July 2019
Volume 60, Issue 8
Open Access
Genetics  |   July 2019
Association of Polymorphisms at the SIX1-SIX6 Locus With Primary Open-Angle Glaucoma
Author Affiliations & Notes
  • Shi Yao Lu
    Department of Ophthalmology and Visual Sciences, The Chinese University of Hong Kong, Hong Kong, China
  • Zong Ze He
    Department of Neurosurgery, Sichuan Academy of Medical Sciences and Sichuan Provincial People's Hospital, School of Medicine, University of Electronic Science and Technology of China, Chengdu, Sichuan, China
  • Jia Xin Xu
    School of Clinic Medicine, Southwest Medical University, Luzhou, Sichuan, China
  • Chen Yang
    The Key Laboratory for Human Disease Gene Study of Sichuan Province and Department of Laboratory Medicine, Sichuan Academy of Medical Sciences and Sichuan Provincial People's Hospital, School of Medicine, University of Electronic Science and Technology of China, Chengdu, Sichuan, China
  • Li Jia Chen
    Department of Ophthalmology and Visual Sciences, The Chinese University of Hong Kong, Hong Kong, China
  • Bo Gong
    The Key Laboratory for Human Disease Gene Study of Sichuan Province and Department of Laboratory Medicine, Sichuan Academy of Medical Sciences and Sichuan Provincial People's Hospital, School of Medicine, University of Electronic Science and Technology of China, Chengdu, Sichuan, China
    Institute of Chengdu Biology, Sichuan Translational Medicine Hospital, Chinese Academy of Sciences, Chengdu, Sichuan, China
  • Correspondence: Li Jia Chen, Department of Ophthalmology and Visual Sciences, The Chinese University of Hong Kong, 147K Argyle Street, Hong Kong, China; [email protected]
  • Bo Gong, The Key Laboratory for Human Disease Gene Study of Sichuan Province, Sichuan Academy of Medical Sciences and Sichuan Provincial People's Hospital, 32 Road West 2, the First Ring, Chengdu, Sichuan 610072, China; [email protected]
  • Footnotes
     SYL and ZZH contributed equally to the work presented here and should therefore be regarded as equivalent authors.
Investigative Ophthalmology & Visual Science July 2019, Vol.60, 2914-2924. doi:https://doi.org/10.1167/iovs.18-26489
  • Views
  • PDF
  • Share
  • Tools
    • Alerts
      ×
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      Shi Yao Lu, Zong Ze He, Jia Xin Xu, Chen Yang, Li Jia Chen, Bo Gong; Association of Polymorphisms at the SIX1-SIX6 Locus With Primary Open-Angle Glaucoma. Invest. Ophthalmol. Vis. Sci. 2019;60(8):2914-2924. https://doi.org/10.1167/iovs.18-26489.

      Download citation file:


      © ARVO (1962-2015); The Authors (2016-present)

      ×
  • Supplements
Abstract

Purpose: To evaluate the association of single-nucleotide polymorphisms (SNPs) in the SIX1-SIX6 locus with primary open-angle glaucoma (POAG) through a systematic review and meta-analysis from 22 studies.

Methods: To our knowledge, all case-control association studies on SNPs in the SIX1-SIX6 locus and POAG reported up to August 30, 2018, in PubMed, Embase, and Web of Science were retrieved. Unadjusted and adjusted odds ratios (ORs) and 95% confidence intervals (95% CIs) for each SNP were calculated using a fixed- or random-effect model according to interstudy heterogeneity.

Results: This meta-analysis involved 12 SNPs in SIX1-SIX6 reported in 22 studies. The association of rs10483727 with POAG has been presented in 16 studies involving 14,402 patients and 27,425 controls, whereas rs33912345 has been investigated in 12 studies involving 10,563 patients and 16,740 controls. Meta-analyses revealed significant associations of these two SNPs with POAG in the pooled populations under all genetic models. Stratified analyses by population detected significant association of both SNPs in the East Asian and Caucasian subgroups, but not in South Asian or African subgroups. Among the other SNPs that were reported by up to four cohorts of East Asian and African ancestries, only rs12436579 showed a significant association in the meta-analysis (OR = 0.79, P = 1.08 × 10−4).

Conclusions: This meta-analysis confirmed the association of rs10483727 and rs33912345 in SIX1-SIX6 with POAG. The associations of both SNPs were specifically detected in East Asian and Caucasian cohorts, rather than in South Asian and African cohorts, suggesting an ethnic difference. SNP rs12436579 is a candidate variant for the disease that awaits validation in other populations.

