Determining Genetic Cause of Posterior Staphylomas in Eyes with Pathologic Myopia by Whole Exome Sequencing

Abstract

Purpose

To identify genetic variants in posterior staphylomas in eyes with pathologic myopia using whole exome sequencing and to determine possible molecular mechanisms contributing to the pathogenesis.

Design

An observational, case-control study.

Participants

Two hundred sixty-four unrelated Japanese patients with myopia (≤ –0.50 diopters) and posterior staphyloma, which was diagnosed by ultra-widefield OCT, 3-dimensional magnetic resonance imaging, and Optos imaging.

Methods

Whole exome sequencing was performed on genomic DNA from peripheral blood. After variant filtering, the allelic frequencies were compared with control data obtained from East Asian subsets of the 1000 Genomes Project Phase III, the Exome Aggregation Consortium, and the Japanese Multi-Omics Reference Panel using Fisher exact test. A gene panel was constructed based on 13 staphyloma-associated disorders. Variants showing significant frequency differences (P ≤ 0.05) and an overlap of the gene panel were analyzed using gene set enrichment analysis with the DAVID Knowledgebase (v2023q4). Protein–protein interaction analysis was performed to assess functional associations.

Main Outcome Measures

The statistically associated variants and genes, gene set enrichment analysis results, protein–protein interaction networks, and involvement of basement membrane structures, including the inner limiting membrane (ILM) and Bruch membrane, were studied.

Results

Whole exome sequencing identified 16 656 missense variants in 8628 genes. Comparative allele frequency analyses with public databases revealed 3925 variants that had significantly higher allelic frequencies in the subjects. Of these, 81 genes overlapped with a curated staphyloma-related gene panel and were subjected to gene set enrichment analysis. The findings showed enrichment in basement membrane, extracellular matrix, and collagen-related pathways. The COL4A5, COL18A1, COL2A1, and COL9A3 genes are concurrently enriched across these pathways. A missense variant in COL4A5 was identified in 27 patients, and 96.3% of whom were females. Protein–protein interaction analysis demonstrated functional connections among these 4 genes.

Conclusions

Variants in the COL4A5, COL18A1, COL2A1, and COL9A3 genes probably contribute to the pathogenesis of a posterior staphyloma through the disruption of collagen synthesis and basement membrane integrity. This was especially effective for the ILM and Bruch membrane. The COL4A5 variant may cause an ocular-predominant phenotype in heterozygous female carriers, independent of the classical features of Alport syndrome.

Financial Disclosure(s)

The authors have no proprietary or commercial interest in any materials discussed in this article.

The prevalence of myopia and high myopia is increasing worldwide and especially in East Asian countries.^1,2 Among eyes with myopia, those with pathologic myopia (PM) and myopic maculopathy are the major causes of visual impairments and blindness.³ High myopia (spherical equivalent ≤ –6.0 diopters [D]) and extreme myopia (spherical equivalent ≤ –10.0 D) are defined by a high degree of myopic refractive error. Pathologic myopia differs from these myopic eyes by having structural complications. According to the Pathologic Myopia Study Group (META-PM), PM is defined as eyes having myopic maculopathy equivalent to or more severe than diffuse atrophy, i.e., Category 2 in the META-PM classification, or those having a posterior staphyloma.^3,4

Genetic studies have identified multiple loci in several candidate genes that are associated with myopia, high myopia, and extreme myopia.5, 6, 7 However, due to differences in the definition between high myopia and PM and due to difficulties in the imaging methods that correctly identify PM, there has been a scarcity of genetic studies on PM.

Posterior staphylomas are relatively uncommon in the general population, with a prevalence of approximately 2%, and are more frequent in highly myopic eyes, where the prevalence reaches about 50%.^8,9 A posterior staphyloma is defined as an outpouching of a limited area of the posterior segment of the eye with a radius of curvature shorter than that of the surrounding ocular wall. It is widely regarded as a hallmark feature of PM.⁹ It is considered to be the most critical feature distinguishing PM from physiologic myopia.¹⁰ However, a posterior staphyloma can occur in eyes without a long axial length.¹¹ There is increasing evidence that supports a genetic basis for the development of a posterior staphyloma, as it is not only observed in hereditary high myopia but also in nonhighly myopic eyes with inherited retinal disorders such as retinitis pigmentosa.^9,12 In addition, posterior staphylomas have also been reported to be present in other genetic conditions including Leber congenital amaurosis and Stickler syndrome.^13,14 These findings suggest a role for genetic determinants in the pathogenesis of posterior staphylomas, although the specific variants involved have not been identified.

