Molecular Vision 2017; 23:1048-1080
<http://www.molvis.org/molvis/v23/1048>
Received 08 May 2017 |
Accepted 29 December 2017 |
Published 31 December 2017
Jiali Li, Qingjiong Zhang
State Key Laboratory of Ophthalmology, Zhongshan Ophthalmic Center, Sun Yat-sen University, Guangzhou, China
Correspondence to: Qingjiong Zhang, State Key Laboratory of Ophthalmology, Zhongshan Ophthalmic Center, Sun Yat-sen University, 54 Xianlie Road, Guangzhou 510060, China; Phone: (+86)-20-87333393; FAX: (+86)-20-87333271; email: zhangqji@mail.sysu.edu.cn or zhangqingjiong@gzzoc.com
Myopia is the most common cause of visual impairment worldwide. Genetic and environmental factors contribute to the development of myopia. Studies on the molecular genetics of myopia are well established and have implicated the important role of genetic factors. With linkage analysis, association studies, sequencing analysis, and experimental myopia studies, many of the loci and genes associated with myopia have been identified. Thus far, there has been no systemic review of the loci and genes related to non-syndromic and syndromic myopia based on the different approaches. Such a systemic review of the molecular genetics of myopia will provide clues to identify additional plausible genes for myopia and help us to understand the molecular mechanisms underlying myopia. This paper reviews recent genetic studies on myopia, summarizes all possible reported genes and loci related to myopia, and suggests implications for future studies on the molecular genetics of myopia.
Myopia is the most common cause of visual impairment worldwide. Myopia is a condition in which parallel light passes through the eye and focuses in front of the retina. Myopia can be classified as common myopia with refractive error less than −6 diopters (D) and high myopia with refractive error equal or greater than −6 D. Myopia can sometimes be an accompanying symptom in other diseases. Thus, myopia can also be classified as non-syndromic myopia (if it occurs alone) and syndromic myopia (if it is an associated sign of another ocular or systemic disease). As common myopia is too common to be a sign of a specific disease, usually high myopia has been reported as a specific phenotype of a syndrome or other ocular or systemic diseases; thus, syndromic myopia usually means syndromic high myopia. Myopia may also be classified as physiologic myopia (usually low-grade myopia) and pathological myopia (mostly associated with degenerative changes in the retina). Sometimes, common or physiologic myopia in early childhood may develop as high myopia or pathological myopia in adult or older individuals. Most individuals with high myopia may not have pathological changes in the retina, especially in younger ages.
Genetic and environmental factors are well known components that contribute to the development of myopia. Common myopia is more likely to be a complex trait, resulting from the effects of genetic and environmental factors. High myopia may be transmitted as a complex trait (especially late-onset high myopia commonly seen in university students) or a Mendelian trait (such as most early-onset high myopia that is not related to extensive near work) [1]. Efforts to decipher the hereditary determinants of myopia began in the 1960s, and the important role of genetic factors has been implicated in several studies, including familial aggregation, pedigree analysis, twin studies, and population studies. Until now, many loci and genes associated with myopia have been identified with linkage analysis, association studies, whole exome sequencing, and experimental myopia studies. This paper reviews the genetic determinations of myopia and the molecular genetics of non-syndromic and syndromic myopia according to the different approaches.
2.1 Hereditability and familial aggregation of myopia-- Familial aggregation has provided strong evidence to support the important role of genetic factors in causing the pathogenesis of myopia [2-6]. Studies on individuals with myopia from different populations have shown positive indications. For example, Yap et al. performed refraction examinations on 2,888 children in China to investigate the risks of passing myopia from parents to children [2]. This study showed that children with myopic parents tend to have a greater chance of developing myopia than those without myopic parents. Several subsequent studies also supported this finding. Ip et al. conducted a similar study on 2,353 children in Australia with a mean age of 7 years [4]. The results suggested that the prevalence of myopia increased statistically significantly with the number of myopic parents; the associated development risk rates are 7.6% for no myopic parents, 14.9% for one myopic parent, and 43.6% for two myopic parents. In addition, the recurrence risks of myopia among siblings with myopia are also higher than siblings without myopia. A report based on 6,497 inhabitants in Tehran, Iran, confirmed that the heritability of refractive error is 61%, and the ratio of sibling recurrence odds for myopia is 2.25–3.00 [6]. Apart from refraction status, the impacts of myopia in parental history on children’s ocular growth were also observed. Lam et al. followed up with 7,560 children in China for 1 year and reported the effects of myopia in parental history, including eye size and growth. The change in annual axial length increased with the number of myopic parents: 0.20 mm, 0.26 mm, and 0.37 mm for children with no, one, and two myopic parents, respectively [3]. This line of evidence suggests that the genetic risks are higher in families with myopia.
