Association between FAS gene −670 A/G and −1377 G/A polymorphisms and the risk of autoimmune diseases: a meta-analysis

Abstract Objectives: FAS plays a critical role in the extrinsic apoptosis pathway in autoimmune diseases. Previous studies investigating the association between FAS gene −670 A/G and −1377 G/A polymorphisms and the risk of autoimmune diseases reported controversial results. We performed the meta-analysis to evaluate the possible association. Methods: Relevant studies were identified by searching the PubMed, Embase, CNKI, and Wanfang databases up to December 2018. Odds ratios (ORs) and corresponding 95% confidence intervals (CIs) were calculated to determine the association. Results: A total of 43 articles including 67 studies (52 studies for FAS −670 A/G and 15 studies for −1377 G/A) were included in the meta-analysis. Our meta-analysis showed that the FAS −670 A/G polymorphism was associated with the risk of autoimmune diseases (GG vs. GA: OR = 1.079, 95% CI = 1.004–1.160, P=0.038), especially in Caucasians (GG vs. GA: OR = 1.12, 95% CI = 1.03–1.23, P=0.012), Asians (G vs. A: OR = 0.89, 95% CI = 0.83–0.96, P=0.002), systemic lupus erythematosus (SLE) (G vs. A: OR = 0.85, 95% CI = 0.77–0.94, P=0.001), multiple sclerosis (MS) (GG+GA vs. AA: OR = 0.83, 95% CI = 0.70–0.99, P=0.043), systemic sclerosis (SSc) (GG vs. GA: OR = 1.20, 95% CI = 1.07–1.36, P=0.003) and Hashimoto’s thyroiditis (HT) (G vs. A: OR = 1.45, 95% CI = 1.10–1.90, P=0.008); the FAS −1377 G/A polymorphism was associated with the risk of autoimmune diseases (A vs. G: OR = 1.11, 95% CI = 1.03–1.20, P=0.008), especially in Asians (A vs. G: OR = 1.15, 95% CI = 1.05–1.25, P=0.002) and high quality studies (A vs. G: OR = 1.14, 95% CI = 1.05–1.24, P=0.002). Conclusion: This meta-analysis demonstrated that the FAS –670A/G and –1377 G/A polymorphisms were associated with the risk of autoimmune diseases.


Introduction
Autoimmune diseases are chronic disorders characterized by the loss of immune tolerance to self-antigens, leading to immune-mediated tissue destruction. They affect 4-5% of adults, the majority of whom are women [1]. Co-occurrence of distinct autoimmune diseases within a single family and genome-wide association studies (GWASs) support the hypothesis that these diseases share common genetic risk factors [2][3][4][5][6]. The etiology of autoimmune diseases is attributed to complex interactions of genetics, epigenetics, and environmental factors that remain to be elucidated [7][8][9][10][11][12].
FAS (also known as APO-1, CD95, or TNFSF6) is a cell surface receptor that belongs to the tumor necrosis factor (TNF) receptor superfamily [13]. FAS is widely expressed in normal human tissues. To maintain self-tolerance, the binding of FAS-ligand (FASL) to FAS on the cell surface initiates the extrinsic apoptosis pathway [14]; thus, autoreactive lymphocytes are normally eliminated. However, abnormal apoptosis may lead to a failure to eliminate autoreactive lymphocytes, which can induce the appearance and development of autoimmune diseases [15]. The FAS gene is located on chromosome 10q24.1 in humans and is highly polymorphic [16]. In some individuals, there is an A to G substitution at position 670 and a G to A substitution at position 1377 in the FAS promoter region [17]. The FAS −670 A/G and −1377 G/A polymorphisms may destroy signal transducer and activator of transcription protein 1 (STAT1) and stimulatory protein 1 (SP1) transcription factor binding sites, resulting in reduced promoter activity and FAS expression [18]. Abnormal apoptosis mediated by the FASL interaction with the FAS receptor is involved in the pathogenesis of several autoimmune diseases and cancers [19].