A diagnosis of a posterior staphyloma requires special imaging methods, such as 3-dimensional magnetic resonance imaging (3D-MRI) and, more recently, ultra-widefield OCT (UWF-OCT).15, 16, 17, 18 In some earlier studies, a generalized bowing of the posterior sclera was confused with a posterior staphyloma. However, a posterior staphyloma is defined as an outpouching of a limited area of the posterior segment of the eye. Thus, the detection of the staphyloma edges is critical for the diagnosis.¹⁷ Observations of the fundus in the ultra-widefield images of the eyes are necessary because the borders of most staphylomas extend beyond the edges on conventional fundus photographs.¹⁹ As best we know, there has not been a systematic study on the genetic basis of posterior staphylomas.

Thus, the purpose of this study was to identify genes that are associated with the presence of posterior staphyloma in eyes with PM. To accomplish this, we focused on patients with PM strictly characterized by the presence of a posterior staphyloma as determined by UWF-OCT, 3D-MRI, and widefield fundus imaging. We investigated the genetic etiology of the posterior staphylomas in PM eyes by whole exome sequencing (WES).

Methods

Patients and Controls

The procedures used in this study conformed to the Declaration of Helsinki and were approved by the Ethics Committee of the Institute of Science Tokyo. A signed informed consent was obtained from all patients.

A total of 264 unrelated Japanese patients diagnosed with a posterior staphyloma in the Advanced Clinical Center for Myopia at the Institute of Science Tokyo between September 2006 and August 2023 were studied. The inclusion criterion for pathologic myopia was a myopic refractive error of ≤ –0.50 D and the presence of a posterior staphyloma, according to the Meta-analysis for Pathologic Myopia (META-PM) classification.³ Eyes with poor-quality imaging or a history of scleral procedures that may alter the morphology of the posterior pole of the eye, such as scleral buckling, were excluded.

A posterior staphyloma was confirmed to be present by 3D-MRI, UWF-OCT, or pseudocolor images, autofluorescent images, and infrared images of widefield fundus imaging (Fig 1). The UWF-OCT images were obtained with the Xephilio OCT-S1 instrument (Canon), and widefield fundus imaging was conducted with Optos (Optos, PLC). In the UWF-OCT images, a posterior staphyloma was characterized by a gradual thinning of the choroid from the periphery toward the staphyloma edge, followed by a rethickening toward the posterior pole. There was also a progressive thickening and inward protrusion of the sclera at the staphyloma border.^17,19 In the 3D-MRI images, a posterior staphyloma appeared as an outward protrusion of the posterior segment of the eye when viewed from the inferior, the nasal, and the back of the eye.¹⁹ The Optos images had pigmentary abnormalities or abnormal reflectance along the staphyloma edges.^19,20

Figure 1.

Diagnosis of posterior staphyloma in different patients. Abnormal findings (arrowheads) are observed along the edge of staphyloma from the nasal of the optic disk to temporal of the macula in the pseudocolor image (A) and fundus autofluorescence (B) in the Optos images. Ultra-widefield OCT image across the fovea (C: horizontal; D: vertical) shows 2 staphyloma edges, superior and inferior to the macula (arrows). A 3-dimensional magnetic resonance image shows a posterior outpouching (arrowheads) due to a staphyloma in the image viewed nasally (E) and posteriorly (F).

For the case-control comparisons, publicly available databases, including the 1000 Genomes Project phase III (1000GP3, East Asian population, approximately 500 samples, https://www.internationalgenome.org/home), the Exome Aggregation Consortium (ExAC, East Asian population, approximately 4000 samples, http://exac.broadinstitute.org), and the Japanese Multi-Omics Reference Panel (jMorp, approximately 40 000 Japanese samples, https://jmorp.megabank.tohoku.ac.jp/), were used. The allele frequencies (AFs) from these databases were referenced for the statistical analyses.