2.2 Genetic transmission traits of myopia-- Pedigree analysis is widely used in myopia and high myopia studies. High myopia is often found to be transmitted through families in Mendelian patterns, including autosomal dominant (AD), autosomal recessive (AR), and X-linked recessive (XL) inheritance. Based on linkage analysis, 18 myopia and high myopia loci have been discovered and documented in the Online Mendelian Inheritance in Man database (OMIM), including nine high myopia loci in AD inheritance (MYP2: Gene ID 4658, OMIM 160700 [7], MYP3: Gene ID 8782, OMIM 603221 [8], MYP5: Gene ID 404682, OMIM 608474 [9], MYP11: Gene ID 594832, OMIM 609994 [10], MYP12: Gene ID 664780, OMIM 609995 [11], MYP15: Gene ID 100294716, OMIM 612717 [12], MYP16: Gene ID 100270641, OMIM 612554 [13], MYP17: Gene ID 100359401, OMIM 608367 [14], and MYP19: Gene ID 100653370, OMIM 613969 [15]), one high myopia locus in AR inheritance (MYP18, Gene ID 100359406, OMIM 255500 [16]), two XL recessive high myopia loci (MYP1: Gene ID 4657, OMIM 310460 [17] and MYP13: Gene ID 677764, OMIM 300613 [18]), and six myopia loci (MYP6: Gene ID 9997, OMIM 608908 [19], MYP7: Gene ID 553190, OMIM 609256 [20], MYP8: Gene ID 553192, OMIM 609257 [20], MYP9: Gene ID 553194, OMIM 609258 [20], MYP10: Gene ID 553195, OMIM 609259 [20], and MYP14: Gene ID 100359407, OMIM 610320 [21]). The identification of these 18 loci in the human genome through linkage studies not only indicates the contribution of genetic factors to myopia but also provides clues for further screening of candidate genes.
2.3 Twin study: Monozygotic twins have more similar refraction than dizygotic twins-- A twin study is an ideal way to estimate the hereditary component of myopia. This study design provides cogent evidence for the important role of genetic factors in myopia. Monozygotic twins are identical in genetic material, while dizygotic twins share 50% of their genetic material. Therefore, monozygotic twins are considered to have more similarity in phenotype for genetic diseases. A comparison of similarity in refraction power between monozygotic and dizygotic twins can be used to evaluate the heritability of myopia. Karlsson et al. systemically reviewed a series of twin studies and found that 95% of monozygotic twins appear to have a similar refraction power, and only 29% of dizygotic twins have such similarity [22]. Subsequently, several large-scale twin studies from different countries confirmed this finding, and the heritability varied from 75% to 94% [23-25]. In addition, the impacts of genetic and environmental factors on myopia have been observed by analyzing refraction power in 1,152 monozygotic twins and 1,149 dizygotic twins: 77% of the difference in refraction power was explained by genetic components, while 7% of the difference was explained by environmental factors [26]. This result once again supports the importance of genetic factors in the development of myopia.
2.4 SNPs in myopia-related genes from population-based association studies-- Population-based association studies include genome-wide association studies (GWASs) and case-control studies. Population-based association studies have identified that many single nucleotide polymorphisms (SNPs) are statistically significantly associated with myopia, suggesting the involvement of multiple gene effects. The traits used in an association study include spherical equivalent (SE), axial length, and corneal curvature. Although there are some controversial issues and arguments related to the results of association studies, some SNPs have been replicated and confirmed to be statistically significantly associated with myopia or high myopia in separate independent studies from different populations. The replication of significant SNPs is necessary before further analysis is conducted. Disease-associated SNPs would be more convincing if they could be replicated in different large-scale GWASs. For example, Kiefer et al. [27] and Verhoeven et al. [28] identified several significant associations of SNPs in 45,771 and 45,758 participants with myopia and refractive error, respectively. Twelve SNPs were discovered in these two different studies, which provided strong evidence that the 12 SNPs are statistically significantly associated with myopia. Association studies, similar to pedigree analysis, have indicated the importance of genetic factors in myopia and provided clues for discovering new causative genes.
The role of molecular genetics in myopia has been investigated mainly in population- and family-based studies. GWASs, as one of the major forms, have been widely used to identify associations between the characteristics of myopia (i.e., refraction, axial length, or corneal curvature) and genetic variants across the whole genome. Meanwhile, linkage analysis was traditionally applied in family-based studies to exploit linkage regions within families, especially for Mendelian high myopia. Aside from the approaches above, whole exome sequencing and experimental myopia studies have also been applied to reveal the molecular genetic basis of myopia. Overall, molecular genetic studies on myopia have identified numerous loci and genes, which are of great importance in implicating the mechanisms underlying myopia.
3.1 Linkage analysis of the genetic loci for myopia-- Myopia is generally heterogeneous. Through linkage analysis of families with myopia and high myopia, about 18 loci have been identified. The details of these 18 loci were mentioned above and are listed in Table 1. The candidate genes inside the linkage intervals have been screened based on their potential functions involved in the development of myopia.