Statistical analysis
The chi-square test was applied to examine whether the observed genotype frequencies in controls conformed to HWE, and P<0.05 was considered to deviate from HWE. The ORs with their 95% CIs were used to assess the strength of associations between the FAS −670 A/G and −1377 G/A polymorphisms and autoimmune diseases. The statistical significance of the pooled ORs was determined by the Z test. The between-studies heterogeneity was assessed by Q test and quantified by I 2 test [69]. When P≥0.1 or I 2 < 50%, there was no heterogeneity, and pooled OR estimates were combined using the fixed-effects model (Mantel-Haenszel method); otherwise, the random-effects model (Mantel-Haenszel method) was used to combine summary data [70]. To detect the main sources of heterogeneity, subgroup analyses were performed by ethnicity, disease type and quality score. Sensitivity analysis was carried out by excluding studies deviating from HWE to assess the stability of the meta-analysis. Egger's test was used to assess publication bias [71]. If there was publication bias, we recalculated the adjusted ORs using the trim-and-fill method [72] to evaluate the possible impact of publication bias. The trim-and-fill method was used to impute hypothetical missing studies. For significant results observed in the current meta-analysis, the false-positive report probability (FPRP) test was utilized to examine positive associations. An FPRP threshold of 0.5 and a prior probability of 0.1 were set to detect an OR of 0.67/1.50 (protective/risk effects) for an association with the tested genotypes. FPRP values less than 0.5 were considered as noteworthy associations [73]. All statistical analyses were conducted using Stata 15 software (Stata Corporation, College Station, TX, U.S.A.). Results with P<0.05 were considered significant.

Trial sequential analysis
Traditional meta-analysis may yield type I errors due to dispersed data or repetitive significance testing when new studies are added to it [74,75]. Trial sequential analysis (TSA) was used to minimize the risk of type I errors by calculating required information size (RIS) (meta-analysis sample size) and adjusted threshold for statistical significance [76]. TSA was performed by using TSA software 0.9.5.10 Beta (http://www.ctu.dk/tsa/) in the allelic model with the overall included studies by setting an overall type I error of 5%, power of 80%, relative risk reduction (RRR) of 20%, and control event proportion [77]. If the cumulative Z-curve crosses the trial sequential monitoring boundary or the RIS line, a reliable and conclusive evidence has been reached and further studies are not needed. Otherwise, more studies are needed to reach a firm conclusion.

Characteristics of the included studies
A flowchart of the selection of eligible articles is presented in Figure 1. The initial search identified 2552 articles through the search strategy, and a total of 43 articles [15,17,, consisting of 67 studies comprising 13340 patients and 14547 controls, were finally included in the meta-analysis according to the inclusion and exclusion criteria. Fifty-two studies examined the FAS −670 A/G polymorphism, and 15 studies examined the FAS −1377 G/A polymorphism. The characteristics of the articles included in the meta-analysis are summarized in Table 1.

Meta-analysis results of the FAS −670 A/G and −1377 G/A polymorphisms and autoimmune diseases
A summary of the meta-analysis of the association between the FAS −670 A/G and −1377 G/A polymorphisms and autoimmune diseases is shown in

Stratification analyses by ethnicity, disease type, and quality score
Based on ethnicity, disease type, and quality score, we performed stratification analyses.
The results of the meta-analysis of the association between the FAS −670 A/G and −1377 G/A polymorphisms and autoimmune diseases risk stratified by ethnicity, disease type, and quality score are shown in Table 2.
On the basis of ethnicity, the stratified meta-analysis showed an association between FAS −670 A/G polymorphism and the risk of autoimmune diseases in Caucasians (GG vs. GA: OR = 1.12, 95% CI 1.03-1. 23 On the basis of quality score, the stratified meta-analysis suggested that the FAS −670 A/G polymorphism might not be associated with autoimmune diseases in high-or low-quality studies. However, the association between FAS −1377 G/A polymorphism and the risk of autoimmune diseases was observed in high-quality studies (A vs. G: OR = 1.14, 95% CI 1.05-1.24, P=0.002) but not in low-quality studies.
Stratification analysis showed that ethnicity, disease type, and quality score might be the factors of heterogeneity across all studies of association between FAS −670 A/G polymorphism and autoimmune diseases risk, and quality score may be the factor of heterogeneity across all studies of association between FAS −1377 G/A polymorphism and autoimmune diseases risk. A r a s t e he ta l . 2 0 1 0 I r a n C a u c a s i a n A S O -P C R 2 4 9 / 2 1 2 7 4 9 3 8 2 5 8 9 8 5 6 0 . 2 7 3 7 X ue ta l . 2 0 0 4 C h i n a A s i a n P C R -R F L P 1 0 3 / 1 1 0 1 5 5 9 2 9 2 3 6 1 2 6 0 . 2  C o a k l e y e t a l . 1 9 9 9 U . S . A . C a u c a s i a n P C R

Publication bias
The Egger's test was performed to assess the publication bias under all genetic models of the meta-analysis and the results are shown in  (Table 3).