WES

Genomic DNAs were extracted from peripheral blood leukocytes with the QuickGene-610L DNA extraction kit (FUJIFILM Co). Whole exome sequencing was outsourced to AZENTA Life Sciences, where sequencing was performed on the NovaSeq 6000 platform using a 150 bp paired-end configuration, following the manufacturer's instructions (Illumina). The data processing pipeline included the following steps.^21,22 Sequencing reads from Fastq files were mapped to the human reference genome (GRCh38) using BWA-MEM (version 0.7.1) and converted from SAM to BAM using SAMtools (version 1.10). Polymerase chain reaction-derived duplicates were marked and removed using Picard (version 2.23.8). Base quality scores were recalibrated using GATK BaseRecalibrator (version 4.2.0.0) to identify potential misreads. Data reconfiguration was performed using GATK ApplyBQSR. Variants were called using GATK HaplotypeCaller to generate Variant Call Format files. The Variant Call Format file was separated into 2 files: single nucleotide variants and insertion/deletion (indel) files using GATK SelectVariants. Probable sequencing artifacts were filtered out using GATK VariantFiltration. The single nucleotide variant and indel files were merged into a single Variant Call Format file using Picard MergeVcfs. Variant annotations were generated using SnpEff.

Variant Filtering

Variant filtering was performed by the following steps: (1) Variants were filtered based on sequencing depth (≥10) and quality control criteria (filter label “passed”) using SnpSift.jar filter; (2) exclude in-house data as experimental errors; (3) variants with AFs of 0–0.1 (wild-type homozygous), 0.4–0.6 (heterozygous), and 0.9–1.0 (mutant homozygous) were retained, while variants detected in all subjects were excluded; (4) functional prediction scores and annotations were added using the database for nonsynonymous functional predictions (https://sites.google.com/site/jpopgen/dbNSFP); (5) only nonsynonymous variants were retained using SnpSift.jar filter; (6) single nucleotide variants with a minor allele frequency ≥0.01 in any of the following databases were excluded: 1000GP3, jMorp, Exome Sequencing Project, Human Genetic Variation Database, and Genome Aggregation Database; and (7) variants classified as “damaging” in ≥4 of 8 functional prediction algorithms (MutationAssessor, MutationTaster, PROVEAN, PolyPhen2_HDIV, PolyPhen2_HVAR, SIFT, FATHMM, and CADD) in the database for nonsynonymous functional predictions (https://sites.google.com/site/jpopgen/dbNSFP) were extracted. A schematic workflow of the variant filtering process is provided in Figure S1 (available at www.ophthalmologyscience.org).

Statistical Analyses

Fisher exact test was used to evaluate the differences in the AF between the study patients and individuals from the jMorp database, as well as East Asian populations in the 1000GP3 and ExAC databases. Odds ratios and 95% confidence intervals were calculated to assess the strength of associations. A P value ≤0.05 was considered statistically significant. All statistical analyses were performed using the R software (version 4.4.2; R Foundation for Statistical Computing).

Staphyloma-Related Gene Panel

A staphyloma-related gene panel composed of 388 genes was constructed based on 13 diseases reported to be associated with a staphyloma (Table S1, available at www.ophthalmologyscience.org).^9,12, 13, 14^,23, 24, 25, 26, 27, 28, 29, 30, 31 The panel consisted of genes implicated in these diseases. To identify potential staphyloma-associated genes, the panel was cross-referenced with genes identified through WES, and overlapping genes were extracted for further analysis.

Gene Set Enrichment Analysis

Gene set enrichment analysis was performed on genes that overlapped between the results of Fisher exact test and the staphyloma-related gene panel, using DAVID Knowledgebase (v2023q4).³² Gene set enrichment analysis was used to investigate the shared biological functions, domains, pathways, posttranslational modifications, and sequence features among these genes.

Sanger Sequencing

All candidate variants were validated using Sanger sequencing that was performed by Azenta Life Sciences. Primer sequences are listed in Table S2 (available at www.ophthalmologyscience.org). Variants that could not be conclusively detected by Sanger sequencing were subsequently validated by restriction fragment length polymorphism assay using XhoI with the same primers used for the polymerase chain reaction amplification.

Protein–Protein Interaction Network Analyses

To further explore the potential functional associations among the candidate genes identified by the enrichment analysis, protein–protein interactions were investigated using the STRING database (version 12.0; https://string-db.org/).