Glaucoma, a group of optic neuropathies, is a leading cause of irreversible blindness worldwide.1,2 It is characterized by typical optic nerve head damage associated with optic disc cupping and retinal ganglion cell (RGC) death and visual field defect.3,4 The global glaucoma prevalence in the population aged ≥40 years old is approximately 3.54%.5 By 2020, the estimated number of people diagnosed with primary glaucoma will increase to 76 to 79.6 million worldwide.5,6 
Primary open-angle glaucoma (POAG), a major form of glaucoma, is characterized by an open anterior iridocorneal chamber angle.7 POAG is a complex disease with multiple risk factors. Genetic factors play an important role in the pathogenesis of POAG.8 So far, more than 25 candidate loci have been linked to POAG, including MYOC,9 OPTN,10 WDR36,11 CAV1-CAV2,12 TMCO11,13 CDKN2B-AS1,13 SIX1-SIX6,14 ABCA1,15,16 PMM2,15 AFAP1,16 GMDS,16 FNDC3B,17 TGFBR3,17,18 TXNRD2,19 ATXN2,19 FOXC1,19 ANKRD55-MAP3K1,18 LMX1B,18 LHPP,18 HMGA2,18 MEIS2,18 LOXL1,18 and chromosomal regions 8q22 and 11p11.2.14,20 
Among these loci, the SIX1-SIX6 locus is one of the early findings in the genome-wide association studies (GWASs) of POAG. Single-nucleotide polymorphisms (SNPs) in this locus have been associated with POAG in multiple GWASs and subsequent studies in specific ethnic groups,14,17,2138 especially two variants, rs10483727 and rs33912345. However, the results were inconsistent across these studies, and the role of the SIX1-SIX6 locus in POAG remains inconclusive. Associations of other variants in this locus with POAG have also been reported, but the results were inconsistent. Moreover, variants in SIX1-SIX6 have been correlated with the retinal nerve fiber layer (RNFL) thickness, an endophenotype of glaucoma, in both POAG patients29 and population-based cohorts.35,39 Furthermore, SIX6 plays an important role in eye formation. It is expressed in the retina, choroid, ciliary body, and sclera.40 Deficiency of SIX6 homologs caused abnormal eye development (e.g., anophthalmia and microphthalmia) in Drosophila and zebrafish,41,42 and mutations (e.g., c.532_536del and c.493A>G) of SIX6 had been observed in congenital anophthalmia and microphthalmia patients.43,44 In contrast to the mutations that cause eye malformation, the SIX6 variants with POAG association should have less damaging effect. It has been proposed that SIX6 risk variants may disrupt the development of the neural retina, leading to a reduced number of RGCs and thereby increasing the risk of glaucoma.29 Thus, it is worthwhile to summarize the reported SIX1-SIX6 risk variants in POAG for a better understanding of their roles. Meta-analysis, as a commonly used method, can increase the overall sample size and the statistical power in a certain effect size for a more precise estimation.45 Thus, we conducted the present systematic review and meta-analysis with a view to elaborate the associations of SNPs in the SIX1-SIX6 locus with POAG. 
Methods
Searching Strategy
A systematic literature search in the databases PubMed, Embase, and Web of Science was conducted to identify all published studies on the association of polymorphisms in the SIX1-SIX6 locus with POAG from the starting dates of respective databases to August 30, 2018, using Boolean logic and controlled vocabularies. The logical expression for searching was (“SIX1” OR “SIX6” OR “SIX1-SIX6” OR “SIX1/SIX6” OR “SIX homeobox 1” OR “BOS3” OR “TIP39” OR “DFNA23” OR “SIX homeobox 6” OR “SIX9” OR “ODRMD” OR “OPTX2” OR “MCOPCT2”) AND (“polymorphism(s)” OR “SNP(s)” OR variant(s) OR mutation(s) OR locus OR loci OR gene(s) OR genotype(s)) AND (“glaucoma” OR “primary open-angle glaucoma” OR “POAG”). Reference lists of the retrieved articles and reviews were manually checked for additional articles that were potentially omitted by the electronic searches. Studies applying genome-wide screening methods (i.e., GWAS) can detect the association of our targeted polymorphisms, but some of them might have provided the data in supplementary materials rather than in main text. Therefore, we also added the POAG GWASs (n = 7) that were missed by the literature search from the databases. Furthermore, we added data of our recent study that investigated the association of SIX6 with POAG.38 
Inclusion and Exclusion Criteria
Eligible articles were considered if they (1) evaluated the associations between one or more SNPs in the SIX1-SIX6 locus and POAG; (2) used a case-control design to compare the allelic or genotypic frequencies between POAG cases and normal controls in defined populations; (3) reported an odds ratio (OR) with 95% confidence intervals (CIs), or provided available data (e.g., allele frequency and sample size) from which the OR and 95% CI can be estimated; and (4) calculated the association using genotyping data instead of data from imputation analysis. Only original research articles were included in the meta-analysis. Abstracts from conferences, full texts without raw data available, republished data, duplicate studies, and reviews were excluded. The SNPs to be included in a meta-analysis should be reported by more than two studies. There was no limitation in language. 
Data Extraction
Two reviewers (BG and JXX) performed the literature search. Two observers (SYL and ZZH) independently extracted the data from all eligible publications into a predesigned data collection form. Disagreements were resolved by discussion until a consensus was achieved. Otherwise, a third investigator (LJC) was involved to resolve the discrepancies. The following items were collected from each study: first author's surname, year of publication, statistical data, ethnicity of subjects, Hardy-Weinberg equilibrium in controls (if available), genotyping method, sample sizes of cases and controls, respectively. For calculating the unadjusted association, we extracted the number and/or frequency of the alleles from all the included studies. In order to calculate the adjusted association results, we also extracted in the adjusted ORs from 13 studies,13,14,19,2224,27,2932,38,46 which conducted logistic regression analysis under an additive model adjusted mainly for sex and age, as well as ocular parameters28 and top principal components14,19,22 in some studies. 
Statistical Analysis
A pooled OR with corresponding 95% CI were used to measure the association of the SIX1-SIX6 SNPs with POAG. The χ2 test was used to compare the allele frequencies between cases and control in the unadjusted association (allelic, a versus A). For the studies that did not provide ORs with 95% CIs for the allelic associations or obtained the association results by other tests, such as logistic regression, we used the χ2 test to reassess the associations based on the data from the original articles. Analyses of other genetic models, that is, additive (aa versus AA), recessive (aa versus Aa+aa), and dominant (aa+Aa versus AA), were also performed. Missing genotype counts were estimated according to the Hardy-Weinberg equation. The summary outcomes of unadjusted ORs were estimated by using inverse-variance–weighted meta-analysis. The 13 studies13,14,19,2224,27,2932,38,46 that contained adjusted ORs and 95% CIs were included in the meta-analysis of adjusted associations. The pooled effect sizes in the adjusted associations were also meta-analyzed by using inverse-variance weighting. Heterogeneity (between-study inconsistency) was measured using the I2 statistic. An I2 value of <40% indicated an absence of significant heterogeneity among studies. In this case, the fixed-effect model (Mantel-Haenszel method) was used to calculate the pooled OR. In contrast, if the I2 value was ≥40%, which indicates high heterogeneity between studies, the random-effect model (DerSimonian-Laird method) was applied. Sensitivity analysis was conducted by excluding the studies one at a time to explore whether the results were influenced by a specific study. The Egger liner regression test was used to assess the potential publication bias, where P value of <0.05 was considered statistically significant. All statistical analyses were performed using software (Review Manager [RevMan], version 5.3; the Nordic Cochrane Centre, the Cochrane Collaboration, Copenhagen, 2012, and ProMeta 3.0, available in the public domain from IDoStatistics; https://idostatistics.com/prometa3/). A two-sided P value of <0.05 was considered nominally statistically significant. P values in multiple testing were corrected by using Bonferroni method. A P value of ≤5.2 × 10−4 (= 0.05/(4*2*12)), where 4 was the number of genetic models tested (allelic, addictive, recessive, dominant), 2 was the number of data types (unadjusted and adjusted), and 12 was the number of SNPs analyzed, was considered statistically significant. 
Results
Literature Search and Characteristics of Studies
The flow of study selection is shown in Figure 1. The initial search strategy yielded 169 papers about the SIX1-SIX6 locus in POAG from the databases and reference lists. Eighty-six records were evaluated after removing duplicates, and 35 among them were excluded because they addressed unrelated topics or were conference papers. The full texts of the remaining 51 articles were retrieved and assessed. Six articles were excluded after full-text review: five publications were reviews and one was about cell biology. In the next step of eligibility evaluation, 23 articles were excluded due to inadequate data provided. Finally, 22 studies were included for the meta-analysis (Table 1). The data extracted for the meta-analyses were shown in Table 2 (rs10483727), Table 3 (rs33912345), and Supplementary Table S1 (other SIX1-SIX6 SNPs). These SNPs span from position 60,322,458 to 60,622,618 (GRCh38) in chromosome 14 (Supplementary Fig. S1). Supplementary Table S2 summarized the SNPs that were reported in single POAG association studies. 
Figure 1
 