3.1.1 Susceptibility loci for common myopia-- Common myopia is usually considered a complex trait because genetic and environmental factors contribute to the susceptibility risk. The genetic mapping of complex traits has identified six susceptibility loci for myopia. Among them, four loci (MYP7–MYP10) were identified in a genome-wide scan with a multipoint linkage analysis in 226 monozygotic and 280 dizygotic twins, using refraction as a quantitative trait ranging from −12.12 D to + 7.25 D [20]. In these four loci, MYP10 was replicated in a GWAS with high myopia (rs189798) [29], and MYP7 was replicated in another genome-wide scan in dizygotic twins [20,30].
For two other loci, MYP6 and MYP14, MYP6 was mapped in a genome-wide scan in 44 Ashkenazi Jewish families with mild to moderate myopia [19]. The scan was replicated in another genome-wide scan of 486 extended families with refractions ranging from −12.13 D to +8.38 D [31]. MYP14 was identified in a genome-wide scan with multipoint regression-based quantitative trait locus (QTL) linkage in 49 Ashkenazi Jewish families (mean SE=−3.46 D ± 3.29 D) [21]. In a large international collaborative study of high myopia, MYP14 (1p36.32) was replicated in a subset of 24 families from Denmark with a maximum logarithm (base 10) of odds (LOD) score of 1.8 at rs1870509 in a multipoint linkage analysis [32].
Candidate genes within these six myopia loci have been screened according to their related functions. PAX6 (Gene ID: 5080, OMIM: 607108), located in the MYP7 locus, has been of great concern and research interest. PAX6 is an important transcriptional regulator involved in the development of the eye [33] from the surface ectoderm to the cornea and the lens [34,35] and from the neuroectoderm to the iris, ciliary body, and retina [36,37]. Mutations in PAX6 are associated with a wide range of abnormalities in humans, including aniridia, Peter’s anomaly, microcornea, cataract, coloboma of the optic nerve, microphthalmia, and nystagmus [38-43]. Studies have reported an association between PAX6 and high myopia [44-49], but these results are not consistent with other studies [50-52]. Therefore, PAX6 should be evaluated with caution as a candidate gene for myopia. MYP8, MYP9, and MYP10 have not been investigated yet as candidate genes.
3.1.2 Genetic loci and genes identified in families with high myopia-- In genetic analysis of non-syndromic high myopia, 12 loci have been identified using linkage studies, as mentioned above (Table 1). These loci were frequently observed in families with Mendelian high myopia, characterized by onset in early childhood (≤7 years old) and monogenic inheritance. Several loci have been replicated or refined in independent families using the same approach.
Of these 12 loci, MYP1, located at Xq28, is a locus for non-syndromic and syndromic high myopia (the latter will be discussed in the section on syndromic myopia). In studies on non-syndromic high myopia, MYP1 was mapped in a Chinese family [17] with high myopia, and the finding was replicated in two Indian families [53]. MYP3 was revealed by Young et al. in a large German–Italian family [8] and was further replicated in at least four independent studies involving different populations [32,54-56]. Moreover, a genome-wide scan in 254 families with high myopia also discovered the MYP1, MYP3, MYP6, MYP11, MYP12, and MYP14 loci and identified a novel locus at chromosome 9q34.11 [32]. In addition, several loci have been replicated in association studies on myopia, including MYP11, MYP15, MYP16, and MYP17. However, compared with the high prevalence of high myopia in the general population, these 12 loci account for fewer than 5% of individuals with high myopia, suggesting that more loci and genes await to be identified.
Numerous candidate genes in these 12 loci have been screened in different kinds of studies, including direct sequencing, association analysis, and experimental animal analysis. Although several candidate genes or SNPs have been investigated frequently in association studies, inconsistent results were often noted. No mutations in candidate genes have been identified in mapped families using Sanger sequencing (Table 2) [10,11,13,15,17,53,57-59]. For example, TGIF (Gene ID 7050, OMIM 602630), located in the MYP2 locus, is expressed strongly in the sclera, retina, and optic nerve. TGIF is involved in the regulation of the transforming growth factor (TGF)-beta pathway [60], which has a close relationship with the development of myopia [61,62]. However, screening mutations in TGIF among affected members in families mapped to MYP2 did not identify any cosegregated mutations; no significant associations between SNPs in TGIF and high myopia were reported [3,63,64]. Although knockout of LUM (Gene ID 4060, OMIM 600616; located in MYP3) in mice showed a series of myopia-related phenotypes [65], Sanger sequencing of LUM in an MYP3-mapped family did not identify any mutations [57]. Other candidate genes within these loci are not listed here because of their undetermined relationship with myopia. The exact mutations among these genes need to be further studied. Especially with the method of whole exome sequencing, which fully screens all genes within linkage regions for mapped families, new genes are expected to be revealed in the future.
