To date, 22 loci of glaucoma (Table I) have been identified and designated as GLC1A-Q and GLC3A-E. POAG is linked to 17 loci; GLC1A, 1C, 1E-H, 1O-P, for which the responsible genes are MYOC, interleukin 20 receptor subunit β (IL20RB), OPTN, ASB10, WDR36, EGF containing fibulin-like extracellular matrix protein 1 (EFEMP1), NTF4 and TBK1, respectively; and GLC1B, 1D, 1I-N, and 1Q, for which the responsible genes remain unidentified. There are five loci linking to primary congenital glaucoma (PCG), GLC3A and 3D-E, for which the responsible genes are CYP1B1, latent transforming growth factor-β-binding-protein 2 (LTBP2) and TEK tyrosine kinase endothelial (TEK), respectively; the responsible genes of GLC3B-C remain unidentified. GLC1A-Q, except GLC1A, 1J, 1K, 1M and 1N, which contribute only to juvenile open angle glaucoma (JOAG), contribute to adult-onset POAG. All of GLC3A-E have been implicated in PCG. Glaucoma-causing mutations may be classified into two groups. One is autosomal dominant, including POAG-causing genes (MYOC, IL20RB, OPTN, EFEMP1 and TBK1) and a PCG-causing gene (TEK). The other is autosomal recessive, including a PCG-causing gene (CYP1B1). Of the 22 loci, GLC1A (MYOC) and GLC3A (CYP1B1) are the most important for glaucoma; they have correspondingly been the most investigated in research.

Only four pathogenic genes, MYOC (64,65), NTF4 (65,70), OPTN (65,77) and WDR36 (65,78), have been definitively linked to POAG. Furthermore, it was reported that mutations in OPTN, MYOC or WDR36 account for ~4% of all glaucoma (79). The link between ASB10, IL20RB and EFEMP1, and POAG, is less certain. TBK1 is controversial, since GLC1P covers three other genes, n-acetylglucosamine-6-sulfatase, ras association domain family protein 3 and exportin-t (80); however, TBK1 has been suggested to be the most possible glaucoma-causing gene for GLC1P (80).

Only one pathogenic gene for PCG, CYP1B1 (6), has been clearly identified in the locus GLC3A. Numerous genes have been observed in 1p36 that contain GLC3B, however, none have been demonstrated to be associated with PCG (6). To date, it remains to be investigated whether LTBP2 is associated with the GLC3C or GLC3D loci. LTBP2 is ~1.3 Mb proximal to GLC3C (82), thus there is a hypothesis that LTBP2 may be the GLC3C gene; however, the possibility that it may be an adjacent gene associated with PCG may not be ruled out. Another study suggested that GLC3D is distal to GLC3C without overlapping (83). Furthermore, there is evidence that LTBP2 is a candidate for GLC3D (82); therefore, in the present review LTBP2 is presented as the GLC3D gene. There is strong evidence (OMIM) that mutations of TEK may result in GLC3E, and the locus of TEK is GLC3E.

To the best of the authors' knowledge, besides the 22 loci of glaucoma mentioned, there are 74 genes that are more closely associated with glaucoma presented in Table II. Of those 74, 48 (64%) are associated with POAG, followed by PACG (16%), PCG (4%) and pseudoexfoliation glaucoma (PEXG; 4%). Toll-like receptor 4, sine oculis homeobox Drosophila homolog of 1, doublecortin-like kinase 1, RE repeats-encoding gene, retinitis pigmentosa GTPase regulator-interacting protein, lysyl oxidase-like protein 1 (LOXL1), heat-shock 70-kD protein 1A (HSP70-1), baculoviral IAP repeat-containing protein 6, 5,10-methylenetetrahydrofolate reductase (MTHFR) and nitric oxide synthase 3 (ENOS) are human genes involved in more than one phenotype of glaucoma. Nanophthalmos 1 is identified to be the only human gene known to cause PACG (140). For other genes (ATP-binding cassette subfamily C member 5, SPARC-related modular calcium-binding protein 2, matrix metalloproteinase 9, membrane-type frizzled-related protein, hepatocyte growth factor, HSP70-1, pleckstrin homology domain-containing protein family A member 7, collagen type XI α-1, MTHFR and ENOS) identified to be associated with PACG in Table II, it remains unclear whether they are pathogenic genes for PACG; however, they may be a risk factor for the development of PACG. Among genes associated with PEXG and PEX, the majority of research has been conducted on LOXL1 to determine whether it is pathogenic and how it contributes to the two diseases. PEX, characterized by the accumulation of protein fibers in the eyes, may have a genetic basis. The accumulation of protein obstructs aqueous humor (AH) outflow, and that results in PEXG. Previously, two studies (111,141) have confirmed that LOXL1 is significantly associated with PEXG and PEX. A decrease in LOXL1 expressionmay cause degenerative tissue alterations in LC, and consequently results in patients with PEX being more vulnerable to optic nerve damage caused by pressure (141), a risk factor for PEXG development.