Flow diagram of the study inclusion process on the association between the SIX6 locus and POAG.
Figure 1
 
Flow diagram of the study inclusion process on the association between the SIX6 locus and POAG.
Table 1
 
Characteristics of Studies Included for the Meta-Analysis
Table 1
 
Characteristics of Studies Included for the Meta-Analysis
Table 2
 
Allelic Frequencies and Effect Sizes of rs10483727 Among POAG Cases and Controls in the 16 Included Studies
Table 2
 
Allelic Frequencies and Effect Sizes of rs10483727 Among POAG Cases and Controls in the 16 Included Studies
Table 3
 
Allelic Frequencies and Effect Sizes of rs33912345 Among POAG Cases and Controls in the 12 Included Studies
Table 3
 
Allelic Frequencies and Effect Sizes of rs33912345 Among POAG Cases and Controls in the 12 Included Studies
Associations Between SIX1-SIX6 SNPs and POAG
Sixteen studies provided results on the POAG association of rs10483727 in 20 cohorts (Table 2), involving a total of 41,827 subjects (14,402 patients and 27,425 controls).14,22,2428,3034,36,38,46,47 There was moderate interstudy heterogeneity (I2 = 43.7%) among the 20 cohorts when analyzing the unadjusted ORs under the allelic model (Fig. 2 and Supplementary Table S3). Meta-analysis showed a significant association of rs10483727 with POAG in the overall pooled population (OR = 1.23, P = 1.46 × 10−12). We then stratified the cohorts into five subgroups (nine in East Asian, two in South Asian, one in West Asian, four in African, and four in Caucasian subgroups) according to the ethnicity information provided in the articles. Among them, only one study involved West Asians (Saudi Arabian; the original data are shown in Fig. 2);25 therefore, meta-analysis was not eligible. The association in this cohort is significant (P = 0.013).25 There was high heterogeneity (I2 = 51.6%) between the two cohorts in the South Asian subgroup, while there were low heterogeneities (I2 ≤ 5%) in other subgroups. Meta-analysis showed significant allelic associations of rs10483727 with POAG in East Asian (OR = 1.23, P = 4.02 × 10−16) and Caucasian (OR = 1.32, P = 2.90 × 10−16), but not in South Asian (OR = 0.96, P = 0.68) or African (OR = 1.05, P = 0.55; Fig. 2) cohorts. In the analysis of other genetic models using unadjusted ORs (Supplementary Table S3), the heterogeneities among all the cohorts were moderate in the additive (I2 = 40.1%) and recessive (I2 = 31.7%) models and were relatively low in the dominant model (I2 = 11%). Two African cohorts had no homozygous minor allele in the estimated genotype count, so they were removed from the analysis of additive and dominant models. Meta-analysis showed significant associations of rs10483727 with POAG in the overall population (P ≤ 1.09 × 10−8), East Asians (P ≤ 3.75 × 10−5), and Caucasians (P ≤ 2.69 × 10−10), but not in South Asians (P > 0.6) or Africans (P > 0.5) in different unadjusted models (Supplementary Table S3). The trend of the ORs under all genetic models was the same (>1) in all the populations except South Asian, in which the pooled ORs were less than 1 (Supplementary Table S3). This might imply a difference in the allele effect in the South Asian population, although this cannot be proved by the current data as the pooled P value was not statistically significant. 
Figure 2
 
Forest plot of the association between rs10483727 and POAG expressed as unadjusted OR. Squares indicate risk estimate in each study; horizontal lines represent 95% CIs and ORs; diamonds indicate summary estimate OR with corresponding 95% CI in each subgroup or pooled overall population. Effects of the SNP were weighted by inverse variance under fixed or random-effects model.
Figure 2
 
