Genetic Architecture and Clinical Phenotypes of Primary Open-Angle Glaucoma: An Updated Review of Mendelian Genes, GWAS Loci, and Polygenic Risk

Abstract

Purpose

To synthesize the genetic architecture of primary open-angle glaucoma (POAG) across the full variant spectrum and map each genetic subtype to its clinical phenotype, IOP profile, optic nerve morphology, systemic associations, and treatment response.

Methods

PubMed/MEDLINE was searched from inception through April 2025 for POAG genetics, susceptibility genes, GWAS loci, polygenic risk scores, and gene-specific clinical correlations. Eligible designs included peer-reviewed original research, GWAS meta-analyses, candidate-gene studies, family-based linkage analyses, cohort and case-control studies, and PRS development in human subjects. Genetic association studies were appraised using Q-Genie; observational clinical studies were appraised using the Newcastle-Ottawa Scale.

Results

Seven high-penetrance Mendelian genes (MYOC, OPTN, TBK1, WDR36, NTF4, FOXC1, CYP1B1) account for fewer than 10% of cases, each defining a clinically distinct subtype: juvenile high-tension glaucoma (MYOC), normal-tension neurodegeneration (OPTN, TBK1), and syndromic developmental glaucoma (FOXC1, CYP1B1). GWAS meta-analyses have identified 100 to 127 susceptibility loci converging on five biological axes: trabecular meshwork dysfunction, retinal ganglion cell survival, autophagy and protein homeostasis, neuroinflammation, and extracellular matrix integrity. Polygenic risk scores combined with clinical variables achieve an AUC of 0.89.

Conclusion

POAG comprises genetically and clinically distinct subtypes defined by dominant molecular axes. Gene-specific data reveal actionable pharmacogenomic signals: ROCK inhibitors for ARHGEF12 and AFAP1 carriers, nicotinamide for OPTN and TXNRD2 subtypes, and goniotomy for CYP1B1-associated congenital glaucoma. IOP-centric management is insufficient for normal-tension subtypes driven by autophagy failure and neuroinflammation. The framework supports genotype-guided decisions on screening intensity, IOP targets, and neuroprotective adjuncts.

Plain Language Summary

Glaucoma is the world’s leading cause of permanent blindness, and the most common type, primary open-angle glaucoma (POAG), runs strongly in families, suggesting an important genetic basis. This review brings together current knowledge on the genes and genetic variants associated with POAG, and explains what each tells us about who gets the disease, how it behaves, and which treatments may work best. We describe two broad categories of genetic risk: rare mutations in a small number of genes (such as MYOC, OPTN, and TBK1) that almost always cause glaucoma and define distinct clinical subtypes, and hundreds of common low-risk variants that, in combination, determine most of the inherited risk in the general population. Combining these into a polygenic risk score now allows clinicians to identify high-risk individuals before symptoms appear. Importantly, different genetic subtypes respond differently to treatment: some benefit most from lowering eye pressure, while others need neuroprotective strategies independent of pressure. This review provides a practical framework for matching genetic subtype to clinical management, supporting a move toward precision medicine in glaucoma care.

Introduction

Glaucoma is the leading cause of irreversible blindness worldwide, with primary open-angle glaucoma (POAG) the most prevalent subtype. POAG is defined by progressive optic neuropathy, visual field loss, an open anterior chamber angle, and the absence of a secondary cause.¹ At the cellular level, POAG is characterized by selective death of retinal ganglion cells (RGCs) and degeneration of their axons.² Because this damage is irreversible, pre-symptomatic identification of at-risk individuals is the central opportunity to prevent blindness Figure 1. A PRISMA flowchart tracks study filtering from 2,874 database records plus 97 hand-searched items, leaving 2,683 after deduplication and 642 for full-text review after 2,041 title/abstract exclusions Figure 2.

Figure 1.

PRISMA flow diagram of literature screening and study selection for Arm A (POAG genetics, IOP endophenotypes, and polygenic risk scores).

Figure 2.

PRISMA flow diagram of literature screening and study selection for Arm B (gene-stratified clinical studies of MYOC, OPTN/TBK1, CYP1B1/FOXC1, WDR36/NTF4, common GWAS loci, and PRS validation).

Approximately 80 million people were living with POAG as of 2020, projected to reach 112 million by 2040, with African and Hispanic populations bearing the highest risk of bilateral blindness.^3,4 Heritability estimates for IOP and optic disc morphology reach 70% in twin cohorts, and a positive family history is reported in 4 to 16% of cases.^5,6 Environmental factors contribute but do not explain this pattern; over 90% of the genetic architecture of POAG remains unaccounted for.⁷

Genetic contributors fall into two classes. The first comprises rare, high-penetrance mutations in MYOC (myocilin), OPTN (optineurin), TBK1, and related genes, causing early-onset or familial glaucoma through protein misfolding and ER stress, disrupted autophagy, and altered NF-kappaB signaling.⁸ MYOC mutations alone account for approximately 3.5% of POAG cases.⁹ Animal models overexpressing mutant myocilin consistently fail to replicate the human phenotype, raising unresolved questions about model validity and disease mechanism.¹⁰

The second class consists of common low-penetrance variants identified through genome-wide association studies (GWAS). Meta-analyses across multiple ancestries have implicated 100 to 127 loci in POAG susceptibility, spanning extracellular matrix remodeling (LOXL1, AFAP1), cell cycle control (CDKN2B-AS1), vascular biology (CAV1/CAV2), calcium homeostasis (TMCO1), and neuronal survival (SIX6, ATOH7, NTF4).¹¹ In aggregate these variants define a polygenic architecture accounting for the majority of heritable POAG liability.^11,12