3.2 Genetic association studies on myopia-- To investigate the genetic factors for myopia in the general population, GWASs and a series of replications in follow-up association studies have identified numerous associations with myopia, although some are uncertain. A literature search was conducted in PubMed (until 11/18/2016) using the following terms: (association) and myopia/genetics, myopia and gene and association, and (association) and refractive errors/genetics. Only association studies related to myopia are included and reviewed in this section. There were some controversial issues in the GWASs and case-control association studies, such as the level of the test and the sample sizes. Thus, this part of the review mainly focuses on genetic loci that were detected by GWASs and met the significance standard of p<5 × 10−8.
3.2.1 Evidence of associations between polymorphisms and common myopia-- Thus far, around 82 loci for myopia have been examined in GWASs and case-control association studies using refractions between −0.50 D and −6.00 D as parameters (Appendix 1) [3,27,52,66-94]. Of these 82 loci, 41 loci were identified by seven GWAS and GWAS meta-analyses (Appendix 1) that met the requirement of p<5 × 10−8. An important principle for an association study is replication, replication, and replication. The 15q14 locus has been confirmed frequently in replication studies over the past few years (Table 3). Until now, about 20 association studies have aimed to test the relationship between the 15q14 locus and myopia in different populations, and only three showed no statistically significant association due to the small size of the samples. The SNP rs634990, reaching the greatest p value of 9.20 × 10−23, was identified by a GWAS meta-analysis in 55,177 samples in Caucasian and Asian populations. Known genes near the 15q14 locus include GJD2 (OMIM 607058), ACTC1 (Gene ID 70, OMIM 102540), GOLGA8B (Gene ID 440270, OMIM 609619), LPCAT4 (Gene ID 254531, OMIM 612039), and CHRM5 (Gene ID 1133, OMIM 118496). Of these genes, gap junction protein, delta 2 (GJD2), is located closest to the most significant SNP rs634900. GJD2 was first screened in 47 individuals with refractive error, but no pathogenic variants were identified. Later in our previous study, a sequence analysis of GJD2, which had been detected in significant association with myopia in the 15q14 locus within the cohort, did not identify any causative variants. As for the other genes in the 15q14 locus, none have been reported in statistically significant association with myopia. Therefore, the mechanism underlying this significant p value needs to be studied further.
An important and often mentioned discovery is related to two recent large-scale GWASs, which identified 24 and 22 statistically significant associations from independent cohorts. One study was a GWAS meta-analysis involving 37,382 individuals from 27 studies of European ancestry and 8,376 individuals from five Asian cohorts. The other study was a GWAS of 45,771 participants in a USA population of European ancestry. The findings are remarkable for genetic studies on myopia. These two GWASs not only replicated the previously associated loci but also identified several novel significant associations. The findings, if replicated from each other, would lead to less controversial results. There were 12 overlapping statistically significant associations with myopia and refractive error (p<5 × 10−8), and the candidate genes nearby were shown as follows: protease, serine, 56 (PRSS56, OMIM 613858); bone morphogenetic protein 3 (BMP3, OMIM 112263); potassium voltage-gated channel, KQT-like subfamily, member 5 (KCNQ5, OMIM 607357); laminin, alpha 2 (LAMA2, OMIM 156225); thymocyte selection-associated high mobility group box (TOX, OMIM 606863); tight junction protein 2 (TJP2, OMIM 607709); retinol dehydrogenase 5 (11-cis/9-cis; RDH5, OMIM 601617); zinc family member 2 (ZIC2, OMIM 603073); Ras protein-specific guanine nucleotide-releasing factor 1 (RASGRF1, OMIM 606600); GJD2; RNA binding protein, fo×-1 homolog (C. elegans) 1 (RBFOX1, OMIM 605104); and shisa family member 6 (SHISA6, Gene ID 388336, OMIM 617327). These genes are involved in neurotransmission, ion transport, retinoic acid metabolism, extracellular matrix remodeling, and eye development. They have common networks in the protein–protein interactions related to cell cycle and growth pathways, such as the well known molecular mechanisms of refractive error, which are the TGF-beta, mitogen-activated protein kinase (MAPK), and SMAD pathways. Overall, the findings of GWASs provide us with clues to discover more candidate genes and enrich our understanding of the pathogenesis of myopia.
Studies on the association of SNPs with myopia are generally based on refractive error after adjusting for age and gender. The association of SNPs with age at the onset of myopia or gender in myopia is not commonly seen but also provided us with new ideas to some extent. Association studies on SNPs with the age of myopia onset suggest that variants associated with myopia might have age-dependent genetic effects [27,95-98]. A study of the association between the age of myopia onset and 39 previously reported associated loci among 5,200 children ages 7–15 years provided evidence [95]. The results categorized the variants by ages into three groups: early-onset effect remaining stable, early-onset effect progressed with increasing age, and later-onset effect. The analysis of the genetic score for the 39 loci revealed a greater percentage of variants explained the phenotype at older age (2.3% at age 15 versus 0.6% at age 7), suggesting an increased genetic effect at older age. All of this evidence implies that the effects of the associated variants on the risk of myopia vary with age. This finding provided us with the new idea that association studies should be performed in a more specific sub-phenotype [95], and different biologic processes might contribute to the development of myopia at different ages. The gender association of SNPs with myopia was also observed in some studies [96-99]. For example, an association analysis of the GSTP1 Val/Val genotype with myopia identified this genotype as having a higher frequency in female probands [96], the CC genotype of TGF-beta 1 gene polymorphism (T869C) was statistically significantly associated with low myopia cases among male probands [97], and the DN genotype in endostatin gene polymorphism (D104N) increased the risk of myopia for women [98]. This evidence provided the information that gender is an important factor affecting the results and should be controlled in association analyses.
