology of psychotic disorders [8]. Schwarz et al. [6] suggested that the RGS4 gene, localized to chromosome 1q23, might be an important part of a larger biological system contributing to schizophrenia risk. Mirnics et al. [9] showed that RGS4 expression was down regulated in schizophrenia [10,11]. However, the association between RGS4 and the risk of schizophrenia remains controversial [12][13][14][15].
Meta-analysis is a useful tool for the detection of disease − gene relationships [16]. In the Chinese Han population, one meta-analysis showed no association between the RGS4 gene and the risk of schizophrenia [15]; however, in another meta-analysis, the SNP, rs951436, was found to be associated with the risk of schizophrenia [17]. Therefore, the association between RGS4 and the risk of schizophrenia remains complicated and controversial [17][18][19]. Additional articles have since been published; thus, an updated meta-analysis is needed. Here, we conducted an updated meta-analysis to detect the association between RGS4 gene polymorphisms and the risk of schizophrenia.

. 1 . L i t e r a t u r e s e a r c h
The systematic review and meta-analysis were conducted in adherence to the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines [20]. A search was performed in English (PubMed, SZGene) and Chinese (CNKI, Wanfang and Weipu) databases with the following keywords: "the regulator of G-protein signaling 4" or "RGS4" and "schizophrenia". References to related articles were also reviewed for further data.

Identification and eligibility of relevant studies
The inclusion criteria were: 1) studies with a case-control design; 2) involvement of patients with schizophrenia; 3) available allele or genotype frequencies; and 4) published before February 20th, 2019. The authors were emailed if there was no genotype frequency mentioned in the article. The exclusion criteria were: 1) family-based studies; 2) no control group data; 3) no detailed genotype frequency data after emailing the authors; and 4) duplicate samples [21]. Information regarding the author, year, country, ethnicity, controls source, mean age of the control group, number of samples, diagnostic criteria, gender index the of cases and controls, and genotypes of the cases and controls were collected.

Statistical analysis
The meta-analysis was conducted using Stata version 10.0 (Stata Corp., College Station, TX). In the control group, the P-value of Hardy − Weinberg equilibrium (P HWE ) was calculated. Parameters, such as the odds ratios (ORs), 95% confidence intervals (CIs), and P-values of association (P z ,), were calculated to detect the association in five genetic models [22], using the random effects model [21,23]. The heterogeneity of the studies (P h ) was determined by Cochran's chi-square-based Q-statistic test. To assess the heterogeneity, subgroup analyses by ethnicity and control source were performed [24].
The studies were classified by control source into community-based (participants from the general population) and hospital-based (participants from a hospital) groups [25]. The Egger's test was conducted to detect the publication bias (P e ), which could be visualized using a funnel plot. To assess the impact of each study on the pooled results, sensitivity analysis was performed by removing single studies in turn. The power was calculated using the PS program [26]. The threshold for statistical significance was P < 0.05 in all tests.

Description of studies
A total of 106 English and 43 Chinese articles were found, with 20 articles being eligible for analysis following exclusion (Fig. 1). The data regarding the genotypes in [11,14,27] were unavailable. Table 1 described the detailed characteristics of the 20 eligible studies.

There was no association between rs951439 and the risk of schizophrenia
To evaluate the relationship between rs951439 and the risk of schizophrenia, 7,995 cases and 8,646 controls were included in the pooled and subgroup analyses ( Table 6). Detailed genotype frequencies were not available in [43]; thus, these data were only included in the allele contrast. No relationship between rs951439 and the risk of schizophrenia was detected in the pooled analysis (P z = 0.414, OR = 1.036, 95%CI = 0.952 − 1.128) using the dominant model (Fig. 4) or in the subgroup analyses by ethnicity and control source. No significant heterogeneity was observed in the pooled or subgroup analyses.