These two tiers are not independent. Digenic and epistatic interactions between Mendelian genes and common alleles complicate both mechanistic interpretation and clinical risk stratification.¹³ Whether IOP-dependent or IOP-independent pathways primarily drive RGC loss remains debated.¹⁴ This review updates and extends the authors’ 2015 synthesis.¹⁵ Since that publication, GWAS meta-analyses have expanded the locus catalogue to more than 127, polygenic risk scores have reached clinically actionable predictive accuracy, and mechanistic evidence for pathway convergence has consolidated. Beyond cataloguing variants, we systematically map each genetic subtype to its IOP profile, optic nerve morphology, systemic associations, and treatment response. Genetic testing integration into routine clinical practice, has limited value in POAG right now because the genetics is still poorly understood. The handful of high penetrance genes like MYOC explain only a small share of cases, maybe 5% or less, so a negative result tells you almost nothing. The common variants found by GWAS each carry tiny effect sizes, mostly odds ratios near 1.1 to 1.5, which means no single SNP predicts who will develop disease.^16–18 Polygenic risk scores try to combine them but were built mainly in European cohorts and lose accuracy in African, Asian, and Middle Eastern patients, the groups where POAG is often most severe. And even a positive result rarely changes anything, since management still comes down to lowering intraocular pressure regardless of genotype, so the test adds cost without yet earning its place in routine care.

Materials and Methods

A systematic PubMed/MEDLINE search was conducted from inception through April 2025 using Boolean combinations of MeSH terms and free-text keywords covering POAG genetics, susceptibility genes, GWAS loci, polygenic risk scores, and specific gene symbols (MYOC, OPTN, TBK1, WDR36, FOXC1, CYP1B1, CAV1/CAV2, CDKN2B-AS1, TMCO1, and related terms). For each confirmed gene, a targeted secondary search retrieved clinical literature on phenotype, diagnosis, treatment outcomes, and prognosis. Reference lists of key reviews were hand-searched. Variant pathogenicity was cross-referenced against OMIM, ClinVar, and the NHGRI-EBI GWAS Catalog.

Eligible designs included peer-reviewed original research, GWAS and GWAS meta-analyses, candidate-gene studies, family-based linkage analyses, cohort and case-control studies, polygenic risk score development, and gene-specific mechanistic studies in human subjects. Related endophenotypes (IOP, vertical cup-to-disc ratio, optic disc morphology, RNFL thickness) were eligible. Clinical correlation studies required a confirmed POAG diagnosis, a genetically characterized sample with a pathogenic variant or risk allele, and at least one reported clinical outcome. Studies lacking full text, duplicates, non-English publications, preprints, animal-only studies without translational relevance, and studies focused on secondary glaucoma without POAG-relevant genetic data were excluded. Quality of genetic association studies was assessed with the Q-Genie tool; observational clinical studies were appraised with the Newcastle-Ottawa Scale. This systematic review was conducted and reported in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) 2020 guidelines.

Heritability and Genetic Architecture

POAG is highly heritable but inherited in a non-Mendelian pattern in most patients. High-penetrance single-gene mutations account for less than 10% of cases; the remainder reflects cumulative burden of common variants with modest effect sizes, defining a polygenic architecture. GWAS in multiple populations has identified roughly 100 to 127 loci linked to POAG susceptibility or related endophenotypes, most enriched for genes regulating extracellular matrix, cell adhesion, vascular development, and neuronal support.^11,19 Linkage analyses in familial cases mapped GLC1A through GLC1P loci, confirming substantial locus heterogeneity.¹⁹

High-Penetrance Mendelian Causative Genes

Seven genes carry rare, high-penetrance variants that cause familial or early-onset POAG. Their combined characteristics are summarized in Table 1.

Table 1.

High-Penetrance Mendelian Causative Genes and Their Contribution to POAG Pathogenesis

Gene Symbol	Gene Name	Contribution to POAG Pathogenesis	Reference No.
MYOC	Myocilin	Mutations cause protein misfolding, ER stress, and TM dysfunction, elevating IOP. Accounts for ~3.5% of POAG; autosomal dominant, high penetrance. p.P370L induces mitochondrial dysfunction and TM cell death. Gain-of-toxic-function mechanism proposed.	[9,10]
OPTN	Optineurin	Adaptor protein regulating autophagy, NF-kB signaling, and Golgi maintenance. Primarily linked to NTG. E50K triggers RGC apoptosis; p.H486R disrupts NF-kB regulation. Shares a convergent pathway with TBK1; overexpression upregulates MYOC in TM cells. Autosomal dominant
TBK1	TANK-Binding Kinase 1	CNV duplications account for ~1% of NTG. Phosphorylates OPTN; regulates NF-kB and autophagy. Overexpression impairs autophagosome clearance, causing RGC apoptosis. Shares pathway with OPTN. Role in high-tension POAG unestablished. Autosomal dominant	[20,21]
WDR36	WD Repeat-Containing Protein 36	Accounts for 1.6–17% of POAG. Expressed in TM, retina, and optic nerve. Mutations activate the p53 pathway and impair RGC axon formation without IOP elevation. Effect sizes are inconsistent across studies. Autosomal dominant	[22]
NTF4	Neurotrophin 4	Activates TrkB on RGCs; essential for neuroprotection. Mutations impair TrkB signaling and neuronal growth, causing RGC loss. Associated with POAG in European and Chinese populations; not replicated in Caucasian or Indian cohorts. Autosomal dominant	[23]
FOXC1	Forkhead Box C1	Transcription factor for anterior segment and TM development. Mutations impair TM stress response and reduce IOP-lowering therapy responsiveness via altered prostaglandin secretion. Dosage-sensitive: gain and loss of function both cause ocular defects. Autosomal dominant	[24,25]
CYP1B1	Cytochrome P450 Family 1 Subfamily B Polypeptide 1	Digenic modifier of MYOC; associated with congenital glaucoma and anterior segment dysgenesis. Mutations alone are insufficient for POAG; pathogenicity requires co-occurrence with MYOC variants (GLC1A phenotype). Autosomal recessive	[26,27]

Note: Gene’s symbols are in bold.