The pathogenesis of MYOC mutations is unclear; however, the three most possible causes for glaucoma are as follows. In the unhealthy state, there is poor normal MYOC protein secretion. MYOC mutations may lead to accumulation of mutated MYOC proteins within the TM (151–154). Retention of abnormal MYOC protein may be harmful to TM cells and result in their dysfunction or death, which may obstruct AH outflow, and consequently increase IOP (43,64,147,155,156). In addition, accumulation of mutated MYOC proteins in the endoplasmic reticulum activates the unfolded protein response (UPR) in TM cells (157), subsequently leading to apoptosis that may cause high IOP. Over activated UPR may lead to certain neurodegenerative diseases, and inhibiting UPR is a possible therapy for these diseases (158). Thus, this method may additionally be applicable to glaucoma. Normal MYOC is involved in exosome shedding into the aqueous humor, and exosomes are associated with paracrine and autocrine signaling (74), that therefore may serve as vehicles of MYOC protein trafficking. Notably, normal MYOC protein is absent in the aqueous humor of glaucomatous patients with pathogenic MYOC mutations (151). Thus, another prevalent hypothesis on the pathogenesis of MYOC mutations is that they interfere with MYOC protein trafficking and lead to the intracellular aggregation of the misfolded MYOC protein (74). Accumulation of misfolded MYOC proteins decreases AH out flow, and that influences IOP regulation; however, its mechanism is unclear (74). The third hypothesis is regarding specific interactions between MYOC mutations and mitochondria in the TM (159,160). A subsequent study indicated that MYOC mutations lead to dysregulation of calcium channels resulting in mitochondrial depolarization in the TM, consequently resulting in TM contraction, which decreases AH outflow and further causes increased IOP (161).

There is strong evidence that only ~5% of POAG (65) is caused by a single gene, and other cases of POAG are caused by digenic or polygenic cooperation mechanisms, none of which may alone cause glaucoma. Usually, cooperation of MYOC mutations with one or more genes contributes to glaucoma. Mutations in MYOC, OPTN and CYP1B1 are identified to coexist in ~3.59% of POAG cases (162). This demonstrates that mutations in the three genes together may be involved in the pathogenesis of POAG. There are other studies investigating the association between MYOC and OPTN. OPTN and MYOC are observed in POAG (69,162–166), exfoliative glaucoma (164) and exfoliation syndrome (164). Overexpression of OPTN may upregulate MYOC in TM and stabilize MYOC mRNA (167). There is a possible polygenic interaction among MYOC, OPTN and apolipoprotein E (APOE). Disease-causing mutations in MYOC and OPTN contribute to only a small number of Chinese POAG cases (163). However, common polymorphisms in MYOC, OPTN (69,163,166), APOE (69,163) and WDR36 (69) may together account for POAG. Common polymorphisms of these genes are not associated with POAG alone; however, they may cooperatively contribute to the disease, which indicates a polygenic pathogenesis. A study reported that the mean onset age of carriers with only MYOC mutations is 51 years; however, the mean onset age of carriers with MYOC and CYP1B1 mutations is 27 years (168). This indicates that mutations in the two genes may interact to advance the onset age of glaucoma. Notably, in a study (169) Gln48His, a MYOC mutation, was observed in POAG and PCG; however, one patient with PCG had a CYP1B1 mutation (Arg368His), and the other patient with PCG had none of the CYP1B1 mutations. These results demonstrate that there is a possible digenic interaction between MYOC and CYP1B1, without excluding the possibility that there has been an unidentified gene associated with glaucoma. However, another study reported that none of the CYP1B1 mutations was observed in all five POAG cases with MYOC mutations (170). Forkhead box C1 (FOXC1) may regulate MYOC secretion through modulation of RAB3 GTPase-activating protein catalytic subunit 1 RAB, synaptosomal-associated protein 25-kD and RAB3 GTPase-activating protein noncatalytic subunit (144). A different study (171) suggested that MYOC and FOXC1 mutations are not associated with the pathogenesis of PCG. The mutations Leu486Phe in MYOC and Val108Ile in UDP-Gal: β GlcNAc β-1,4-galactosyltransferase polypeptide 3 may cooperatively contribute to the pathogenesis of POAG (94).