3.2.2 Identification of susceptibility loci for high myopia using association analysis-- GWASs and case-control studies have tested at least 27 loci associated with non-syndromic high myopia based on a refraction of SEM < −6.00 D (Appendix 2) [16,29,44-48,50,63,64,66,68,72,75,99-178]. Of these studies, eight loci were identified by GWASs and GWAS meta-analyses in Chinese, French, and Japanese populations and met the requirement of p<5 × 10−8 (Appendix 2). These loci were 2q22.3, the best SNP (rs13382811) near the ZFHX1B (Gene ID 9839, OMIM 605802) gene; 4q25 (MYP11), the best SNP (rs10034228); 7p36.3 (MYP4), the best SNP (rs2730260) near the VIPR2 (Gene ID 7434, OMIM 601970) gene; 8q24.12, the best SNP (rs4455882) near the SNTB1 (Gene ID 6641, OMIM 600026) gene; 8q23 (MYP10), the best SNP (rs17155227) near the MIR4660 (Gene ID 100616350) and PPP1R3B (Gene ID 79660, OMIM 610541) genes; 13q12.12 (MYP20), the best SNP (rs9318086) near the MIPEP (Gene ID 4285, OMIM 602241), C1QTNF9B-AS1 (Gene ID 542767, OMIM 617122), and C1QTNF9B (Gene ID 387911, OMIM 614148) genes; 15q14, the best SNP (rs11073058) near the GJD2 gene; and 22q13.31, the best SNP (rs10453441) near the WNT7B (Gene ID 7477, OMIM 601967) gene. Some were consistent with the findings in linkage studies, as seen in 7p36.3 (MYP4), 4q25 (MYP11), 13q12.12 (MYP20), and 8q23 (MYP10).
The SNP rs4455882 near SNTB1 has been reported to be statistically significantly associated with high myopia in a Chinese population and reached a best p value of 2.13 × 10−11 using GWAS meta-analysis (665 high myopia cases and 960 controls) followed by replications in three additional cohorts (a total of 2,128 cases and 3,683 controls) [154]. Another SNP, rs6469937 in SNTB1, was also found to be statistically significantly associated with myopia in another meta-analysis study [179]. SNTB1 is expressed in the human retina and sclera [179]. In mice with induced myopia, the expression of SNTB1 was downregulated in the retina/RPE compared with that in control mice [179]. This line of evidence indicates that SNTB1 might be the best candidate gene for high myopia as suggested by GWASs. Other loci were less replicated or replicated in controversial studies. None of the causative variants were detected in the myopia-associated genes.
3.2.3 Association studies of quantitative trait analysis-- Axial length and corneal curvature are key determinants of the refractive power in the eye. Association studies on these quantitative traits have tested 21 susceptible loci for ocular axial length and corneal curvature (Table 4) [3,69,79,80,141,180-184]. Of these loci, 16 were identified as statistically significantly associated with phenotype by GWASs, reaching p<5 × 10−8.
3.2.3.1 The role of axial length in association studies on myopia-- In a meta-analysis of three GWASs on axial length involving 1,118 cases and 5,433 controls from Chinese and Malay populations, SNP rs4373767 in ZC3H11B (Gene ID 643136) was found to be statistically significantly associated with axial length (p=2.69 × 10−10) and high myopia (p=4.38 × 10−7) [185]. ZC3H11B was also replicated to be associated with axial length and high myopia in another large meta-analysis with GWASs involving a total of 12,531 Europeans and 8,216 Asians [69]. Together with ZC3H11B, six other genes, RSPO1 (Gene ID 284654, OMIM 609595), C3orf26 (Gene ID 84319), LAMA2, GJD2, CD55 (Gene ID 1604, OMIM 125240), and ZNRF3 (Gene ID 81433, OMIM 612062), were found to be associated with axial length in this study [69]. The differential expression of these six genes has been observed between the induced myopic eyes and the control eyes in tissues, including the sclera, the RPE, and the neural retina. Among these genes, GJD2 and CD55 were previously reported to be associated with refractive error. GJD2 is a candidate gene in the 15q14 locus, which has been replicated in multiple studies. A statistically significant SNP near CD55 (rs1652333; p combined=3.05 × 10−12) was previously detected on a large scale in GWASs. These findings indicate that there is a close relationship between refractive error and axial length, which share about 50% of variants [186].