Abbreviations: AD, autosomal dominant; AR, autosomal recessive; JOAG, juvenile-onset open-angle glaucoma; NTG, normal-tension glaucoma; PCG, primary congenital glaucoma; ARS, Axenfeld–Rieger syndrome; TM, trabecular meshwork; ER, endoplasmic reticulum; ALS, amyotrophic lateral sclerosis; IOP, intraocular pressure; KKESH, King Khaled Eye Specialist Hospital and Research Center.

MYOC (Myocilin)

MYOC (chromosome 1q24, GLC1A) encodes myocilin, expressed in trabecular meshwork (TM) cells and the ciliary body. Over 100 pathogenic variants account for approximately 3.5% of POAG cases.^9,28 Mutant myocilin accumulates in the endoplasmic reticulum (ER), triggering the unfolded protein response and TM cell apoptosis.^8,29 Juvenile-onset POAG presents at ages 3 to 35 with IOP of 26 to 55 mmHg; adult-onset variants (p.Pro370Leu, p.Gln368Stop) cause milder phenotypes (IOP 22 to 30 mmHg) with slower progression.³⁰ The p.Pro370Leu variant causes mitochondrial dysfunction and TM cell death; p.Gln368Stop has notably low penetrance. No systemic associations are confirmed. Prostaglandin analogues reduce IOP by 25% to 35%, but most JOAG eyes require trabeculectomy or glaucoma drainage device (GDD). The chemical chaperone 4-phenylbutyrate (4-PBA) shows preclinical promise in relieving ER stress.¹⁰

OPTN (Optineurin)

OPTN (GLC1E, chromosome 10p13) encodes a 67-kDa cytoplasmic adaptor coordinating selective autophagy, mitophagy, NF-kappaB signaling, and vesicular trafficking in the retina and TM.^8,29,31 The p.Glu50Lys (E50K) variant disrupts vesicular TM trafficking and triggers RGC apoptosis; loss-of-function alleles (p.His486Arg, p.Glu322Lys) impair NF-kappaB regulation.³¹ OPTN-associated disease presents predominantly as normal-tension glaucoma (IOP ≤21 mmHg in approximately 70% of carriers), with deep optic cup excavation and focal inferior notching. Progression is slow but lifetime visual field loss risk is substantial, worsened by myopia >-3 D or disc hemorrhage. Heterozygous loss-of-function variants are associated with amyotrophic lateral sclerosis (ALS) and Paget disease; neurological screening is warranted. Nicotinamide 3 g/day showed significant visual field preservation in a Phase II NTG trial and is the best-evidenced neuroprotective intervention currently available.³²

TBK1 (TANK-Binding Kinase 1)

TBK1 (GLC1P, chromosome 12q14) encodes a serine/threonine kinase that phosphorylates optineurin and regulates autophagy initiation and NF-kappaB signaling.^8,28,33 The pathogenic variants are copy-number duplications, accounting for approximately 1% of NTG cases, which hyperactivate mTORC1 and impair autophagosome clearance. This converges on the same RGC mitochondrial clearance deficit as OPTN loss-of-function. TBK1-NTG presents at mean age 50 to 60 with IOP ≤21 mmHg but severe rim thinning; mean deviation declines at approximately 1.0 to 1.5 dB/year, faster than idiopathic NTG. ALS co-segregation has been reported in TBK1 duplication families; motor neuron screening is warranted. ROCK inhibitors (netarsudil) and mTOR inhibitors have mechanistic rationale; rapamycin protects RGCs in TBK1-overexpression mouse models. Critically, a negative TBK1 point-mutation screen does not exclude TBK1-axis NTG, since pathogenicity rests on copy-number variation.

WDR36

WDR36 (GLC1G, chromosome 5q22) encodes a WD40-repeat protein involved in ribosomal RNA processing and T-cell activation.^8,29 Reported POAG contribution ranges widely (1.6% to 17%), reflecting population heterogeneity and inconsistent replication. WDR36 knockdown in zebrafish reduces 18S rRNA levels and activates the p53 stress-response pathway, causing ocular defects; transgenic mouse models show progressive RGC axon abnormalities without IOP elevation.³⁴ WDR36 is best characterized as a risk modifier rather than a standalone high-penetrance Mendelian cause; failed replication in independent European cohorts supports reclassification. No variant-specific treatment is validated; standard IOP-lowering applies.

NTF4 (Neurotrophin 4)

NTF4 (chromosome 19q13.33) encodes neurotrophin-4, which activates TrkB receptors on RGCs to support survival.^28,35 Pathogenic variants disrupt TrkB signaling and neuronal growth in vitro, causing progressive RGC loss. POAG association has been reported in European and Chinese cohorts but not consistently replicated in Caucasian or Indian populations; NTF4 is best categorized as a low-to-moderate evidence susceptibility modifier rather than a confirmed high-penetrance locus. No variant-specific treatment exists; standard IOP management applies.

FOXC1

FOXC1 (chromosome 6p25.3) is a forkhead transcription factor regulating ocular and craniofacial development.^28,36 Both haploinsufficiency and overdosage disrupt TM homeostasis and impair aqueous outflow through reduced prostaglandin secretion. Heterozygous variants cause Axenfeld-Rieger syndrome (ARS, OMIM #602482): anterior segment dysgenesis, dental anomalies, facial dysmorphism, and umbilical abnormalities; glaucoma affects 50% to 75% of carriers, often severe from childhood. Systemic evaluation should include cardiology (approximately 15% congenital heart defects) and endocrinology. FOXC1 mutations partially explain the reduced responsiveness of ARS patients to prostaglandin analogues (approximately 30% to 40% IOP reduction); GDD (Baerveldt or Ahmed) is preferred over trabeculectomy. Goniotomy is first-line in pediatric ARS. FOXC1-associated glaucoma is best classified as a developmental subtype distinct from classical POAG.