OPTN, widely expressed in retinal ganglion cells (172), the nonpigmented ciliary epithelium, human TM and the retina (77), is an autophagy receptor. Autophagy may remove damaged organelles and proteins via lysosomal degradation (172). Autophagy and membrane vesicle trafficking serve an important role in the regulation of OPTN functions (172). Furthermore, the level of autophagy mediated through OPTN is very important for the survival of retinal cells (172). Mutations in OPTN are involved in POAG (172). Another conclusion contradicted this result, reporting that OPTN is not associated with POAG in Spain (173). Of these disease-causing mutations, two are noteworthy, Glu50Lys and Met98Lys.

The frequency of Glu50Lys in POAG is 13.5% (77). Notably, 81.6% of POAG cases with recurrent Glu50Lys have normal IOP; whereas only 18.4% of those have increased IOP (77). However, another study suggested that OPTN mutations are involved in POAG rather than glaucoma with normal IOP in Japanese patients (174). Glu50Lys impairs autophagy (172) and trafficking (172,175), resulting in the death of retinal cells through apoptosis (172) and disrupting the endocytic recycling that is very important for maintaining homeostasis (175).

Rezaie et al (77) first reported that Met98Lys is a risk-causing mutation for POAG, and the frequency of Met98Lys in POAG (13.6%) is significantly higher compared with controls (2.1%). In another study by Sripriya et al (176), Met98Lys was not identified in controls; however, it was identified in POAG (4.1%) and NTG (6%) (162). Mukhopadhyay et al (177) did not detect Met98Lys in NTG, and the frequency of Met98Lys was 11% in POAG and 5.5% in controls. An alternative study (162) presented the contrary conclusion that Met98Lys may not be a risk-causing factor for POAG on account of a very similar frequency in POAG (7.97%) and controls (7.29%). Met98Lys is usually known as a disease-causing mutation and the majority of POAG cases with Met98Lys additionally have normal IOP, similar to Glu50Lys (77). The possible pathogenesis of Met98Lys is that it may impair autophagy, which consequently leads to the death of retinal cells through apoptosis and transferrin receptor degradation (172). However, the pathogenic mechanism of glaucoma-causing Met98Lys requires further research and examination.

WDR36, located within the POAG linkage locus GLC1G and first identified by Monemi et al (78), is widely expressed in numerous ocular tissues, including the optic nerve, ciliary body, retina, TM, ciliary muscles, iris, lens and sclera. Monemi et al (78) formerly suggested that the frequency of WDR36 mutations in POAG is 1.6–17%. There is a possible association of WDR36 with the pathogenic mechanism of HTG (67,68). WDR36 mutations may alter the cell phenotype supporting the theory that WDR36 is associated with the polygenic pathogenesis of glaucoma (178). To date, WDR36 importance remains unclear; furthermore, its pathogenicity is controversial. As subsequent studies did not demonstrate WDR36 mutations to be POAG-causing mutations, it was demonstrated that WDR36 mutations may only be a risk factor for POAG (67–69). In addition, Fingert et al (179) were not able to confirm the association of WDR36 with pathogenesis of POAG.