Forest plot of the association between rs10483727 and POAG expressed as unadjusted OR. Squares indicate risk estimate in each study; horizontal lines represent 95% CIs and ORs; diamonds indicate summary estimate OR with corresponding 95% CI in each subgroup or pooled overall population. Effects of the SNP were weighted by inverse variance under fixed or random-effects model.
A total of 27,303 subjects (10,563 POAG patients and 16,740 controls) in 17 cohorts from 12 studies were meta-analyzed for the association between rs33912345 and POAG (Table 3).17,19,23,24,26,27,29,32,33,35,37,38 The heterogeneity in the overall population (I2 = 37.2%) was slightly lower than that for rs10483727, but higher in South Asian and Caucasian cohorts (I2 = 70.4% and I2 = 28.1%, respectively; Fig. 3 and Supplementary Table S4). In stratification analysis, rs33912345 was reported in three Caucasian cohorts, two African cohorts, and one Latino cohort, but there was no cohort from West Asia. Under the allelic model using unadjusted ORs, this SNP was significantly associated with POAG in the overall pooled subjects (OR = 1.32, P = 4.72 × 10−34), as well as in East Asian (OR = 1.33, P = 2.27 × 10−16) and Caucasian (OR = 1.36, P = 2.30 × 10−20), but not in South Asian or African (Fig. 3) cohorts. The study in Latino cohorts also showed a significant association (OR = 1.57, P = 0.016).17 In other genetic models, the association patterns were similar (Supplementary Table S4). 
Figure 3
 
Forest plot of the association between rs33912345 and POAG expressed as unadjusted OR. Squares indicate risk estimate in each study; horizontal lines represent 95% CIs and ORs; diamonds indicate summary estimate OR with corresponding 95% CI in each subgroup or pooled overall population. Effects of the SNP were weighted by inverse variance under fixed or random-effects model.
Figure 3
 
Forest plot of the association between rs33912345 and POAG expressed as unadjusted OR. Squares indicate risk estimate in each study; horizontal lines represent 95% CIs and ORs; diamonds indicate summary estimate OR with corresponding 95% CI in each subgroup or pooled overall population. Effects of the SNP were weighted by inverse variance under fixed or random-effects model.
The pooled measures of adjusted ORs in the association between both SNPs (rs10483727 and rs33912345) and POAG are shown in Table 4. Three population subgroups, South Asian, West Asian, and Latino, were removed from the analysis because none of the studies reported the adjusted association results. The I2 for overall pooled cohorts was 16.5% for rs10483727 and 27.4% for rs33912345. In the stratification analysis, high heterogeneity was found in East Asian cohorts of rs33912345 (I2 = 49.7%), while others were not heterogeneous. Similar to the unadjusted analysis, both SNPs were strongly associated with POAG in the overall pooled population (P = 2.38 × 10−24 for rs10483727 and 1.03 × 10−16 for rs33912345), East Asian (P = 1.28 × 10−12 for rs10483727 and 6.19×10−6 for rs33912345) and Caucasian (P = 4.97 × 10−15 for rs10483727 and 1.55 × 10−8 for rs33912345), but not in African (P = 0.77 for rs10483727 and 0.26 for rs33912345) subgroups. The effect directions of both SNPs in all the subgroups were positive (i.e., OR > 1; Table 4). 
Table 4
 
The Meta-Analysis Results of Adjusted Association in Overall and Different Population Subgroups*
Table 4
 
The Meta-Analysis Results of Adjusted Association in Overall and Different Population Subgroups*
POAG associations of 10 other SNPs in the SIX1-SIX6 locus were meta-analyzed (Table 5). Only three studies contained the data of other SNPs that were reported more than once.32,38,46 These studies involved five cohorts: Southern Chinese, Japanese, South African, African American, and Ghanaian.32,38,46 Two SNPs (rs11849906 and rs1266416) had relatively high heterogeneities (I2 from 31.8% to 64.4%) in both unadjusted and adjusted analyses. None of the SNPs reached the significant level (P < 0.05) in the pooled association analysis except rs12436579 (ORunadjusted = 0.79, Punadjusted = 1.08 × 10−4; and ORadjusted = 0.79, Padjusted = 1.65 × 10−4). This SNP was examined in three East Asian cohorts38 and the South African cohort.46 Although only the Chinese cohorts demonstrated a significant association between this SNP and POAG, the African and Japanese cohorts showed the same direction of effect (Supplementary Table S1). The heterogeneity within these four cohorts was low (I2 = 16.1% in the unadjusted and I2 = 0% in the adjusted analysis; Table 5). Therefore, the significance of the association was enhanced by the meta-analysis. 
Table 5
 
The Meta-Analysis Results of Unadjusted and Adjusted Association of the Other SNPs in SIX1-SIX6 Locus
Table 5
 