CYP1B1

CYP1B1 (chromosome 2p22.2) encodes a cytochrome P450 enzyme essential for TM and Schlemm canal formation.^28,37 Biallelic loss-of-function variants cause primary congenital glaucoma (OMIM #231300), accounting for 20% to 40% of cases in outbred populations and 85% to 100% in consanguineous Middle Eastern and South Asian families. CYP1B1 also acts as a digenic modifier of MYOC, worsening the GLC1A phenotype.¹³ PCG presents at birth to age 3 with buphthalmos, photophobia, and epiphora; IOP is typically 28 to 50 mmHg. Goniotomy achieves 75% to 85% surgical success at 5 years; gonioscopic severity, not genotype, predicts outcome. Genetic counseling is essential (autosomal recessive; carrier frequency approximately 1:50 in Saudi Arabia). Table 1, list high-penetrance Mendelian genes contributing to POAG.

Common Variant Susceptibility Loci

GWAS meta-analyses across multiple ancestries have implicated 100 to 127 common loci in POAG susceptibility, converging on five mechanistic axes.^11,38 Representative loci are summarized in Table 2.

Table 2.

GWAS Loci, Candidate Gene SNPs, and Their Contribution to POAG Risk

SNP / Variant	Gene	Contribution to POAG Risk and Pathogenesis	Ref.
rs4236601, rs1052990	CAV1/CAV2	Caveolin proteins regulate TM transport, eNOS activity, and aqueous outflow; variants alter IOP and ECM integrity. Associated with POAG in Caucasian, Chinese, Korean, and Pakistani cohorts. Odds ratio =1.36 in Icelandic discovery. Meta-analysis pooled OR 1.23 (95% CI 1.12 to 1.34).	[12,39]
rs4656461	TMCO1	ER transmembrane protein maintaining calcium homeostasis. Risk allele confers 1.7× POAG risk per allele and advances onset by 3–4 years; linked to IOP elevation and TM vulnerability. Combined OR 1.51 (95% CI 1.35 to 1.68).	[40,41]
rs33912345 (p.N141H)	SIX6	Homeobox transcription factor; p.N141H causes loss of function impairing retinal development and increasing RGC susceptibility to POAG. Associated with VCDR in Japanese and Caucasian populations. OR 1.32	[41,42]
Multiple SNPs	CDKN2B-AS1 (ANRIL)	lncRNA at 9p21.3 regulating CDKN2A/CDKN2B. Variants associate with IOP/VCDR; risk alleles sensitize RGCs to normal IOP. Most replicated POAG locus across ancestries; strongly linked to NTG. Combined OR 1.39 (95% CI 1.28 to 1.51).	[41]
Multiple SNPs	ABCA1	RGC-expressed membrane transporter modulating IOP via Cav1/eNOS/NO signaling. Elevated in glaucomatous tissue; implicated as both a risk factor and therapeutic target. OR 1.31	[12]
Multiple SNPs	AFAP1	Actin filament-associated protein in TM and retinal astrocytes. Variants disrupt actin cytoskeletal integrity under cellular stress, impairing aqueous outflow regulation. OR 1.20	[43]
Multiple SNPs	ARHGEF12	RhoGEF activating RhoA/ROCK signaling; reduces Schlemm’s canal permeability and raises IOP. Links ABCA1, CAV1/CAV2, and GAS7 in a shared outflow network; ROCK inhibitors show therapeutic promise. OR 1.19 (95% CI 0.71 to 2.0),	[43]
rs9913911	GAS7	Expressed in TM, ciliary body, and retina; regulates TM remodeling via microfilament rearrangement. Interacts with MYOC and CAV1 through RhoA/beta-catenin; associated with elevated IOP and cup volume. A risk allele (p = 0.0004) but reported no clean OR.	[12]
Multiple SNPs	ATXN2	Multi-tissue protein regulating calcium homeostasis, cytoskeletal organization, and autophagy. SNPs associate with IOP and POAG; autophagy role links ATXN2 mechanistically to the TBK1/OPTN pathway in NTG. OR 1.17	[43]
p.R140W	EFEMP1	Fibulin-3 ECM protein; p.R140W causes intracellular aggregation disrupting retinal/RPE ECM integrity. Associated with JOAG/POAG in autosomal dominant pattern; severity correlates with retention degree. No OR	[43,44]
CNV deletion	GALC	Galactosylceramidase required for myelin maintenance. Heterozygous CNV deletion confers 4× POAG risk by impairing optic nerve myelination; proposed biomarker and therapeutic target. OR 5.0 (95% CI 1.6 to 16.4)	[43]
rs7916697, rs1900004	ATOH7	bHLH transcription factor for RGC differentiation and optic nerve development; variants associate with optic disc morphology and POAG. Epistatic interaction with Raftlin SNPs amplifies incidence. OR 1.08,	[44]
Multiple SNPs	TXNRD2	Mitochondrial thioredoxin reductase; first mitochondrial gene linked to POAG. Reduced ROS neutralization impairs RGC oxidative defense; associated in African, European, and Brazilian populations. 0.78. Below 1, so the indexed allele is protective, not a risk allele.	[45]
p.T104M	IL20RB	Cytokine receptor in TM cells; p.T104M abolishes IL-19/IL-20/IL-24 binding, disrupting TM signaling. Found in 10 affected members of a POAG family; epistasis suspected. No odds ratio.	[36]
p.T255T, p.A290A	ASB10	Ubiquitin-pathway negative regulator in TM cells. RNAi silencing reduced aqueous outflow by ~50%; non-synonymous variants account for 6% of POAG in US and German cohorts. No stable odds ratio	[33,34]
148Asp/Glu	APE1	DNA repair endonuclease. Variant associates with POAG risk in homozygote/recessive models (562/644 controls); impaired oxidative DNA repair in RGCs and TM cells proposed as mechanism. No OR	[15]
rs449647	APOE	Apolipoprotein in lipid transport and neurodegeneration. POAG association across multiple genetic models (1,937/1,958 controls); may affect optic nerve lipid homeostasis or amyloid pathways. OR 1.15	[15]
rs1799750	MMP1	Matrix metalloproteinase regulating ECM degradation and collagen remodeling. Variant implicated in altered TM outflow resistance through dysregulated ECM turnover. OR 1.64 (95% CI: 1.05–2.56)	[15]
rs7081455	PLXDC2	Transmembrane receptor in angiogenesis and vascular biology. Significant allelic association (1,428/1,222 controls); putative role in optic nerve head vascular integrity. Global OR 1.3–1.5	[15]
rs1279683	SLC23A2	Ascorbic acid transporter. Variant associated with POAG in allelic model (400/400 controls); reduced ascorbate transport may impair antioxidant defense in TM and RGCs. OR 1.67 (95% CI: 1.03–2.71)	[15]
372T/C	TIMP1	MMP inhibitor regulating ECM homeostasis. Variant associated with POAG in recessive model (451/509 controls); altered MMP inhibition may disrupt TM and lamina cribrosa remodeling. No OR, single study	[15]
rs1927911–rs10759930	TLR4	Innate immune toll-like receptor. Five SNPs associate with POAG across multiple cohorts (669–769 cases), implicating neuroinflammation and microglial activation in RGC loss, particularly NTG. OR 1.34 (95% CI: 1.06–1.71)	[15]
399Arg/Gln	XRCC1	DNA single strand break repair protein. Allelic, heterozygote, and dominant POAG associations (742/840 controls); impaired repair may predispose TM and RGCs to cumulative genotoxic damage. OR 2.50 (95% CI, 1.54–4.07).	[15]
rs540782, rs547984, rs693421	ZP4	Zona pellucida glycoprotein. All three SNPs show allelic POAG association across two to three independent cohorts; pathogenic mechanism remains under investigation. OR 1.34 (95% CI: 1.61–1.54).	[15]