NTF4, located within the POAG linkage locus GLC1O, is localized to RGCs (70). Pasutto et al (70) suggested that the frequency of NTF4 mutations in POAG is 1.7% and there is strong genetic evidence that NTF4 mutations are involved in POAG of European origin. Liu et al (180) additionally identified coding alterations in five POAG cases and 12 controls of European origin from Southeastern USA, of which two mutations were previously detected by Pasutto et al (70). Therefore, Liu et al concluded that these NTF4 coding alterations are not significantly associated with the pathogenesis of POAG. In addition, another study by Chen et al (181) suggested that NTF4 does not mainly contribute to the molecular genetics of POAG. From the above, the association of NTF4 mutations with POAG pathogenesis remains to be investigated. In addition, besides the European origin, NTF4 mutations have been identified in Chinese populations (182). This indicates that NTF4 mutations may derive from multiple ancestors.

ASB10, located within the POAG linkage locus GLC1F (183), influences AH outflow (71). ASB10 is most highly expressed in the iris, followed by human TM, RGCs, the ciliary body, choroid, optic nerve, retina, lamina and a little in the lens (71). Among patients with POAG and controls, the frequency of ASB10 mutations is 6 and 2.8%, respectively (71). To test whether ASB10 influences AH drainage, Pasutto et al (71) applied RNA interference silencing for knockdown of ASB10 mRNA expression in perfused human anterior segment cultures. The results revealed that the decrease in AH outflow facility was ~50%. In addition, ASB10 may be involved in the ubiquitin-mediated degradation pathways through interactions of ASB10 with the α4 subunit of the 20s proteasome and with HSP70 in TM (184).

EFEMP1, located within the POAG linkage locus GLC1H, is a plausible candidate for POAG (185). Although there have been a few efforts to confirm the linkage to GLC1H, it remains uncertain. Mutations in EFEMP1 are involved in decreasing the optic disc area (186). Another mutation, c.418C>T in EFEMP1 may be predictive for POAG (185). Expression of EFEMP1 may be influenced by transforming growth factor (TGF)-β2. A study by Junglas et al (187) reported that TGF-β2 is more highly expressed in AH of POAG and maybe associated with the increase in AH outflow resistance in POAG. Higher amounts of TGF-β2 inhibit the expression of EFEMP1 (188).

IL20RB, located within the POAG linkage locus GLC1C, has a role in POAG pathogenesis (189). An IL20RB mutation, Thr104 Met, lying in an active binding site of IL20RB (190), has been observed in a large POAG family (189); therefore, this additionally demonstrates that IL20RB may be implicated in the pathogenesis of POAG. According to OMIM (no. 605621), IL20RB is highly expressed in human skin and testes, and less expressed in the muscle, placenta, heart, ovary and lung. Recently, IL20RB was detected to be additionally expressed in human TM (191). To the best of the authors' knowledge, thus far, little research effort has been made to investigate IL20RB as a POAG-causing gene.