Results

General Information

We studied 453 eyes of 264 participants with 77.7% females. The mean age of the participants was 60.1 ± 11.5 years (range: 28–86 years). The mean axial length of the 264 patients was 29.68 ± 2.29 mm (range: 24.29–39.92 mm), and the mean refractive error (spherical equivalent) was –11.43 ± 6.99 D (range: –27.75 to –0.5 D), except for 5 eyes with intraocular lenses. According to the staphyloma classification,¹⁹ 247 eyes were categorized as the wide macular type, 148 as the narrow macular type, 4 as the nasal type, 8 as the peripapillary type, 13 as the inferior type, and 33 as the unclear type.

WES and Statistical Analyses

After WES and variants filtering, a total of 16 656 missense variants involving 8628 genes were identified in the 264 staphyloma patients. Among these, 5748 variants were found in 1000GP3, 10 998 in ExAC, and 13 736 in jMorp. Fisher exact test was used to compare the AF between the patient cohort and each database. A total of 141, 1,426, and 3120 variants in 1000GP3, ExAC, and jMorp, respectively, had a P < 0.05 and odds ratio >1. After filtering out variants with AF lower than any of the 3 databases, 3925 variants in 3138 genes remained (Table 1).

Table 1.

Summary of Variant Counts after Statistical Analysis and Low Frequency Filter

	1000Gp3 Database	ExAC Database	Jmorp Database	Significant Variants	Significant Genes
Total variant counts∗	5748	10 998	13 736
P < 0.05† and OR >1	141	1426	3120	4208	3305
Low Frequency Filter‡				3925	3138

1000Gp3 = 1000 Genomes Project Phase III; ExAC = Exome Aggregation Consortium; Jmorp = Japanese Multi Omics Reference Panel; OR = odds ratio.

^∗The number of variants identified in the study and found in each database.

^†Fisher exact test.

^‡Variants with a lower allele frequency in patients than in any of the 3 databases were excluded.

Staphyloma-Related Gene Panel

Within the staphyloma-related gene panel, 195 genes overlapped with the WES results. Of the 3138 significant genes identified through Fisher exact test and the 195 staphyloma-related genes from the gene panel, 81 genes overlapped (Table S3, available at www.ophthalmologyscience.org). These 81 overlapping genes were selected for gene set enrichment analyses.

Gene Set Enrichment Analyses

DAVID was used to identify functional gene groups and functional clusters enriched with pathogenic genes. Gene set enrichment analysis revealed significantly enriched gene function clusters (Table 2). Enrichment score of 1.3 is equivalent to the average of P values of the terms in cluster being 0.05. The count and percentage (%) in the table showed the number of genes that are involved in the annotation term and the ratio of genes related with the term to total genetic variants.

Table 2.

Significantly Enriched Gene Function Clusters from GSEA Results (Based on DAVID)

Enrichment Score: 2.80	Count	%	P Value	Genes List	Bonferroni	Benjamini	FDR
Term
GO:0005604∼Basement membrane	7	8.64	2.17E-06	COL18A1, COL2A1, SERPINF1, P3H2, COL4A5, COL9A3, USH2A	4.93E-04	1.23E-04	1.16E-04
KW-0272∼Extracellular matrix	10	12.35	2.21E-06	ADAMTSL1, FBN2, COL18A1, COL2A1, MYOC, COL11A1, ADAMTS18, COL4A5, COL9A3, LTBP2	6.63E-05	6.63E-05	6.40E-05
GO:0062023∼Collagen-containing extracellular matrix	10	12.35	1.81E-05	FBN2, COL18A1, COL2A1, MYOC, COL11A1, SERPINF1, COL4A5, COL9A3, LTBP2, AEBP1	0.004108	6.52E-04	6.12E-04
GO:0030020∼Extracellular matrix structural constituent conferring tensile strength	5	6.17	3.61E-05	COL18A1, COL2A1, COL11A1, COL4A5, COL9A3	0.01012	0.010171	0.010099
GO:0030198∼Extracellular matrix organization	7	8.64	5.89E-05	ADAMTSL1, COL18A1, COL2A1, COL11A1, ADAMTS18, COL4A5, COL9A3	0.034722	0.007068	0.006962
GO:0005788∼Endoplasmic reticulum lumen	8	9.88	1.91E-04	ADAMTSL1, COL18A1, COL2A1, COL11A1, P3H2, COL4A5, COL9A3, ARSG	0.04248	0.005405	0.005072
GO:0005581∼Collagen trimer	5	6.17	5.65E-04	COL18A1, COL2A1, COL11A1, COL4A5, COL9A3	0.120471	0.011667	0.010947
KW-0379∼Hydroxylation	6	7.41	0.003521	COL18A1, COL2A1, COL11A1, COL4A5, COL9A3, LTBP2	0.064812	0.035205	0.035205

FDR = false discovery rate; GO = gene ontology; GSEA = gene set enrichment analysis; KW = keyword.