The Meta-Analysis Results of Unadjusted and Adjusted Association of the Other SNPs in SIX1-SIX6 Locus
Publication Bias and Sensitivity Analysis
The Egger's test showed that all the meta-analyses of all the SNPs rejected the null hypothesis of publication bias (Egger's test P > 0.05; Tables 4 and 5, Supplementary Tables S3 and S4). In stratified analysis, there were only two studies investigated in some subgroups (West Asian and Latino in the unadjusted analysis, African and Caucasian in the rs33912345 adjusted analysis); we therefore did not evaluate the publication bias. In sensitivity analysis, we conducted sequential omission of individual cohorts, that is, removing one at a time and recalculating the pooled ORs. There was no SNP with alteration in the significance of the pooled OR in any of the analyses, suggesting that the associations results were robust. 
Discussion
In this meta-analysis, we revealed significant associations between POAG and three SNPs, rs10483727, rs33912345, and rs12436579, in the SIX1-SIX6 locus. Stratification analysis revealed that these associations remained significant in Caucasians and Asians, but not in Africans, suggesting population-specific effects. Sensitivity analysis suggested the robustness of the association results. Therefore, these SNPs should be genuine susceptibility genetic markers for POAG, at least for specific ethnic groups. Previous studies17,29,48,49 provided some insights into the mechanisms of the SIX6 gene and rs33912345 underlying POAG pathogenesis. Further functional studies are therefore warranted for the other variants in this locus. 
The pooled ORs of rs33912345 were higher than the ORs of rs10483727 for both unadjusted and adjusted associations in different population subgroups, except in the adjusted association among Caucasians (Supplementary Table S3 and S4, Table 4). SNP rs33912345 had stronger associations compared to rs10483727 in the unadjusted analyses but weaker in the adjusted ones. This might be caused by reduced sample size and lower statistical power in the adjusted analysis: 7162 cases and 10,081 controls for rs33912345 versus 11,889 cases and 26,028 controls for rs10483727. SNP rs10483727 was the first SIX1-SIX6 SNP identified with POAG association,14,31 and rs33912345 was more frequently investigated in recent studies.23,37,39 SNP rs10483727 is located in the intergenic region between SIX1 and SIX6, whereas rs33912345 is a missense variant (p.His141Asn) in SIX6. The two SNPs are in strong linkage disequilibrium (LD) (R2 > 0.96; 1000 Genomes Project Phase 3 v5, provided in the public domain and accessed from http://www.internationalgenome.org/home). Additionally, a Japanese GWAS exhibited the association of rs6573307 and POAG (Supplementary Table S2).18 This SNP is also in strong LD, with rs33912345 (R2 = 0.903) and rs10483727 (R2 = 0.879) (1000 Genomes Project Phase 3 v5) in Japanese populations. These findings suggested that the LD block (Supplementary Fig. S1) defined by SNPs rs33912345 and rs10483727 could be the responsible region in the SIX1-SIX6 locus for POAG. 
The human SIX gene family consists of six members (SIX1 to SIX6), all of which contain two shared protein domains: a DNA-binding homeobox domain and a SIX domain, which encode homeobox protein transcription factors and may be involved in regulating the development of the visual system.50 Recent studies showed that the risk allele of the SIX6 missense variant rs33912345 was associated with RNFL thinning,29,35,51 which is a typical pathologic feature of glaucoma.7 This association suggested that this variant might affect neural retina development or degeneration, such as RGC death in glaucoma pathophysiology. The impacts of this variant were also observed on early eye development in zebrafish29,48,49 and RGC death under elevated intraocular pressure (IOP) in a mouse model, possibly by affecting p16INK4a (also known as CDNK2A) expression.17 Loss of SIX6 homologs also upregulated CDNK2A/CDNK2B in zebrafish. Of note, SNPs in the CDNK2B-AS1 locus were also associated with POAG.13,18,19,22,52 Based on the genetic and biological evidence, this association is more likely triggered by the alterations in SIX6 through the mechanism involving CDNK2A/CDNK2B and RGC loss. Another explanation is the effect from the haplotype, in which multiple variants that were inherited together may exert impact on the disease risk together. The SNP rs12436579 is 3.5 kb downstream from the SIX6 coding region. The linkage between rs12436579 and the two SNPs above is relatively weak (R2 = 0.38 with rs10483727 and R2 = 0.40 with rs33912345, ALL population in 1000 Genomes Project Phase 3 v5). However, the pooled effect sizes of the three SNPs are comparable (Supplementary Tables S3 and S4, Tables 4 and 5). Therefore, which of these SNPs is responsible for the POAG association in this region remains uncertain. Also, the LD block including rs10483727 and rs33912345 could be a proxy for the causal variant that is yet to be identified. Nevertheless, rs33912345 (His141Asn) might participate in the potential mechanism underlying POAG according to the previous evidence.17 
In this study, the stratification analysis revealed different associations across populations. SNPs rs10483727 and rs33912345 were significantly associated with POAG in East Asian and Caucasian populations under all genetic models. Meanwhile, rs10483727 was associated with POAG in Saudi Arabian populations25 in West Asia, and rs33912345 had association in Mexican populations (Latino).17 However, there was only one study reported in each one of these populations, and further validation is therefore required to confirm the associations. In contrast, SNPs rs10483727 and rs33912345 were not associated with POAG in South Asians (Indian and Pakistani)33,36,37 and the populations with African ancestry (Ghanaian, African American, and Barbadian).32,34 These findings suggested that the association of the SNPs in SIX1-SIX6 with POAG could be specific in certain ethnic populations, such as Caucasians and East Asians. Population-specific effects of these SNPs might be the result of different genetic architectures and interactions with local environmental factors. Minor allele frequencies of both SNPs were low in African (0.032 and 0.034; 1000 Genomes Project Phase 3 v5); therefore, a larger sample size is needed for sufficient statistical power to detect a significant association. Boundaries of haplotype blocks may vary across different populations; in particular, the block size (>22 kb) in African cohorts is small, whereas the block sizes in Caucasian and Asian cohorts are relatively large and comparable (>44 kb).53 This implies that the linkage in this region might be weaker in African compared to Caucasian and Asian cohorts. Similarly, the associations of the SNPs, such as rs4977756, in the CDKN2B-AS1 gene reported by GWASs13,18,19,22,52 could not be replicated in some African cohorts,32,34,46 but another two SNPs (rs1063192 and rs10120688) in this locus showed significant associations in Africans.32,34 Both rs1063192 and rs10120688 have stronger LD with the GWAS SNP rs4977756 in European (R2 = 