Rs2661319 might be a risk factor for schizophrenia
Pooled and subgroup analyses of 8,320 cases and 9,440 controls were performed (Table 7). Of the five genetic models, significant differences were detected when using allele contrast (C vs T, P z = 0.023), homozygous codominant (CC vs TT, P z = 0.034 ), dominant (CC + CT vs TT, P z = 0.016), and recessive (CC vs CT + TT, P z = 0.046 ).
According to the dominant model (Fig. 5  3.2.5. Genotype CC + CA of rs10759 might be a risk factor for schizophrenia A total of 2,752 cases and 2,866 controls were analyzed in pooled and subgroup analyses (Table 8). Significant differences were observed in two of the genetic models, allele contrast (C vs A, P z = 0.046) and dominant (CC + CA vs AA, P z = 0.016). Using the random effects model, the dominant model was selected (Fig. 6). The genotype CC + CA of   Figure 1. Article selection process in the present meta-analysis. Figure 2. Forest plot of the association between rs10917670 and schizophrenia using a recessive model (GG vs GA + AA). Figure 3. Forest plot of the association between rs951436 and schizophrenia using a recessive model (TT vs TG + GG). Figure 4. Forest plot of the association between rs951439 and schizophrenia using a dominant model (GG + GA vs AA). Figure 5. Forest plot of the association between rs2661319 and schizophrenia using a dominant model (CC + CT vs TT). Figure 6. Forest plot of the association between rs10759 and schizophrenia using a dominant model (CC + CA vs AA).

Sensitivity analysis
Sensitivity analysis was conducted by omitting each study in turn. The results showed that pooled ORs did not change significantly; thus, the results were considered stable and reasonable.

Publication bias
Publication bias could be visualized using funnel plots. No evidence of publication bias was found in the pooled analysis ( Figures S1-S5).

Discussion
No association between rs10917670 and rs951439 and the risk of schizophrenia was detected in the present study, which was consistent with previous meta-analyses [17][18][19].
In the East Asian and hospital-based subgroup analyses, an association between the genotype TT of rs951436 and the risk of schizophrenia was found; however, this relationship was not detected in the pooled analysis. Therefore, the geographical environment, culture, lifestyle, and genetic background might affect polymorphisms [30,34,42]. It was studied that rs951436 was associated with magnetic resonance imaging measurements of functional activation and connectivity related to working memory, an intermediate phenotype of schizophrenia [44]. Moreover, Prasad et al. reported that rs951436 was related the volume of dorsolateral prefrontal cortex (DLPFC) [32]. But the mechanism remained unclear.
Rs2661319 and rs10759 were found to be associated with the risk of schizophrenia in the present study, which was inconsistent with previous meta-analyses. It was detected by subgroup analyses that the East Asian population contributed to this association. It was previously reported that rs2661319 was related to RGS4-1 mRNA level, which was decreased in the postmortem DLPFC of schizophrenic patients [11]. Moreover, rs2661319 was demonstrated to be associated with a more severe baseline total PANSS score and the treatment effect of perphenazine [45]. The rs10759 polymorphism was suggested to increase the risk of schizophrenia by altering the binding of miRNA-124 to its target [46].
MiRNA-124 might bind to the 3′UTR of mRNAs containing target sites, resulting in miRNAmediated gene silencing, translational inhibition, and induction of mRNA de-adenylation or decay [47]. The level of RGS4 might be decreased, leading to dysfunction of neurotransmission.
More relevant data were included in our meta-analysis than those in previous metaanalyses, for instance, an increased number of more SNPs (5), and databases ((PubMed and SZGene, CNKI, Wanfang and Weipu). However, the results described herein should be interpreted with caution. First, in the present study, the East Asian population contributed to the association between the RGS4 gene and the risk of schizophrenia; however, the sample size was relatively small, and the power was low. Further articles are needed to form a representative and comprehensive conclusion. Second, family-based and functional studies were not included in the present meta-analysis. In addition, it was reported that there was an association between DLPFC volume and RGS4 genotype interacting with COMT rs4818 [ 48]; thus, this association warrants further gene-gene interaction [49,50] and functional studies.

Conclusion
No association between rs10917670 and rs951439 and the risk of schizophrenia was found. In the East Asian and hospital-based subgroup analyses, an association between rs951436 and the risk of schizophrenia was demonstrated. The genotypes CC + CT of rs2661319 and CC + CA of rs10759 might be risk factors for schizophrenia, and the East Asian population contributed to this association. Further updated gene-gene interaction and functional studies are needed.

Ethics approval and consent to participate
Not applicable.

Consent for publication
Not applicable.

Availability of data and materials
This was an evidence synthesis study; all data were available from the primary research studies or could be provided by the corresponding author.
BW designed the study and wrote the protocol. FX and LL managed the literature search, which was checked by XW, YL, and XX. FX performed analyses. The manuscript was written by FX, and corrected by JY. All authors have read and approved the manuscript.

Figure 1
Article selection process in this meta-analysis.

Supplementary Files
This is a list of supplementary files associated with the primary manuscript. Click to download. supplementary.docx