Gene set enrichment analysis identified COL4A5, COL2A1, COL9A3, and COL18A1 were all significantly enriched in the following biological processes: basement membrane, extracellular matrix, endoplasmic reticulum lumen, collagen trimer formation, and hydroxylation pathways.

A significant variant in COL4A5 (Fisher P < 0.05) was identified in 27 patients in the WES results. Similarly, a single variant in COL2A1 and COL18A1 met this statistical threshold, and each was detected in 1 patient. In COL9A3, 2 distinct variants were also statistically significant, and each was found in 1 patient (Table 3).

Table 3.

Detailed Information of 5 Significant Variants in COL4A5, COL2A1, COL18A1, and COL9A3

CHROM	POS	rs ID	REF	ALT	Gene	Effect	Patient Count	Patient Frequency
chr23	108603032	rs104886164	C	G	COL4A5	missense	27	10.2%
chr12	47975553	rs1938662330	C	T	COL2A1	missense	1	0.38%
chr20	62817592	rs1390736361	G	A	COL9A3	missense	1	0.38%
chr20	62837209	rs140377811	G	A	COL9A3	missense	1	0.38%
chr21	45468350	.	C	A	COL18A1	missense	1	0.38%

ALT = alternate allele; CHROM = chromosome; POS = position; REF = reference allele; rs ID = reference SNP ID.

Among the 27 patients with the COL4A5 variants, female patients accounted for 96.3% (26/27) of the cases. Further genotype analyses revealed that among the 26 female patients, 23 were heterozygous and 3 were homozygous, while the remaining patient was a hemizygous male. These findings are consistent with an X-linked inheritance pattern.

To further characterize the phenotypes of these heterozygous females, the images of representative cases are shown in Figure 2 and Figure S2 (available at www.ophthalmologyscience.org). A comparison between staphyloma patients with and without the COL4A5 variant is shown in Table 4. A significant difference in the sex distribution was observed, with a higher proportion of females in the COL4A5 variant group. However, no significant differences were found between the 2 groups in terms of age, refractive error, and axial length.

Figure 2.

Representative ocular images of a 73-year-old heterozygous female carrier with the COL4A5 variant. Both eyes exhibit wide macular staphyloma. A–H, Right eye with a refractive error of –7.9 D and an axial length of 28.0 mm. I–P, Left eye with a refractive error of –4 D and an axial length of 26.2 mm. Pseudocolor images (A, I), fundus autofluorescence (B, J), and infrared images (C, K) acquired using Optos show abnormal findings along the staphyloma border (arrowheads). Ultra-widefield OCT image across the fovea (D, L: horizontal; E, M: vertical) reveals the staphyloma edges (arrows). Three-dimensional magnetic resonance imaging shows a posterior outpouching (arrowheads) due to a staphyloma in the image viewed nasally (F, N), inferiorly (G, O), and posteriorly (H, P). D = diopters.

Table 4.

Comparison between Staphyloma Patients with and without COL4A5 Variants

	With COL4A5 (n = 27)	Without COL4A5 (n = 237)	P Value
Gender, female (%)	26 (96.3%)	179 (75.5%)	0.027∗
Age (yrs, Median [IQR])	57.00 [48.00, 71.50]	62.00 [54.00, 68.00]	0.333†
Refractive error (D, median [IQR])	–11.06 [–15.02, –5.06]	–11.62 [–16.25, –5.88]	0.372†
Axial length (mm, median [IQR])	28.82 [27.11, 30.45]	29.51 [28.39, 30.78]	0.049†

D = diopters; IQR = interquartile range.

^∗Chi-squared test.

^†Mann–Whitney U test.