Abbreviations: AD, autosomal dominant; bHLH, basic helix-loop-helix; CNV, copy-number variant; ECM, extracellular matrix; eNOS, endothelial nitric oxide synthase; ER, endoplasmic reticulum; JOAG, juvenile-onset open-angle glaucoma; LncRNA, long non-coding RNA; NTG, normal-tension glaucoma; POAG, primary open-angle glaucoma; RGC, retinal ganglion cell; RNFL, retinal nerve fiber layer; ROS, reactive oxygen species; SNP, single-nucleotide polymorphism; TM, trabecular meshwork; VCDR, vertical cup-to-disc ratio.

Trabecular Outflow Regulation

CAV1/CAV2 (rs4236601, 7q31.2) suppress endothelial nitric oxide synthase and reduce nitric oxide production, increasing outflow resistance; the risk allele raises IOP by 0.4 to 0.6 mmHg per allele and POAG risk by 20% to 30% across Caucasian, Chinese, Korean, and Pakistani populations.⁴⁴ ARHGEF12 (LARG, 11q23.3) activates RhoA/ROCK signaling and reduces Schlemm canal cell permeability, with one of the largest per-allele IOP effects in the GWAS catalog; ROCK inhibitors (netarsudil, ripasudil) are the biologically rational treatment for carriers and link ABCA1, CAV1/CAV2, and GAS7 in a shared outflow network.⁴⁵ AFAP1 (4p16.1) encodes an actin filament-associated protein that modulates TM cytoskeletal tension under cellular stress. ABCA1 (9q31.1) regulates TM ECM lipid composition through Cav1/eNOS/NO signaling; expression is elevated in glaucomatous tissue and animal models of high IOP, suggesting both a risk factor and therapeutic target.²⁰ ASB10 (7q36.1) is a TM-expressed ubiquitin ligase; RNAi silencing reduces aqueous outflow by approximately 50%, and non-synonymous variants account for 6% of POAG in US and German cohorts.²¹

Mitochondrial and ER Homeostasis

TMCO1 (1q24.1, rs4656461) encodes an ER calcium-release channel; the risk allele confers 1.7-fold POAG risk and 3 to 4 years earlier onset, with biallelic variants causing cerebrofaciothoracic dysplasia.²² TXNRD2 (22q11.21) is the primary mitochondrial antioxidant; reduced activity increases RGC reactive oxygen species and is enriched in NTG. Nicotinamide and CoQ10 are candidate neuroprotective adjuncts. Notably, TXNRD2 association was not replicated in a Saudi cohort, illustrating ancestry-dependent variability.

Optic Nerve Structural Vulnerability

SIX6 p.His141Asn (rs33912345, 14q23.1) reduces developmental RGC reserve, predisposing to cupping at lower IOP; lower IOP targets (≤12 mmHg) and early OCT baselines are warranted.²³ CDKN2B-AS1/ANRIL (9p21.3) regulates TM and lamina cribrosa cell-cycle progression; risk alleles independently associate with IOP, VCDR, and RNFL thickness, and are among the most replicated POAG loci with strong correlation to NTG.²⁴ ATOH7/Math5 (10q21.3, rs1900004) governs RGC fate specification; risk carriers begin adult life with reduced RGC reserve and warrant aggressive IOP targets. Importantly, the ATOH7 association seen in Australian, Singaporean, and Latino cohorts was not replicated in Saudi patients, and an epistatic interaction with Raftlin lipid linker 1 amplifies POAG incidence.