3.2.3.2 Association of variants with corneal curvature-- GWASs in corneal curvature have identified several loci near four genes, including PDGFRA (Gene ID 5156, OMIM 173490; 4p12), CMPK1 (Gene ID 51727, OMIM 191710; 1p32), RBP3 (Gene ID 5949, OMIM 180290; 10q11.2), and FRAP1 (Gene ID 2475, OMIM 601231; 1p36.2), that are statistically significantly associated with corneal curvature. FRAP1 and PDGFRA were first identified in a GWAS in an Asian population from Singapore and were replicated in two subsequent GWASs. Together with FRAP1 and PDGFRA, one of the subsequent GWAS discovered two other novel genes, CMPK1 and RBP3, associated with corneal curvature. These studies have opened a new path for investigating the molecular genetic basis of myopia, although the relationship between myopia and the genes associated with corneal curvature is unclear. Thus, it might be of interest for further study in the future.
3.2.4 Association studies on myopia: Principles and directions for the future-- More myopia loci have been identified by GWASs, but these loci account for a small proportion (less than 1%) of myopia in the general population. All of these significant associations tend to have a small effect due to their high allele frequencies. Thus, genotyping genomic regions near the significant locus in a larger cohort might detect variants with very rare frequencies and high penetrance, as seen in the discovery of highly penetrant variants in the CFH (Gene ID 3075, OMIM 134370) gene associated with age-related macular degeneration. Our previous study of 12 myopia-associated genes identified several rare variants in families with high myopia using whole exome sequencing [122]. Although the variant types and allele frequencies did not show statistically significant differences between the cases and controls, it opened up a new approach to analyze significant associations implicated by GWASs.
Most of the significant associations are located outside the protein-coding regions. Genes close to the significant regions are considered potential causative genes and chosen for further study. Even though no mutations in candidate genes have been identified, systemically reviewing genes near significant regions can enrich our understanding of pathways and gene–gene networks in the development of myopia.
In summary, all SNPs related to myopia, high myopia, axial length, and corneal curvature have been summarized according to their positions in the chromosome. SNPs identified by GWASs must be consistent with the principle of replication and reach the significant association level of p≤5 × 10−8. Recruiting larger cohorts or conducting an international collaboration would be helpful to improve statistical power and enable the discovery of additional significant associations.
3.3 Unraveling causative genes for high myopia with whole exome sequencing-- Whole exome sequencing has been widely used in studies on genetic defects for high myopia. With exome sequencing, ten genes in total have thus far been reported as associated with high myopia (Table 5), including four genes related to AD high myopia (ZNF644: Gene ID 84146, OMIM 614159, SLC39A5: Gene ID 283375, OMIM 608730, SCO2: Gene ID 9997, OMIM 604272, P4HA2: Gene ID 8974, OMIM 600608), three genes related to AR high myopia (LRPAP1: Gene ID 4043, OMIM 104225, LEPREL1: Gene ID 55214, OMIM 610341, CTSH: Gene ID 1512, OMIM 116820), two genes for XL high myopia (OPN1LW: Gene ID 5956, OMIM 300822 and ARR3: Gene ID 407, OMIM 301770), and one gene (BSG: Gene ID 682, OMIM 109480) inherited by de novo mutations [68,187-199].
Among these genes, mutations detected in genes with the AD trait were mostly missense, while those in genes with AR were mainly frameshift mutations. All of these genes are common in the expression spectrum in eye tissues, including the sclera, retina, and RPE; these genes also have a close relationship with the mechanism underlying the development of myopia. LRPAP1 and SCL39A5 were identified to be involved in the well known pathway TGF-beta/BMP. The mice that were homozygous deficient in LRPAP1 had reduced expression of LRP in the liver and brain [200], which, in turn, caused the activation of TGF-beta [201]. LRPAP1 was assumed to be associated with myopia through regulating the expression of TGF-beta. A loss of function in SLC39A5 was detected as involved in the expression of Smad1, a key phosphate protein downstream of the TGF-beta/BMPs. Other genes were involved with collagen synthesis (LERPEL1), ATP metabolism in the retina (SCO2), the transcription factor (ZNF644), and the degradation of lysosomal proteins (CTSH). The exact mechanisms of these genes remain unclear, and functional studies are anticipated in the future.
A recent analysis of whole exome sequencing data from 298 probands with early-onset high myopia discovered that one-fourth of the probands carried mutations in the genes responsible for retinal diseases and myopia-related syndromes [202]. This finding provided further clues for candidate gene screening using whole exome sequencing. Overall, in the era of whole exome sequencing, more genes are expected to be identified, which will provide more information on mechanisms of high myopia.
3.4 Gene analysis in the experimental animal models of myopia-- Animal models of myopia are used to identify the altered expression of genes in the retina, RPE, choroid, and sclera or to observe the features of myopia after knocking out the targeted gene. Methods for establishing myopia models include the form deprivation myopia (FDM), lens-induced myopia (LIM), and knockout animal models. The animal models are varied and can include chicks, mice, guinea pigs, rabbits, monkeys, or tree shrews, as well as the recently used zebrafish.