Validation of Candidate Variants

All candidate variants identified by WES were subsequently validated. Four variants were confirmed by Sanger sequencing. One variant in COL9A3 (rs140377811) could not be conclusively detected by Sanger sequencing due to high background noise and was therefore validated by restriction fragment length polymorphism using XhoI.

Discussion

COL4A5 Variant (c.1411C>G, p.Pro471Ala; rs104886164) as a Candidate Contributor to Posterior Staphyloma Development

The COL4A5 gene, located on the X chromosome, encodes type IV collagen, which is a major component of basement membranes and is critical for structural integrity and cellular signaling.³³ Basement membranes are specialized extracellular matrices found beneath epithelial and endothelial cells, where they maintain tissue architecture and function.³⁴

Pathogenic variants in the COL4A5 gene are well-recognized to cause X-linked Alport syndrome (AS), a hereditary basement membrane disorder characterized by progressive renal disease, hearing loss, and ocular abnormalities.³⁵ The common ocular characteristics include corneal opacities, anterior lenticonus, fleck retinopathies, and thinning of the temporal retina. These alterations are attributed to basement membrane fragility due to collagen IV deficiency in Descemet and Bowman membranes of the cornea, capsule of the lens, and the inner limiting membrane (ILM) and Bruch membrane of the retina.^33,36

Anterior lenticonus, a hallmark of X-linked AS, results from the protrusion of the crystalline lens through a thinned anterior capsule lacking sufficient type IV collagen due to COL4A5 mutations.³⁶ While the link between COL4A5 and anterior lenticonus is established, its association with posterior staphylomas has not been reported. To the best of our knowledge, staphylomas have not been documented in AS.

Based on these observations, we suggest that dysfunction of COL4A5 weakens the ILM and Bruch membrane, predisposing the posterior eyewall to biomechanical deformation and staphyloma formation. This is supported by the eyeIntegration v2.12 gene expression database, which showed high COL4A5 expression in ocular tissues including the cornea, retina, and retinal pigment epithelium (Fig 3). Similarly, Col4a3 knockout mice demonstrate reduced type IV collagen and thinning of the ILM and Bruch membrane, underscoring the essential role of collagen IV in these structures.^37,38

Figure 3.

Expression levels of COL4A5 in human ocular tissues and kidney based on the eyeIntegration v2.12 gene expression database. Expression data are shown for conjunctiva, cornea, kidney, optic nerve, retina, and RPE, demonstrating high COL4A5 expression in both renal and ocular tissues. CPM = counts per million; ESC = embryonic stem cells; iPSC = induced pluripotent stem cells; RGC = retinal ganglion cells; RPE = retinal pigment epithelium.

Although the pathogenesis of staphylomas remains undetermined, Jonas et al³⁹ proposed that Bruch membrane plays a role in their development. Local posterior scleral protrusions resembling staphylomas have been observed in conditions with Bruch membrane defects, such as ocular toxoplasmosis, intrachoroidal cavitation, and lacquer cracks (linear defect of Bruch membrane) in highly myopic eyes. These findings further implicate Bruch membrane in the development of staphylomas.40, 41, 42 Histological findings of ILM bridges over areas lacking Bruch membrane support the idea that the ILM also contributes to the stability of the posterior eyewall through greater resistance to retinal elongation-induced stress compared to other retinal layers.⁴³

Our identified variant, c.1411C>G, p.Pro471Ala, substitutes a proline with alanine. Proline is essential for undergoing hydroxylation to form hydroxyproline that stabilizes the triple-helical structure of collagen by interchain hydrogen bonding.⁴⁴ Proline hydroxylation occurs in the endoplasmic reticulum, and both processes were significantly enriched in our enrichment analysis (Fig 4). Disruption of this step may prevent the proper formation of the triple-helical structure of collagen, which can lead to instability and degradation of collagen IV networks.

Figure 4.

Schematic model of proline hydroxylation and collagen IV stabilization. Proline amino acids in collagen IV are hydroxylated to hydroxyproline in the endoplasmic reticulum lumen, a reaction catalyzed by prolyl hydroxylase with vitamin C and Fe²⁺ as cofactors. Hydroxyproline stabilizes the triple-helical structure of collagen IV, which forms heterotrimers composed of the α3α4α5 or α5α5α6 chains. A COL4A5 variant causing a proline-to-alanine substitution may impair this process, resulting in defective collagen assembly and weakened basement membranes.