Extracellular Matrix Integrity

GAS7 (17p13.1, rs9913911) influences TM cell remodeling via microfilament rearrangement and interacts with MYOC and CAV1 through beta-catenin and RhoA signaling. EFEMP1 (2p16.1) encodes fibulin-3; the missense variant p.R140W disrupts TM basement-membrane ECM and co-causes Doyne honeycomb retinal dystrophy. Macular OCT is indicated in POAG patients carrying this variant.²⁵

Neuroinflammation

IL20RB (GLC1C, 3q22.3) encodes an IL-20 family receptor; the rare p.Thr104Met variant abolishes cytokine signaling in TM cells, linking innate immune pathways to outflow biology. GALC (14q31.3) heterozygous CNV deletions confer approximately 4-fold POAG risk through impaired optic nerve myelination. ATXN2 (12q24.12) connects autophagy and stress-granule clearance to IOP regulation and overlaps mechanistically with the TBK1/OPTN pathway.²⁶ A meta-analysis of candidate genes identified significant POAG associations across 20 SNPs in 12 genes including APE1, APOE, CAV1/CAV2, MMP1, OPTN, PLXDC2, SLC23A2, TIMP1, TMCO1, XRCC1, and ZP4.⁴⁰ Five TLR4 variants showed the most consistent associations, supporting the neuroinflammatory axis. Three ZP4 SNPs (rs540782, rs547984, rs693421) were significantly associated across allelic models in pooled cohorts comprising 919 to 1334 cases.

Selected GWAS and candidate-gene loci associated with POAG risk are listed in Table 2.

Polygenic Risk Scores

A POAG polygenic risk score (PRS) aggregates weighted common-variant effects from GWAS summary statistics. Current scores span more than 127 loci; individuals in the top decile carry approximately 15-fold the average POAG risk.⁴³ An early POAG-specific PRS derived from MTAG analysis integrating POAG, IOP, and VCDR GWAS data incorporated 2673 uncorrelated variants. In the Ocular Hypertension Treatment Study (OHTS) cohort, mean PRS was 0.24 in converters vs −0.12 in non-converters; PRS alone yielded an AUC of 0.72, rising to 0.79 with age and sex, and to 0.89 when combined with IOP and VCDR.^42,43

PRS integration refines POAG management at three decision points. For screening, OHTS post-hoc analysis showed the highest PRS quartile had a 2.35-fold hazard ratio for POAG conversion vs the lowest quartile, independent of IOP and CDR; low-risk quartile patients showed no significant benefit from prophylactic IOP-lowering, supporting PRS-selective intervention. For treatment selection, MYOC high-tension subtypes respond robustly to prostaglandin analogues and surgical drainage; OPTN/TBK1 NTG subtypes are less responsive to IOP reduction alone and warrant neuroprotective co-treatment. PRS also refines penetrance estimates for monogenic carriers, including MYOC p.Gln368Stop. Combining family history, age, and sex gives an AUC of 0.73; adding PRS raises it to 0.80, and the two measures are complementary rather than interchangeable.

The principal limitation is ancestry. Current scores are predominantly trained on European-ancestry cohorts and lose 50% to 70% of predictive accuracy in African-ancestry populations, precisely those carrying the highest disease burden.⁴³ Linkage disequilibrium structure is ancestry-dependent, constraining trans-population transferability. Clinical deployment should be restricted to populations well-represented in training data until trans-ethnic scores are validated. Commercial tools such as Sight Score are emerging; wider adoption depends on closing the ancestry representation gap and validating integrated multifactorial models in diverse prospective cohorts.

Convergence of Genetic Pathways

Mendelian mutations and common GWAS variants reach the same biological pathways through different molecular entry points. Five convergence axes account for most of the accumulated evidence.

Trabecular meshwork dysfunction and aqueous outflow impairment: MYOC mutations cause protein misfolding and ER stress in TM cells; CAV1/CAV2 risk alleles reduce eNOS-mediated nitric oxide production; ARHGEF12 variants increase RhoA/ROCK-mediated cytoskeletal rigidity. All three reduce outflow facility through distinct mechanisms.

Retinal ganglion cell survival: NTF4 and ATOH7 disrupt neurotrophin signaling; TXNRD2 alleles increase mitochondrial oxidative stress; CDKN2B-AS1 and SIX6 promote apoptosis via cell-cycle dysregulation. The net effect is reduced RGC counts before clinical glaucoma is detectable.

Autophagy and protein homeostasis: OPTN loss-of-function impairs selective autophagy of damaged organelles; TBK1 copy-number gains deregulate autophagosome clearance; ATXN2 intermediate repeats disrupt stress-granule dynamics. These three genes define the molecular basis of the NTG neurodegenerative subtype.

Neuroinflammation and NF-kappaB signaling: OPTN E50K and TBK1 (a direct NF-kappaB kinase) link immune dysregulation to RGC loss independently of IOP, as do IL20RB and TLR4 variants. This axis is the most plausible explanation for why some NTG progresses despite well-controlled pressure.

Extracellular matrix and structural integrity: LOXL1, AFAP1, and EFEMP1 variants compromise the mechanical support of the optic nerve head and trabecular meshwork, increasing susceptibility to pressure-induced cupping at IOP levels that would not damage a structurally normal nerve.

These axes are not independent. CYP1B1 modifies MYOC penetrance through digenic inheritance;¹³ OPTN and TBK1 cooperate in the same autophagy pathway such that variants in either gene compound dysfunction;⁴¹ the ATOH7/Raftlin epistatic interaction affects optic nerve development independently of IOP. The clinical corollary is direct: NTG driven by the autophagy-neuroinflammation axis responds differently to IOP reduction than high-tension POAG, and shared pathway nodes, particularly ROCK and mTOR kinases, are rational targets for mechanism-based therapy across genetically heterogeneous patients.