0.778 and 0.643, respectively) than in African cohorts (R2 = 0.0264 and 0.327, respectively; 1000 Genomes Project Phase 3 v5). Therefore, different LD block structures in different ethnic groups might have altered the proxies for the causal variants for the disease, leading to different association patterns. Moreover, environmental factors linked to POAG risk might vary regionally and have different interaction with the genetic locus in glaucoma pathogenesis. Clinical assessment variabilities might lead to the differential association across the studies. Two of the South Asian studies36,37 recruited only POAG cases with IOP > 21 mm Hg, while the other included studies had both high- and normal-tension glaucoma patients. Besides POAG, population-specific genetic effects among Caucasian, Asian, and African populations have also been reported in other diseases (e.g., type 2 diabetes54 and asthma55,56) and quantitative traits (e.g., body mass index57 and plasma B12 concentration58). However, there could be other reasons for such discrepancies, including sampling or selection bias and/or limited sample sizes with insufficient power. Further studies in those populations showing negative results are warranted to confirm the ethnic diversities of SNPs in SIX1-SIX6 in POAG. 
This meta-analysis study provided an increased statistical power to estimate the associations of SIX1-SIX6 polymorphisms with POAG from the published genetic studies. However, there were several limitations. First, this study was mainly based on the East Asian (Japanese and Chinese) and Caucasian populations. It may restrict our conclusions and indicate the need for studies in other ethnic populations. Second, there were high heterogeneities in stratified analyses for some populations in certain genetic models. It is possible that there was underlying ethnic differences in the association, but it may also be due to different clinical characteristics of the patients and/or various definitions of controls. Third, POAG is a multifactorial disease resulting from interactions between various risk factors. Therefore, our results may be influenced by confounding factors, such as age, gender, environment, and lifestyle. Future investigation of gene-environment interactions in different ethnic subgroups should be carried out to acquire more conclusive claims about the association of the polymorphisms in SIX1-SIX6 with POAG. 
In conclusion, this meta-analysis showed that SNPs rs10483727, rs33912345, and rs12436579 in the SIX1-SIX6 locus are associated with POAG, suggesting that this locus may play an important role in affecting individual susceptibility to the disease. The associations of rs10483727 and rs33912345 were widely investigated in different populations. Our stratification analysis demonstrated the significant correlations between the two SNPs and POAG in East Asian and Caucasian, but not in African, cohorts. The associations in some cohorts, such as South Asian, still need to be further confirmed because of high heterogeneity within the reported studies. The association of SNP rs12436579 with POAG is worth validating in other populations. Also, further functional investigations are warranted to understand the mechanism of the SIX1-SIX6 locus in POAG pathogenesis. 
Acknowledgments
Supported by grants from the Natural Science Foundation of China (81670853 and 81371048) (BG); grants from Department of Science and Technology of Sichuan Province (2019JDJQ0031) (BG); the General Research Fund, Hong Kong (14100917 and 14112514) (LJC); foundation for Technology & Science & Technology Bureau of Chengdu (2018-YF05-00348-SN) (BG); and the Endowment Fund for Lim Por-Yen Eye Genetics Research Centre, Hong Kong. 
Disclosure: S.Y. Lu, None; Z.Z. He, None; J.X. Xu, None; C. Yang, None; L.J. Chen, None; B. Gong, None 
References
Weinreb RN, Aung T, Medeiros FA. The pathophysiology and treatment of glaucoma: a review. JAMA. 2014; 311: 1901–1911.
Kingman S. Glaucoma is second leading cause of blindness globally. Bull World Health Organ. 2004; 82: 887–888.
Quigley HA. Glaucoma. Lancet. 2011; 377: 1367–1377.
Kwon YH, Fingert JH, Kuehn MH, Alward WL. Primary open-angle glaucoma. N Engl J Med. 2009; 360: 1113–1124.
Tham YC, Li X, Wong TY, Quigley HA, Aung T, Cheng CY. Global prevalence of glaucoma and projections of glaucoma burden through 2040: a systematic review and meta-analysis. Ophthalmology. 2014; 121: 2081–2090.
Quigley HA, Broman AT. The number of people with glaucoma worldwide in 2010 and 2020. Br J Ophthalmol. 2006; 90: 262–267.
Weinreb RN, Leung CK, Crowston JG, et al. Primary open-angle glaucoma. Nat Rev Dis Primers. 2016; 2: 16067.
Fingert JH. Primary open-angle glaucoma genes. Eye (Lond). 2011; 25: 587–595.
Stone EM, Fingert JH, Alward WL, et al. Identification of a gene that causes primary open angle glaucoma. Science. 1997; 275: 668–670.
Rezaie T, Child A, Hitchings R, et al. Adult-onset primary open-angle glaucoma caused by mutations in optineurin. Science. 2002; 295: 1077–1079.
Footz TK, Johnson JL, Dubois S, Boivin N, Raymond V, Walter MA. Glaucoma-associated WDR36 variants encode functional defects in a yeast model system. Hum Mol Genet. 2009; 18: 1276–1287.
Thorleifsson G, Walters GB, Hewitt AW, et al. Common variants near CAV1 and CAV2 are associated with primary open-angle glaucoma. Nat Genet. 2010; 42: 906–909.
Burdon KP, Macgregor S, Hewitt AW, et al. Genome-wide association study identifies susceptibility loci for open angle glaucoma at TMCO1 and CDKN2B-AS1. Nat Genet. 2011; 43: 574–578.
Wiggs JL, Yaspan BL, Hauser MA, et al. Common variants at 9p21 and 8q22 are associated with increased susceptibility to optic nerve degeneration in glaucoma. PLoS Genet. 2012; 8: e1002654.
Chen Y, Lin Y, Vithana EN, et al. Common variants near ABCA1 and in PMM2 are associated with primary open-angle glaucoma. Nat Genet. 2014; 46: 1115–1119.
Gharahkhani P, Burdon KP, Fogarty R, et al. Common variants near ABCA1, AFAP1 and GMDS confer risk of primary open-angle glaucoma. Nat Genet. 2014; 46: 1120–1125.
Skowronska-Krawczyk D, Zhao L, Zhu J, et al. P16INK4a upregulation mediated by SIX6 defines retinal ganglion cell pathogenesis in glaucoma. Mol Cell. 2015; 59: 931–940.
Shiga Y, Akiyama M, Nishiguchi KM, et al. Genome-wide association study identifies seven novel susceptibility loci for primary open-angle glaucoma. Hum Mol Genet. 2018; 27: 1486–1496.
Cooke Bailey JN, Loomis SJ, Kang JH, et al. Genome-wide association analysis identifies TXNRD2, ATXN2 and FOXC1 as susceptibility loci for primary open-angle glaucoma. Nat Genet. 2016; 48: 189–194.
Hysi PG, Cheng CY, Springelkamp H, et al. Genome-wide analysis of multi-ancestry cohorts identifies new loci influencing intraocular pressure and susceptibility to glaucoma. Nat Genet. 2014; 46: 1126–1130.
Abu-Amero K, Kondkar AA, Chalam KV. An updated review on the genetics of primary open angle glaucoma. Int J Mol Sci. 2015; 16: 28886–28911.
Osman W, Low SK, Takahashi A, Kubo M, Nakamura Y. A genome-wide association study in the Japanese population confirms 9p21 and 14q23 as susceptibility loci for primary open angle glaucoma. Hum Mol Genet. 2012; 21: 2836–2842.