In humans, the CYP1B1 gene encodes cytochrome P450 1B1, and is regulated via the aryl hydrocarbon receptor. CYP1B1 was the first gene identified in PCG-associated loci (GLC3A-3E) (192), and its role has been clearly understood (65). CYP1B1 is widely expressed in the eyes, including the retina, iris, ciliary body and TM (193). However, certain previous studies suggested that CYP1B1 is not expressed in TM at any stage of eye development (194). CYP1B1 has been thought to be significantly associated with human fetal eye development (194). To date, at least 147 CYP1B1 mutations have been identified globally in 542 patients with PCG in various countries, including Brazil, China, India, Iran, Morocco, Russia, Saudi Arabia, Slovak Gypsy populations, Turkey, USA, Spain, Pakistan, Oman, the Netherlands, Mexico, Kuwait, Japan, Israel, Indonesia, Germany, Ecuador, Canada, Britain and Algeria (6,195). Among CYP1B1 mutations, Glu387Lys has been traced to a common genetic origin for PCG (196). CYP1B1 mutations, which have been reportedly associated with a wider range of glaucomatous phenotypes, including PCG (6,169,195–197), POAG (198–200), JOAG (201) and PEXG (199), appear in patients with glaucoma at a higher frequency compared with other glaucoma-associated genes (199). CYP1B1 mutations may confer increased susceptibility to PCG and are the most common pathogenic factors of PCG (6). However, the frequency of PCG-causing mutations in CYP1B1 varies significantly in different populations, including Mexican (<10%) (202), Vietnamese (16.7%) (197), Chinese, Japanese and Indonesian (all 20%) (202), Indian (40%) (169). Furthermore, PCG-causing mutations in CYP1B1 occur with extremely high incidence in Slovak Gypsy and Saudi Arabian populations (202), which supports an additional study reporting that consanguinity is a fundamental mechanism for high PCG incidence in Slovak Gypsy and Saudi Arabian populations (203) Available data demonstrate that CYP1B1 may not be the primary disease-causing gene for glaucoma in East Asians and South East Asians, unlike in Gypsy and Saudi Arabian populations. Furthermore, PCG in Mexicans may not be caused by CYP1B1 mutations. In addition, only ~10% of cases of POAG in Mexico harbor CYP1B1 mutations (198), demonstrating that CYP1B1 mutations may not be the cause of the pathogenesis of POAG; however, dysfunction of CYP1B1 may increase the risk of POAG. A low percentage of JOAG cases (~5%) harbor CYP1B1 mutations (168), and CYP1B1 possibly contributes to JOAG in a monogenic model (201).

An increasing amount of research attention is focusing on the interactions of CYP1B1 with other genes. There is growing evidence that interactions of CYP1B1 with MYOC occur in patients with PCG (169,204). In the process of interactions, MYOC is a potential modifier gene (205). In addition, TEK mutations co-occur with CYP1B1 mutations in patients with PCG; notably, the parents of these patients with PCG harbor either heterozygous CYP1B1 or TEK alleles and are asymptomatic (206). Furthermore, there is strong evidence suggesting that the interaction between CYP1B1 and TEK accounts for the pathogenesis of PCG (206); however, the mode of interaction remains unclear regarding whether an overlapping or independent mode is involved in the pathogenic mechanism of PCG. The interaction of CYP1B1 with MYOC and TEK, respectively, in the pathogenesis of PCG further lends support to the digenic inheritance of PCG.

Although CYP1B1 mutations are the most common cause of PCG, these mutations only contribute to a very small proportion of the total amount of PCG (6). Besides CYP1B1, there are a number of genes demonstrated to be associated with PCG, including LTBP2, FOXC1 and MYOC. Therefore, it is reasonable to speculate that other genes may participate in the pathogenesis of PCG; however, there still remain a large number of unknown genes requiring identification.

LTBP2, located within the PCG linkage locus GLC3D, is the largest member of the latent TGF-β family whose signaling failure in the anterior and posterior eye may cause pathogenic alterations in POAG (84). LTBP2 is most highly expressed in the lens capsule (192), secondly in the TM and ciliary processes (82,192) that are thought to be associated with PCG pathogenesis, with a very small amount in the sclera, corneal stroma and iris (192). LTBP2 mutations are identified in different populations, including Pakistani (82), Indian (207), Gypsy (82), Iranian, Moroccan, and Saudi Arabian populations (OMIM). From the aforementioned data, LTBP2 mutations appear to derive from West Asia and South Asia. Although Morocco is located in Africa, 75% of Moroccans are of Arabic descent; furthermore, the origin of the Gypsy ethnicity is thought to be in Ancient India. To the best of our knowledge, LTBP2 mutations have been not observed in other populations. Therefore, it is reasonable to hypothesize that LTBP2 mutations may have the same ancestor.