Sensitivity analysis
The genotype frequencies in the controls of five articles [22,31,51,52,55] deviated significantly from the HWE, which could cause potential bias. To check the robustness of our results, sensitivity analysis was performed by excluding

FPRP analysis results
The FPRP values were calculated for the main significant associations and the results are shown in Table 4.

TSA results
In the TSA of association of FAS −670 A/G polymorphism and autoimmune diseases risk, the cumulative Z-curve neither crossed conventional boundary nor trial sequential monitoring boundary, however, the sample size reached RIS (3365) in allelic model (Figure 2A). In the TSA of association of FAS −1377 G/A polymorphism and autoimmune diseases risk, the sample size also reached RIS (4387) and the cumulative Z-curve crossed the conventional boundary, although the cumulative Z-curve did not cross trial sequential monitoring boundary in allelic model ( Figure 2B). The TSA results indicated that the cumulative evidence was reliable and sufficient, and no additional studies were required.

Discussion
Our results showed that ethnicity, disease type, and quality score may be the factors of heterogeneity across all studies of association between FAS −670 A/G polymorphism and autoimmune diseases, and quality score may be the factor of heterogeneity across all studies of association between FAS −1377 G/A polymorphism and autoimmune diseases. In the ethnicity stratification analysis, the results of our meta-analysis revealed diverse associations between the FAS −670 A/G and −1377 G/A polymorphisms and various autoimmune diseases in different ethnic groups.   In the disease-type stratification analysis, the FAS −670 G allele was associated with an increased risk of SSc and HT and with a decreased risk of SLE, MS, and AIH (in Asians) but was not associated with other autoimmune diseases. These findings may reflect differences in the risks of various autoimmune diseases due to differences in environmental and genetic backgrounds. The present results indicate that the FAS −670 G allele is associated with a decreased risk of SLE, MS, and AIH (in Asians), which conflicts with a previous finding that the FAS −670 G allele in the FAS promoter was associated with an increased risk of autoimmune diseases [22,78]. One possible mechanism by which this allele may reduce the risk of SLE, MS, and AIH (in Asians) is by a reduction in soluble FAS (sFAS). The FAS protein exists in two isoforms, one a transmembrane protein and the other a soluble protein. sFAS expression is highly regulated at the mRNA transcript level [79,80]. Transcription of both FAS and sFAS is driven by the same gene promoter [22], with alternative splicing of the FAS mRNA resulting in a variant that lacks exon 6, which encodes the transmembrane domain of FAS [81]. Plasma sFAS, an antiapoptotic molecule, has been found to block apoptosis in autoreactive lymphocytes by competing with FAS for FASL or soluble FASL binding in SLE, MS, and AIH (in Asians) [79,[82][83][84][85]. Similarly, this may explain why the FAS −670 G allele was associated with an increased risk of autoimmune diseases in Caucasians and with a decreased risk in Asians. For FAS −1377 G/A polymorphism, subgroup analysis was not performed owing to the limited study number. The FAS −1377 G/A polymorphism occurs at the consensus sequence of transcription factor SP1 binding site in the silencer region [48]. The FAS −1377 A allele may destroy SP1 transcription factor binding sites, resulting in reduced promoter activity and FAS expression [18]. Abnormal apoptosis mediated by the FASL interaction with the FAS receptor is involved in the pathogenesis of several autoimmune diseases [19].
We performed a meta-analysis of data from patients diagnosed with autoimmune diseases (SLE, MS, RA, AIH, LN, SSc, AA, pSS, HT, GBS, PBC, vitiligo, GD, T1D, IAA, JIA, and SPA) and healthy controls. This meta-analysis differs from the seven previous meta-analyses [43,[61][62][63][64][65][66] because the present study included 33 more studies (consisting of new studies with same and different disease types) [15,17,22,[27][28][29][30][32][33][34][35]37,41,[43][44][45]50,[52][53][54][55]59,60] and yielded several novel and distinct findings. One previous meta-analysis [62] including SLE, RA, SSc, pSS, JIA, and SPA demonstrated that the FAS −670 A/G polymorphism might be associated with the risk of rheumatic disease, especially in Asians, SLE and RA, and the FAS −1377 G/A polymorphism was associated with SLE risk. Compared with this meta-analysis, our meta-analysis focused on overall autoimmune diseases risk and showed that FAS −670 A/G polymorphism was associated with autoimmune diseases risk in Caucasians, MS, SSc and HT; and the FAS −1377 G/A polymorphism was associated with autoimmune diseases risk in Asians and high quality studies, which were different from the previous meta-analyses. One meta-analysis [43] showed that the FAS −670 A/G polymorphism may be associated with SLE risk in the Chinese population. Two meta-analyses [64,66] suggested that the FAS −670 A/G and −1377 G/A polymorphisms was associated with the risk of