Shiga Y, Nishiguchi KM, Kawai Y, et al. Genetic analysis of Japanese primary open-angle glaucoma patients and clinical characterization of risk alleles near CDKN2B-AS1, SIX6 and GAS7. PLoS One. 2017; 12: e0186678.
Yoshikawa M, Nakanishi H, Yamashiro K, et al. Association of glaucoma-susceptible genes to regional circumpapillary retinal nerve fiber layer thickness and visual field defects. Invest Ophthalmol Vis Sci. 2017; 58: 2510–2519.
Kondkar AA, Azad TA, Almobarak FA, et al. Polymorphism rs10483727 in the SIX1/SIX6 gene locus is a risk factor for primary open angle glaucoma in a Saudi cohort. Genet Test Mol Biomarkers. 2018; 22: 74–78.
Sang J, Jia L, Zhao B, Wang H, Zhang N, Wang N. Association of three single nucleotide polymorphisms at the SIX1-SIX6 locus with primary open angle glaucoma in the Chinese population. Sci China Life Sci. 2016; 59: 694–699.
Chen Y, Hughes G, Chen X, et al. Genetic variants associated with different risks for high tension glaucoma and normal tension glaucoma in a Chinese population. Invest Ophthalmol Vis Sci. 2015; 56: 2595–2600.
Burdon KP, Mitchell P, Lee A, et al. Association of open-angle glaucoma loci with incident glaucoma in the Blue Mountains Eye Study. Am J Ophthalmol. 2015; 159: 31–36.
Ulmer Carnes M, Liu YFP, Allingham RR, et al. Discovery and functional annotation of SIX6 variants in primary open-angle glaucoma. PLoS Genet. 2014; 10: e1004372.
Dimasi DP, Burdon KP, Hewitt AW, et al. Genetic investigation into the endophenotypic status of central corneal thickness and optic disc parameters in relation to open-angle glaucoma. Am J Ophthalmol. 2012; 154: 833–842.
Fan BJ, Wang DY, Pasquale LR, Haines JL, Wiggs JL. Genetic variants associated with optic nerve vertical cup-to-disc ratio are risk factors for primary open angle glaucoma in a US Caucasian population. Invest Ophthalmol Vis Sci. 2011; 52: 1788–1792.
Liu Y, Hauser MA, Akafo SK, et al. Investigation of known genetic risk factors for primary open angle glaucoma in two populations of African ancestry. Invest Ophthalmol Vis Sci. 2013; 54: 6248–6254.
Shah MH, Tabanera N, Krishnadas SR, Pillai MR, Bovolenta P, Sundaresan P. Identification and characterization of variants and a novel 4 bp deletion in the regulatory region of SIX6, a risk factor for primary open-angle glaucoma. Mol Genet Genomic Med. 2017; 5: 323–335.
Cao D, Jiao X, Liu X, et al. CDKN2B polymorphism is associated with primary open-angle glaucoma (POAG) in the Afro-Caribbean population of Barbados, West Indies. PLoS One. 2012; 7: e39278.
Cheng CY, Allingham RR, Aung T, et al. Association of common SIX6 polymorphisms with peripapillary retinal nerve fiber layer thickness: the Singapore Chinese Eye Study. Invest Ophthalmol Vis Sci. 2014; 56: 478–483.
Micheal S, Ayub H, Khan MI, et al. Association of known common genetic variants with primary open angle, primary angle closure, and pseudoexfoliation glaucoma in Pakistani cohorts. Mol Vis. 2014; 20: 1471–1479.
Mohanty K, Dada R, Dada T. Identification and genotype phenotype correlation of novel mutations in SIX6 gene in primary open angle glaucoma. Ophthalmic Genet. 2018; 39: 366–372.
Rong SS, Lu SY, Matsushita K, et al. Association of the SIX6 locus with primary open angle glaucoma in southern Chinese and Japanese. Exp Eye Res. 2018; 180: 129–136.
Yoshikawa M, Yamashiro K, Nakanishi H, et al. Association of SIX1/SIX6 locus polymorphisms with regional circumpapillary retinal nerve fibre layer thickness: The Nagahama study. Sci Rep. 2017; 7: 4393.
Aijaz S, Allen J, Tregidgo R, van Heyningen V, Hanson I, Clark BJ. Expression analysis of SIX3 and SIX6 in human tissues reveals differences in expression and a novel correlation between the expression of SIX3 and the genes encoding isocitrate dehydrogenase and cadherin 18. Genomics. 2005; 86: 86–99.
Cheyette BN, Green PJ, Martin K, Garren H, Hartenstein V, Zipursky SL. The Drosophila sine oculis locus encodes a homeodomain-containing protein required for the development of the entire visual system. Neuron. 1994; 12: 977–996.
Halder G, Callaerts P, Flister S, Walldorf U, Kloter U, Gehring WJ. Eyeless initiates the expression of both sine oculis and eyes absent during Drosophila compound eye development. Development. 1998; 125: 2181–2191.
Aldahmesh MA, Khan AO, Hijazi H, Alkuraya FS. Homozygous truncation of SIX6 causes complex microphthalmia in humans. Clinical Genet. 2013; 84: 198–199.
Gallardo ME, Rodriguez De Cordoba S, Schneider AS, Dwyer MA, Ayuso C, Bovolenta P. Analysis of the developmental SIX6 homeobox gene in patients with anophthalmia/microphthalmia. Am J Med Genet A. 2004; 129a: 92–94.
Hedges LV, Olkin I. Statistical Methods for Meta-Analysis. Cambridge, MA: Academic Press; 1985.
Williams SEI, Carmichael TR, Allingham RR, Hauser M, Ramsay M. The genetics of POAG in black South Africans: a candidate gene association study. Sci Rep. 2015; 5: 8378.
Mabuchi F, Sakurada Y, Kashiwagi K, Yamagata Z, Iijima H, Tsukahara S. Involvement of genetic variants associated with primary open-angle glaucoma in pathogenic mechanisms and family history of glaucoma. Am J Ophthalmol. 2015; 159: 437–444.
Iglesias AI, Springelkamp H, Van der linde H, et al. Exome sequencing and functional analyses suggest that SIX6 is a gene involved in an altered proliferation-differentiation balance early in life and optic nerve degeneration at old age. Hum Mol Genet. 2014; 23: 1320–1332.
Springelkamp H, Iglesias AI, Mishra A, et al. New insights into the genetics of primary open-angle glaucoma based on meta-analyses of intraocular pressure and optic disc characteristics. Hum Mol Genet. 2017; 26: 438–453.
Kawakami K, Sato S, Ozaki H, Ikeda K. Six family genes–structure and function as transcription factors and their roles in development. Bioessays. 2000; 22: 616–626.
Kuo JZ, Zangwill LM, Medeiros FA, et al. Quantitative trait locus analysis of SIX1-SIX6 with retinal nerve fiber layer thickness in individuals of European descent. Am J Ophthalmol. 2015; 160: 123–130.
Li Z, Allingham RR, Nakano M, et al. A common variant near TGFBR3 is associated with primary open angle glaucoma. Hum Mol Genet. 2015; 24: 3880–3892.
Gabriel SB, Schaffner SF, Nguyen H, et al. The structure of haplotype blocks in the human genome. Science. 2002; 296: 2225–2229.
Cook JP, Morris AP. Multi-ethnic genome-wide association study identifies novel locus for type 2 diabetes susceptibility. Eur J Hum Genet. 2016; 24: 1175–1180.
Igartua C, Myers RA, Mathias RA, et al. Ethnic-specific associations of rare and low-frequency DNA sequence variants with asthma. Nat Commun. 2015; 6: 5965.
Torgerson DG, Ampleford EJ, Chiu GY, et al. Meta-analysis of genome-wide association studies of asthma in ethnically diverse North American populations. Nat Genet. 2011; 43: 887–892.
Salinas YD, Wang L, DeWan AT. Multiethnic genome-wide association study identifies ethnic-specific associations with body mass index in Hispanics and African Americans. BMC Genet. 2016; 17: 78.
Nongmaithem SS, Joglekar CV, Krishnaveni GV, et al. GWAS identifies population-specific new regulatory variants in FUT6 associated with plasma B12 concentrations in Indians. Hum Mol Genet. 2017; 26: 2551–2564.
Figure 1
 