LTBP2 is a disease-causing gene for PCG (192) and is very important in the development of the anterior chamber of the human eye, where LTBP2 possibly serves a role in maintaining ciliary muscle tone (82). Besides PCG, LTBP2 mutations maybe associated with PACG and POAG (208). Therefore, there may be an overlap in the pathogenic mechanism among various types of glaucoma. It is this overlap that may account for the common characteristics among these various types of glaucoma, including optic nerve impairment and decreased vision, and for the common clinical presentation at onset, including eye pain, red-eye, blurred vision, nausea and mid-dilated pupils. However, another study (207) had contrary conclusions that LTBP2 mutations are not implicated in the pathogenesis of PCG. In addition, LTBP2 is not thought to be a disease-causing gene for PCG in the Han Chinese population (209).

TEK, located within the PCG linkage locus GLC3E, is an angiopoietin receptor, additionally termed cluster of differentiation 202B and tyrosine kinase with immunoglobulin-like and EGF-like domains 2, and may regulate vascular homeostasis (210). Although TEK and other vascular growth factors are important for AH outflow and Schlemm's canal development, their association with glaucoma remains unclear (211). A 50% decrease in TEK adequately demonstrated defective Schlemm's canal development and impaired AH outflow (210), and this demonstrates that TEK concentration is important for the AH drainage pathways. Variable expression of TEK is possibly produced by oligogenic or digenic inheritance, in line with other ocular disorders of developmental origin produced by mutations in optic atrophy 1, FOXC1, paired box gene 6 and MYOC (210). In addition, another recent study demonstrated that TEK mutations co-occur with CYP1B1 mutations in PCG (206), and demonstrated that interactions between TEK and CYP1B1 account for digenic inheritance in PCG pathogenesis.

Among the 96 genes, mutations of MYOC (GLC1A) and CYP1B1 (GLC3A) have the closest associations with potential pathological mechanisms in glaucoma. Besides the aforementioned glaucomatous pathogenesis, a novel pathogenic mechanism for MYOC-associated glaucoma is proposed. Extracellular matrix (ECM) proteins of TM are synthesized in the endoplasmic reticulum (ER) and finally secreted into the ECM. Malfunction of the ER during ER stress caused by mutant myocilin accumulation in the ER may affect ECM protein processing and secretion, which results in aberrant intracellular accumulation of ECM proteins in TM (212). The accumulation of ECM proteins may deteriorate ER stress, leading to TM cell dysfunction and obstructing AH outflow, thereby increasing IOP (212).

CYP1B1 defects cause angle abnormalities involving TM and Schlemm's canal (213). CYP1B1 mutations lower activity or stability of the enzyme in the mitochondria (214,215) and reduce expression levels of ECM proteins in TM (215). These may impact the development or filtering function of TM. In addition, abnormal mitochondria caused by CYP1B1 mutations [the same case as with MYOC mutations (161)] may cause dysregulation of calcium channels resulting in mitochondrial depolarization in TM, consequent TM contraction, reduction of AH outflow and an increase in IOP.

Based on previous studies on the potential pathogenic mechanism of MYOC mutation, at present, there have been three novel approaches to treatment for MYOC-associated POAG: i) Using chemical chaperones (based on molecular mechanisms) to decrease misfolding or unfolding of proteins and increase MYOC secretion (216); ii) given the gain-of-function nature of MYOC mutations, another novel approach is targeting MYOC mRNA or the myocilin protein (216); and iii) targeting MYOC by gene editing with clustered regularly interspaced short palindromic repeats-Cas9 technology to reduce ER stress and lower IOP (216).

According to previous studies on potential pathogenic mechanisms of CYP1B1 mutation, researchers developed two novel therapies: One is the approach based on the gene, directly attempting to correct or replace abnormal CYP1B1 (217); the more novel approach differentiates into a specific lineage and transfers stem cells containing wild-type CYP1B1 to stimulate the normal development of TM cells (217).