Type IV collagen assembles into 3 distinct heterotrimeric networks: α1α1α2, α3α4α5, and α5α5α6. The α3α4α5 network, which incorporates the α5 chain encoded by COL4A5, predominates in the adult glomerulus (glomerular basement membrane), cochlea (stria vascularis), cornea (Descemet membrane and Bowman membrane), lens capsule, and retina (ILM and Bruch membrane).³³ The presence of the α3α4α5 network in both the ILM and Bruch membrane strongly supports our hypothesis. Therefore, these findings support a molecular model in which COL4A5 dysfunction impairs proline hydroxylation, disrupts type IV collagen triple-helix stability, and compromises the integrity of basement membranes. These changes have the potential of increasing the susceptibility of the posterior eyewall to biomechanical deformation and staphyloma formation. A schematic representation of the proposed pathogenic mechanism linking COL4A5 dysfunction to posterior staphyloma formation is shown in Figure 5.

Figure 5.

Proposed pathogenic mechanism linking the COL4A5 variant (c.1411C>G, p.Pro471Ala) to posterior staphyloma formation. This model illustrates the hypothesized sequence of events by which this COL4A5 variant may contribute to posterior staphyloma. Left: the variant causes type IV collagen deficiency that weakens the basement membranes of the inner limiting membrane and Bruch membrane. This weakening leads to a loss of structural support and localized posterior protrusion. Right: the variant impairs proline hydroxylation that disrupts the stability of the type IV collagen triple-helical structure and compromises the integrity of the basement membrane. The combined effects may increase the biomechanical susceptibility of the posterior eyewall to deformation, potentially resulting in posterior staphyloma. RPE = retinal pigment epithelium.

None of the 27 patients in our cohort carrying this variant had a clinical diagnosis of AS. Prior studies report that missense or small in-frame variants in COL4A5 tend to cause milder phenotypes than truncating mutations.⁴⁵ Consistent with these observations, all variants identified in this study were missense, which may have resulted in subclinical or mild manifestations insufficient for a clinical diagnosis of X-linked AS. Thus, the possibility of undetected mild or subclinical ocular or renal manifestations cannot be excluded.

There are several possible reasons why staphylomas have not been described in AS patients. First, conventional examinations such as slitlamp biomicroscope and ophthalmoscopy cannot detect posterior staphylomas, which require advanced imaging techniques such as 3D-MRI, UWF-OCT, and widefield fundus imaging. Second, AS is typically diagnosed early before myopic progression. Third, the number of genetically confirmed AS patients remains limited, and posterior ocular segment imaging is not routinely performed in these patients.

Although many pathogenic variants in COL4A5 have been linked to AS, our identified COL4A5 variant has not been associated with AS. It is possible that different COL4A5 mutations lead to different tissue-specific phenotypes, a phenomenon known as allelic heterogeneity. We suggest that this variant may predominantly affect ocular basement membranes, resulting in posterior staphylomas without having renal or auditory involvement. Among the 27 posterior staphyloma patients carrying this variant, 23 (85.2%) were heterozygous females, 3 (11.1%) were homozygous females, and 1 (3.7%) was a hemizygous male. None had overt renal or auditory symptoms or had been clinically diagnosed with AS. These findings raise the possibility that this COL4A5 variant contributes to posterior staphylomas independently of the classical AS phenotype. Further research is required to clarify its tissue-specific effects.

This predominant ocular presentation, which occurred primarily in heterozygous female carriers, is reminiscent of what has been observed in female carriers of RPGR mutations, who do not develop typical retinitis pigmentosa phenotypes but often present with high myopia and distinctive fundus autofluorescence patterns.^46,47 The high proportion of heterozygous females in our cohort suggests that certain COL4A5 variants may result in isolated ocular manifestations in female carriers. These findings raise the possibility that heterozygous carriers of X-linked mutations can present with nonclassical, tissue-specific phenotypes.

In RPGR-associated disease, gene therapy is being explored for female carriers, despite the absence of a fully developed retinitis pigmentosa phenotype.^46,47 Given that staphylomas and pathological myopia can similarly lead to severe vision reduction, gene therapy may also be considered in the future for female carriers of COL4A5 variants with ocular manifestations. Further studies are needed to clarify the tissue-specific effects and phenotypic variability of COL4A5 variants to guide future therapeutic strategies.