Flow diagram of the study inclusion process on the association between the SIX6 locus and POAG.
Figure 1
 
Flow diagram of the study inclusion process on the association between the SIX6 locus and POAG.
Figure 2
 
Forest plot of the association between rs10483727 and POAG expressed as unadjusted OR. Squares indicate risk estimate in each study; horizontal lines represent 95% CIs and ORs; diamonds indicate summary estimate OR with corresponding 95% CI in each subgroup or pooled overall population. Effects of the SNP were weighted by inverse variance under fixed or random-effects model.
Figure 2
 
Forest plot of the association between rs10483727 and POAG expressed as unadjusted OR. Squares indicate risk estimate in each study; horizontal lines represent 95% CIs and ORs; diamonds indicate summary estimate OR with corresponding 95% CI in each subgroup or pooled overall population. Effects of the SNP were weighted by inverse variance under fixed or random-effects model.
Figure 3
 
Forest plot of the association between rs33912345 and POAG expressed as unadjusted OR. Squares indicate risk estimate in each study; horizontal lines represent 95% CIs and ORs; diamonds indicate summary estimate OR with corresponding 95% CI in each subgroup or pooled overall population. Effects of the SNP were weighted by inverse variance under fixed or random-effects model.
Figure 3
 
Forest plot of the association between rs33912345 and POAG expressed as unadjusted OR. Squares indicate risk estimate in each study; horizontal lines represent 95% CIs and ORs; diamonds indicate summary estimate OR with corresponding 95% CI in each subgroup or pooled overall population. Effects of the SNP were weighted by inverse variance under fixed or random-effects model.
Table 1
 
Characteristics of Studies Included for the Meta-Analysis
Table 1
 
Characteristics of Studies Included for the Meta-Analysis
Table 2
 
Allelic Frequencies and Effect Sizes of rs10483727 Among POAG Cases and Controls in the 16 Included Studies
Table 2
 
Allelic Frequencies and Effect Sizes of rs10483727 Among POAG Cases and Controls in the 16 Included Studies
Table 3
 
Allelic Frequencies and Effect Sizes of rs33912345 Among POAG Cases and Controls in the 12 Included Studies
Table 3
 
Allelic Frequencies and Effect Sizes of rs33912345 Among POAG Cases and Controls in the 12 Included Studies
Table 4
 
The Meta-Analysis Results of Adjusted Association in Overall and Different Population Subgroups*
Table 4
 
The Meta-Analysis Results of Adjusted Association in Overall and Different Population Subgroups*
Table 5
 
The Meta-Analysis Results of Unadjusted and Adjusted Association of the Other SNPs in SIX1-SIX6 Locus
Table 5
 
The Meta-Analysis Results of Unadjusted and Adjusted Association of the Other SNPs in SIX1-SIX6 Locus
Supplement 1
Supplement 2
×
×

This PDF is available to Subscribers Only

Sign in or purchase a subscription to access this content. ×

You must be signed into an individual account to use this feature.

×