Clinical Genotype-Phenotype Subtypes

The expanding catalogue of POAG-associated genes reveals one clinical reality: patients grouped under the same diagnosis often have different dominant pathological mechanisms, IOP signatures, and prognoses. Each genetic subtype maps to a characteristic pressure profile, optic nerve morphology, systemic associations, progression rate, and treatment response.

High-Tension Surgical Subtype

Monogenic MYOC variants, particularly p.Pro370Leu, p.Gln368Stop, and p.Tyr437His, define the archetypal high-tension POAG subtype. Patients present with IOP frequently exceeding 30 mmHg, a large cup-to-disc ratio at a young age, and heavily asymmetric structural damage. Despite the severe structural phenotype, central visual acuity is often preserved until advanced disease, and response to IOP-lowering (pharmacological or surgical) is robust. Early trabeculectomy with mitomycin C or GDD implantation is recommended when medical therapy fails to achieve target IOP.

Normal-Tension Neurodegenerative Subtype

OPTN p.Glu50Lys, TBK1 copy-number duplications, and TXNRD2 risk alleles collectively define an NTG sub-phenotype dominated by IOP-independent neurodegeneration. Patients have statistically normal IOP at diagnosis yet present with characteristic inferior notching, parapapillary beta-zone atrophy, and a history of optic disc hemorrhages. Progression rates exceed those predicted by the modest pressure levels, pointing to mitochondrial dysfunction, oxidative stress, and disrupted autophagy as co-drivers. Nicotinamide 3 g/day has Phase II evidence of visual field preservation in NTG and is the most evidence-supported neuroprotective intervention currently available.³² The strength of evidence varies by variant: OPTN E50K has robust familial segregation data, while other OPTN alleles and TBK1 point mutations have inconsistent replication.

Syndromic and Developmental Glaucoma Subtype

FOXC1, CYP1B1, and TMCO1 biallelic variants represent glaucoma embedded within broader developmental syndromes. These patients require coordinated multidisciplinary care: pediatric ophthalmology for amblyopia prevention, pediatric dentistry and craniofacial surgery for ARS, systemic genetics for syndrome delineation, and early surgical planning. The trabecular meshwork is structurally abnormal in all three subtypes, making angle surgery (goniotomy and trabeculotomy) the preferred primary intervention over trabeculectomy in pediatric patients.

Vascular-Dysregulation Subtype

CAV1/CAV2 risk alleles and TXNRD2 variants both impair nitric oxide production and mitochondrial redox balance, converging on a vascular-dysregulation NTG sub-phenotype. Patients often carry systemic vascular markers: low blood pressure, migraines, Raynaud phenomenon, and sleep apnea, and may show flickering visual field defects correlating with episodes of nocturnal hypotension. Management targets the systemic drivers: blood pressure optimization, avoidance of excessive nocturnal dipping, NO-donating IOP-lowering agents (latanoprostene bunod, netarsudil), and mitochondrial antioxidant supplementation.

Structural-Vulnerability Subtype

ATOH7, SIX6, and GAS7 risk variants define individuals who begin adulthood with structurally vulnerable optic nerves: smaller disc area, thinner neuroretinal rim, lower peak RGC counts, and weaker lamina cribrosa, resulting in accelerated cupping at any given IOP. Management should emphasize lower IOP targets, more frequent structural imaging (OCT every 3 to 6 months in the first 2 years after diagnosis), and earlier initiation of medical therapy, even in ocular hypertension without definitive field loss.

Ancestry-Specific Considerations

POAG prevalence is 3 to 5-fold higher in African-ancestry individuals than in European populations, with younger onset and more severe disease at presentation. African-ancestry POAG features larger optic discs (requiring population-specific normative OCT databases), more frequent NTG presentation, and higher disc hemorrhage rates. Current IOP-based diagnostic criteria are calibrated primarily on European populations, creating a diagnostic blind spot.

Saudi Arabia and the broader Arabian Gulf region have among the highest reported frequencies of CYP1B1 biallelic variants, with founder mutations p.Gly61Glu, p.Arg368His, and p.Arg390 accounting for the majority of PCG cases. Data from one tertiary center on more than 800 CYP1B1-confirmed PCG eyes over two decades demonstrate that early goniotomy (ideally within the first three months of life) achieves IOP control in approximately 80% of cases with a single procedure. Consanguinity rates exceeding 50% in affected families underscore the critical need for genetic counseling and cascade screening.

Discussion

A useful framework for integrating these observations is to conceptualize POAG genetic architecture as a four-tier hierarchy. Tier 1 contains the single firmly validated Mendelian gene, MYOC, with reproducible segregation across ancestries and a well-characterized ER-stress mechanism; the key caveat is that animal models do not reliably develop glaucoma, indicating mutation-specific ER accumulation thresholds rather than simple dosage effects. Tier 2 contains genes with strong but ancestry-dependent or variant-specific evidence: OPTN (E50K specifically), TBK1 (copy-number gains), FOXC1 (developmental subtype), and CYP1B1 (biallelic PCG; digenic in POAG). Tier 3 contains modifier genes whose independent pathogenicity is disputed: WDR36 and NTF4, which likely amplify risk in specific genetic contexts rather than cause disease in isolation. Tier 4 contains the GWAS susceptibility loci, approximately 100 to 127 common variants each contributing modest individual risk but collectively defining the polygenic architecture.