Variants in COL2A1, COL9A3, and COL18A1 May Be Associated with Staphylomas

In addition to COL4A5, our enrichment analysis identified COL2A1, COL9A3, and COL18A1 as candidate genes significantly enriched in biological processes related to basement membrane composition and collagen organization. Protein–protein interaction analysis using the STRING database showed that the proteins associated with the COL2A1, COL4A5, COL18A1, and COL9A3 genes form a closely connected network (Fig 6), supporting their potential cooperative roles in posterior staphyloma pathogenesis.

Figure 6.

Protein–protein interaction network of candidate genes. The network shows predicted and known functional associations among the COL2A1, COL4A5, COL18A1, and COL9A3, generated using the STRING database (version 12.0, https://string-db.org/). Nodes represent proteins; edges represent protein–protein associations. Species: Homo sapiens.

COL18A1 encodes type XVIII collagen, a key component of Bruch membrane and the basement membranes of the choroidal vessels.^48,49 Together with type IV collagen encoded by COL4A5, it contributes to the structural integrity of ocular basement membranes. COL2A1 encodes type II collagen, the main structural element of the vitreous membrane that is essential for maintaining the integrity of the vitreous body and providing structural support by forming the fibrillar network that attaches to the ILM.^50,51 COL9A3 encodes type IX collagen, which coats type II collagen fibrils to regulate fibril spacing and stabilize the vitreous gel. It also anchors the vitreous collagen network to the ILM that then contributes to the structural stability of the vitreoretinal interface.⁵¹

Together, COL4A5, COL18A1, COL2A1, and COL9A3 likely contribute to maintaining the structural and biomechanical integrity of the posterior eyewall through complementary mechanisms. COL4A5 plays a central role in stabilizing the ILM and Bruch membrane as a key component of type IV collagen networks. COL18A1 reinforces the Bruch membrane. COL2A1 and COL9A3 maintain the vitreous collagen network and anchor it to the ILM, thereby reinforcing the vitreoretinal interface and indirectly supporting the retina. Disruption of any of these components may weaken the structural resistance to the intraocular pressure, leading to localized protrusion and posterior staphyloma formation.

Elucidating these molecular mechanisms may help identify novel therapeutic targets for posterior staphyloma, such as gene therapies addressing basement-membrane or collagen abnormalities. Detecting these variants in myopic patients could also enable early identification of individuals at risk and facilitate genetics-based risk stratification. Moreover, genotype–phenotype correlations may help predict disease progression and guide personalized management in the future.

In addition to these genetic factors, environmental influences such as near-work activity, educational load, and limited outdoor exposure are known to accelerate myopia and high myopia progression.^1,3 These factors may further exacerbate posterior mechanical stress in eyes with collagen gene variants, potentially amplifying the risk of posterior staphyloma formation. The precise mechanisms underlying these gene–environment interactions warrant further investigation.

Limitations

This study has several limitations. First, the sample size was modest, with WES performed on 264 patients with posterior staphyloma. Second, the candidate variants identified by bioinformatic analysis were not functionally validated; in vitro and in vivo studies are needed to confirm their pathogenicity. Third, population bias and potential technical differences cannot be ruled out, as control AFs were derived from public databases. Fourth, the staphyloma-related gene panel was artificially constructed based on currently available literature and databases, which may have led to the omission of other real causative genes not included in the panel. Despite these limitations, our findings support the association of COL4A5, COL18A1, COL2A1, and COL9A3 variants with posterior staphyloma development.

Conclusions

Variants in the COL4A5, COL18A1, COL2A1, and COL9A3 genes are likely to contribute to the development of posterior staphylomas. Genetic disruptions of proline hydroxylation and collagen triple-helix stabilization may play important roles in the pathogenesis of posterior staphyloma. The ILM and Bruch membrane are likely to be the primary tissues involved in the formation of posterior staphylomas. The COL4A5 variant may be associated with an ocular-predominant phenotype without systemic features of AS, particularly in heterozygous female carriers.

Data Availability

The data that support the findings of this study are not publicly available due to ethical and privacy restrictions but are available from the corresponding author upon reasonable request.

Manuscript no. XOPS-D-25-00514.