Experimental studies on myopia have provided convincing evidence for the role of the RPE, choroid, and sclera in the regulation of ocular growth and the development of myopia [203-212]. Animal models established by FDM and LIM have been widely used to investigate the candidate genetic factors underlying ocular growth and myopia. Studies on RPE have revealed that BMPs (typically BMP2: Gene ID 650, OMIM 112261, BMP4: Gene ID 652, OMIM 112262, BMP7: Gene ID 655, OMIM 112267) have bidirectional regulation during ocular growth [203,211]. For example, the expression of BMP2 was upregulated in a myopic defocus model (induced hyperopia) and downregulated in a hyperopic defocus model (induced myopia) [203]. BMPs encode secreted ligands of the TGF-beta superfamily of proteins involved in the TGF-beta/BMPs pathway. The expression of TGF-beta isoforms (TGF-beta 1, 2, and 3) in chick RPE was also found to be statistically significantly different in the LIM chick model [213]. All of this evidence indicates that TGF-beta superfamily members are involved in the regulation of ocular growth and suggests that the RPE plays a role as a signal relay. Studies on the choroid indicated its active role in conveying signals from the retina to the sclera [210]. Significant differences in gene expression in the choroid were observed between eyes with induced myopia and control eyes [205,206,210,213]. In addition, it is well accepted that the sclera plays a pivotal role in controlling ocular size. Sclera remodeling leads to decreased collage synthesis, scleral thinning, and loss of scleral tissue, all of which underlie the development of myopia [214].Thus, the alteration of scleral gene expression during the development and recovery of induced myopia in animal models has been widely investigated to explore the relationship between candidate genes and myopia [204,207,208,215-223]. Genes involved in scleral remodeling, such as TGF-beta isoforms, BMPs, cAMP: Gene ID 820, OMIM 600474, COL1A1: Gene ID 1277, OMIM 120150, TIMPs, and GAGs, were investigated mostly based on their functions as ocular growth factors, extracellular matrix (ECM) protein and enzymes, and their role in ECM and collagen synthesis. However, no specific mutations were detected among these genes associated with myopia in human beings. Most alterations in gene expression tend to be consistent during the development and recovery of induced myopia. This evidence not only indicates that scleral remodeling might involve the modulation of gene expression at the transcriptional level but also suggests that scleral remodeling is too complex to be explained by a single gene.
The genes identified in experimental myopia studies have features in common. First, they have expression in the sclera, retina, or choroid. Second, increased or decreased changes in expression can be observed in induced myopia models. Third, knockout animal models present with some related features of myopia, such as a thinner sclera, an elongated axial length, and the decreased synthesis of collagen fibers. Importantly, most of these genes are relevant to the signaling pathways possibly underlying myopia (Table 6) [121,203-207,211,217,224-246], although most have yet to be determined. The potential signaling pathways are TGF-beta signaling, scleral remodeling, Wnt signaling, Pax6 in early embryonic growth, Sonic hedgehog, retinoic acid, nitric oxide synthase, neurotransmitter, muscarinic receptor signaling, and retinal dopamine signaling. Among them, the well recognized signaling pathway is the TGF-beta/BMPs pathway, which has been implicated independently in different studies. The TGF-beta/BMPs pathway is considered to have a potential role in controlling ocular growth. Decreased expression of TGF-beta isoforms (the TGF-beta 1 to 3 isoforms) in the sclera of the FDM animal models was intended to reduce the synthesis of collagen, thus making a contribution to the sclera remodeling in the myopic eye. Additional candidate genes related to the TGF-beta/BMPs pathway will be of interest for future investigation.
In addition, animal model studies have provided further evidence to support the findings of association studies, whole exome sequencing, and linkage studies. In human beings, P3H2 is encoded by LERPEL1, in which mutations have been reported to be associated with autosomal recessive high myopia using whole exome sequencing [192]. P3h2 null mice were further observed with altered collagen prolyl 3-hydroxylation [235], which is believed to cause abnormalities in the structure of the sclera, thus resulting in progressive myopia. A GWAS identified ZFHX1B as a susceptible locus for severe myopia [121]. The decreased expression of Zfhx1b was observed from the induced experimental mouse model for myopia, providing further evidence of the involvement of this locus in the development of myopia [121]. However, we still have little knowledge about the exact relationship between induced myopia in animals and physiologic myopia in humans, and the exact variants in most of the genes suggested by animal studies are still unclear.
Myopia, usually high myopia, has been reported as a feature in a variety of ocular and systemic syndromes. These myopic syndromes are often shown in Mendelian inheritance patterns, and the genetic defects underlying these syndromes have been well clarified with certain protein-coding genes. A total of 382 entries were obtained using “myopia” as a search word (until 3/18/2017) in OMIM. Entries associated with syndromic myopia were included and reviewed. A total of 115 genes were identified as the causes of the 131 syndromes accompanied by myopia (Appendix 3) [116,164,192,247-416]. These syndromes shown in Appendix 3 are congenital and inherited by defects in certain genes in a pattern of autosomal recessive, autosomal dominant, or X-linked inheritance.