This four-tier model has direct implications for genetic counseling: Tier 1 mutations warrant cascade family screening; Tier 2 variants require ancestry-informed penetrance counseling; Tier 3 variants should be reported as modifiers with uncertain clinical significance; Tier 4 variants are most useful in aggregate PRS form rather than as individually actionable findings. Risk is not simply additive: a MYOC pathogenic-variant carrier with a high polygenic background score has earlier onset and more severe phenotype than the monogenic variant alone predicts, because common variants modify penetrance. Conversely, a patient with NTG and no identifiable Mendelian mutation but a PRS enriched in OPTN- and TBK1-pathway SNPs is a better candidate for neuroprotective adjuncts than for aggressive IOP reduction. ROCK inhibitors for ARHGEF12/CAV1-axis patients, pharmacological chaperones (4-PBA) for MYOC-ER stress, and nicotinamide for OPTN/TXNRD2 mitochondrial dysfunction are each grounded in the pathway logic the genetics reveals.

Among the 100 to 127 GWAS loci, evidence quality varies. CAV1/CAV2, CDKN2B-AS1, SIX6, TMCO1, and ARHGEF12 are replicated across multiple ancestries with functional support; GALC and ASB10 require further multi-ancestry validation. Statistical association does not establish causation; for most loci, the causal variant remains unmapped and the mechanism unconfirmed. ARHGEF12 is the most translationally actionable: its RhoA/ROCK pathway is directly targeted by approved medications.

Approximately 76% of genetic studies enrolled exclusively European-ancestry cohorts, producing risk models least predictive where prediction is most needed: in African and Asian populations with 3 to 5-fold higher POAG prevalence. The Saudi/Gulf Arab dataset is one of the few non-European anchors in this review, highlighting CYP1B1 founder mutation frequencies in consanguineous populations and the incomplete transferability of European GWAS findings. Future reviews should require pre-specified ancestry stratification as a reporting standard.

Limitations

Five methodological limitations affect interpretation. First, evidence quality is heterogeneous: 16.7% of appraised genetic association studies scored low on Q-Genie and 23.3% of clinical observational studies were high risk of bias on NOS; both sets were retained for qualitative synthesis only. Second, ancestral homogeneity dominates: 76% of genetic studies enrolled exclusively European cohorts and fewer than 5% enrolled African, Middle Eastern, or South Asian participants, constraining PRS generalizability. Third, follow-up is short: median clinical follow-up was 2.4 years, insufficient for a disease that progresses over decades. Fourth, most clinical studies recruit from tertiary centers, skewing phenotypic descriptions toward severe presentations. Fifth, heterogeneity in design, genotyping platform, and outcome definition precluded formal meta-analytic pooling; all effect estimates derive from individual studies.

Conclusion

POAG has no single unifying mechanism. High-penetrance Mendelian genes (MYOC, OPTN, TBK1, FOXC1) account for a small fraction of cases, but each has illuminated a core pathological pathway: trabecular meshwork protein misfolding and ER stress, impaired selective autophagy, NF-kappaB-driven neuroinflammation, and RGC apoptosis. GWAS has extended the locus count above 127, converging on five biological axes that partially overlap those Mendelian pathways: extracellular matrix remodeling, Rho/ROCK cytoskeletal tension, mitochondrial oxidative stress, optic nerve development, and cell cycle dysregulation. No single axis dominates; POAG results from their cumulative intersection, a polygenic architecture that resists single-target intervention.

Two constraints temper clinical translation. The first is ancestral bias: most GWAS data come from European cohorts, and PRS accuracy falls substantially in African and Asian populations who carry a disproportionate share of severe, early-onset disease. The second is the step from statistical association to mechanism; fewer than a third of the 127 loci have validated functional models. Whole-genome sequencing, ancestry-diverse recruitment, and mechanistic validation are the three prerequisites for the next generation of POAG genetics. The canonical IOP-centric model is incomplete; normal-tension cases driven by OPTN, TBK1, and CDKN2B-AS1 confirm that neurodegeneration and RGC susceptibility operate independently of pressure. The ultimate translational goal is converting genetic stratification into targeted neuroprotection before irreversible vision loss occurs.

Funding Statement

This research received no external funding.

Abbreviations

ALS, amyotrophic lateral sclerosis; ARS, Axenfeld-Rieger syndrome; AUC, area under the curve; bHLH, basic helix-loop-helix; CDR, cup-to-disc ratio; CNV, copy-number variant; ECM, extracellular matrix; eNOS, endothelial nitric oxide synthase; ER, endoplasmic reticulum; GDD, glaucoma drainage device; GWAS, genome-wide association study; HTG, high-tension glaucoma; IOP, intraocular pressure; JOAG, juvenile-onset open-angle glaucoma; MTAG, multi-trait analysis of GWAS; NTG, normal-tension glaucoma; OCT, optical coherence tomography; OHTS, Ocular Hypertension Treatment Study; PCG, primary congenital glaucoma; POAG, primary open-angle glaucoma; PRS, polygenic risk score; RGC, retinal ganglion cell; RNFL, retinal nerve fiber layer; ROCK, Rho-associated kinase; ROS, reactive oxygen species; SNP, single-nucleotide polymorphism; TM, trabecular meshwork; VCDR, vertical cup-to-disc ratio.

Data Sharing Statement

All data supporting the findings of this review are derived from published studies cited in the reference list. No new or unpublished datasets were generated or analyzed. The PRISMA 2020 checklist is provided as Supplementary Material (Supplementary Table S1).

Ethics Approval and Informed Consent

This systematic review analyzed previously published, publicly available data. No primary data collection involving human subjects was performed; therefore, ethics approval and informed consent were not required.

Author Contributions

KKA: Conceptualization, methodology, writing – original draft, supervision. SA: Data curation, writing – original draft. HA: Formal analysis, writing – review and editing. GS: Writing – review and editing, validation. All authors reviewed and approved the final manuscript. All authors made a significant contribution to the work reported, whether that is in the conception, study design, execution, acquisition of data, analysis and interpretation, or in all these areas; took part in drafting, revising or critically reviewing the article; gave final approval of the version to be published; have agreed on the journal to which the article has been submitted; and agree to be accountable for all aspects of the work.

Disclosure

The authors report no conflicts of interest in this work.