SLE, stratification by ethnicity indicated an association between the FAS −670 A/G and SLE in Asian populations. Two meta-analyses [61,63] showed that the FAS −670 A/G polymorphism was not associated with the risk of RA. One meta-analysis [65] suggested that the FAS −670 A/G polymorphism was not associated with the risk of AIH. These six meta-analyses focused on the association between FAS polymorphism and a single disease (SLE, RA, or AIH). Compared with these meta-analyses, our meta-analysis covered overall autoimmune diseases, and subgroup analyses were performed by ethnicity, disease type, and quality score, thereby yielding several novel and distinct findings. Furthermore, some previous meta-analyses [63,65] including several studies [25,30,31,40] made some errors when extracting the data. Thus, we here added 33 new studies [15,17,22,[27][28][29][30][32][33][34][35]37,41,[43][44][45]50,[52][53][54][55]59,60] on SLE, MS, pSS, AA, PBC, HT, GBS, LN, vitiligo, T1D, IAA, and GD and corrected the previous errors, providing more reliable results. In addition, FPRP test was performed to support that the evidence of our results was robust and sufficiently conclusive, and the result of TSA showed that there was sufficient evidence and much larger sample size to support these conclusions, thereby increasing the statistical power. We strongly believe that our findings can help resolve many of the controversies of the association of FAS polymorphism and autoimmune diseases. Sensitivity analysis are generally performed to assess the robustness of meta-analyses by excluding and including HWE-deviating studies from genetic association studies, which is a recommended approach [86]. Probable explanations for deviation from HWE include nonrandom mating, population stratification, selection bias, genotyping error, inbreeding, genetic drift, chance, differential survival of marker carriers, or combinations of these reasons [87]. However, key empirical evidence does not support a strong association between estimates of genetic effect and deviations from HWE [88]. Nonetheless, the findings of our meta-analysis should be interpreted with caution in the case of material alterations in results after excluding the HWE-deviating studies.
The present study has several limitations that should be considered when interpreting the conclusions. First, only case-control studies were considered for inclusion. Selection bias and unmeasured confounding can occur at both the design and analysis stages of observational studies. Second, this analysis only included articles published in English and Chinese; this may reduce the credibility of the results because of language bias [89]. Third, our study only analyzed a single locus, single nucleotide polymorphism (SNP) −670 A/G and −1377 G/A in the FAS gene and did not investigate associations between genetic haplotypes containing the FAS −670 A/G and −1377 G/A polymorphisms and the risk of autoimmune diseases because of inadequate haplotype data. It is unknown whether other genetic mutations contribute to changes in the expression or function of the FAS gene. For uncovering the genetic causes of disease, haplotypes provide more information and have a greater influence than genotypes and single SNPs. Fourth, most studies included in our analysis were performed in the Caucasian and Asian populations; therefore, our results may apply only to these ethnic groups. Additional studies of other ethnicities are needed. Fifth, autoimmune diseases are multifactorial diseases caused by interactions between genetic and environmental factors, meaning that the FAS −670 A/G and −1377 G/A polymorphisms may only partially influence the pathogenesis of autoimmune diseases; this may lead to bias in the present results. Finally, the findings of our meta-analysis should be interpreted with caution in the case of heterogeneity observed under some genetic models.
Translating information of genetic associations into clinical diagnostics would help with improved understanding of the autoimmune diseases' etiology. Establishing evidence-based medical evidence of genetic susceptibility to autoimmune diseases risk might facilitate the preventive and therapeutic strategies, which has a beneficial clinical utility for not only clinicians and researchers but also patients.
In summary, our meta-analysis suggested that the FAS −670 A/G polymorphism might be associated with the risk of autoimmune diseases, especially in Caucasians and Asians, SLE, MS, SSc, and HT. Moreover, the FAS −670 A/G polymorphism might be associated with the risk of autoimmune diseases in Asian patients with SLE or AIH and Caucasian patients with SLE, MS, or SSc. The FAS −1377 G/A/ polymorphism might be associated with the risk of autoimmune diseases, specifically for Asians and high quality studies. Stratification analysis showed that ethnicity, disease type and quality score might be the factors of heterogeneity across all studies of association between FAS −670 A/G polymorphism and autoimmune diseases risk, and quality score might be the factor of heterogeneity across all studies of association between FAS −1377 G/A polymorphism and autoimmune diseases risk.