Myopic syndromes have variable clinical manifestations, ranging from mild to severe, that affect patients’ daily lives in various aspects. Some even lead to death after birth or in childhood. Myopia or high myopia is one of these variable features, and other clinical features involve many different systems, especially other segments of the ocular system, nervous system, or locomotor system. For example, some ocular abnormalities include astigmatism, nystagmus, night blindness, strabismus, microphthalmia, microcornea, brittle cornea, keratoconus, cataract, ectopia lentis, posterior staphyloma, glaucoma, vitreoretinopathy, retinitis pigmentosa, and retinal detachment. In association with the nervous system, mental retardation, microcephaly, cerebellar hypoplasia, special facial appearance, and hearing loss are often seen. In ophthalmology, myopic syndromes, such as Stickler syndrome, Marfan syndrome, congenital stable night blindness (CSNB), retinitis pigmentosa (RP) associated with RP2 (Gene ID 6102, OMIM 312600) and RPGR (Gene ID 6103, OMIM 312610), Bornholm eye disease (BED), and cone-rod dystrophy (CORD), have been well established, and high myopia is their common feature. Others involving primarily non-ocular systems are rare and have not been thoroughly investigated.
Systematic reviews of myopia-related syndromes have great implications in diagnosis, prevention, and treatment. Two studies [417,418] revealed that most patients diagnosed with simple myopia showed other ocular or systemic symptoms after full assessment, especially children under 10 years of age. Stickler syndrome, Marfan syndrome, and retinal dystrophy were predominantly seen. If early diagnosis and early prevention were made available for patients with conditions such as Stickler syndrome, it would reduce their chances of developing retinal detachment using prophylactic cryotherapy. Practically, high myopia might be an earlier feature in some syndromes or a major complaint in a patient's first visit. In routine clinical practice, syndromic high myopia was generally considered to be rare and might be simply cosidered as high myopia alone due to unawareness or atypical manifestation of other major signs. This situation may not well recorgnized until recently. Sun et al. performed a systematic analysis of the genes associated with retinal dystrophy and myopia-related syndromes in 298 probands with early-onset high myopia (eoHM) [202]. Mutations were identified in one-fourth of the probands, of whom 62% have mutations in the genes that contribute to myopia-related syndromes. Mutations in genes COL2A1 (Gene ID 1280, OMIM 120140) and COL11A1 (Gene ID 1301, OMIM 120280) associated with Stickler syndrome, CACNA1F (Gene ID 778, OMIM 300110) associated with CSNB, and RPGR associated with XL-RP were predominantly discovered among these probands. Follow-up examination on some of the eoHM with the mutations revealed that: 1) some of the eoHM are still hardly recorgnized as related syndromes due to atypical or variable manifestation, and 2) a small portion of those eoHM should be syndromic high myopia if systemic examination of the eyes as well as other related organs could be thoroughly performed. The investigation of the genes associated with syndromic high myopia has also become an additional approach for the candidate gene screening of isolated high myopia based on a common molecular genetic function, such as involving the ECM and connective tissues. Mutations in NYX (Gene ID 60506, OMIM 300278) have been reported to cause high myopia without congenital stationary night blindness (CSNB1) [419]. A recent study conducted with seven families with BED, two of whom were previously mapped to the MYP1 locus, discovered two rare haplotypes (LIAVA and LVAVA) underlying BED [334]. Li et al. later found that one of the haplotypes (LVAVA) also contributes to two large families with X-linked non-syndromic high myopia [191]. These observations suggest that there is a close relationship between the genetic basis for syndromic and non-syndromic high myopia, and thus, it is a good approach for investigating the genes implicated in studies of syndromic high myopia.
This review provides a comprehensive overview of studies on the molecular genetics of myopia. For the reported genetics factors associated with myopia listed in this review, some might be very preliminary while a few may already have contradictory evidence. Therefore, many of the myopia-related genes or loci need to be further validated. In addition, for most loci or associated SNPs, the exact genetic defects and their related functional consequence that contributes to myopia await disclosure. No specific pathways underlying myopia have been clarified that would impede our prioritization of candidate genes. Therefore, in conjunction with functional studies and a combination of different technologies will be helpful for enhancing our understanding of the genetic factors in myopia, ultimately leading to improvements in the prediction of onset, prevention, and treatment in the future.
Appendix 1. Loci or genes tested in associated with complex myopia
Appendix 2. Loci or genes tested in association with nonsyndromic high myopia
Appendix 3. Reported genes associated with syndromic high myopia.
The authors acknowledge the grant support from the Natural Science Foundation of Guangdong (S2013030012978), the National Natural Science Foundation of China (81770965), the Science and Technology Planning Projects of Guangzhou (201607020013), and the Fundamental Research Funds of the State Key Laboratory of Ophthalmology