ABSTRACT
Background
To investigate single‐nucleotide polymorphisms (SNPs) reported in the largest up‐to‐date systematic review and meta‐analysis on primary angle‐closure disease (PACD), on their associations with primary angle‐closure glaucoma (PACG) and disease progression.
Methods
This study involved a case–control design for PACG risk and a case‐only design for PACG progression risk, including 628 PACG patients and 564 controls for disease association and 386 PACG patients with up to 10‐year follow‐up for PACG progression analysis. Associations of 17 SNPs in 15 genes with PACG were analysed using logistic regression. Sex‐stratified association analysis was performed, followed by the Breslow‐Day test. Genetic risk for PACG progression was evaluated using logistic regression. Bonferroni correction of p values was adopted for multiple comparisons.
Results
LOXL1 rs3825942 (G153D; p = 0.0026; OR = 0.65) was significantly associated with PACG, while ABCC5 rs1401999 showed a nominal association (p = 0.023; OR = 1.32). ABCA1 rs2422493 was significantly associated with PACG in females (p = 0.0016; OR = 0.70) but not in males (p = 0.95; OR = 0.99); and the Breslow‐Day Test (p = 0.046) suggested a sex‐specific association in females. VAV3 rs6689476 showed nominal associations with PACG progression at 3‐year (p = 0.045; OR = 2.86), 5‐year (p = 0.037; OR = 2.84) and 10‐year follow‐ups (p = 0.03; OR = 2.74), but the p values could not withstand Bonferroni correction.
Conclusion
This study demonstrated a role of LOXL1 in PACG and a sex‐specific effect of ABCA1 in the Hong Kong Chinese population while suggesting a potential role of VAV3 in PACG progression, which has yet to be further confirmed.
Keywords: candidate gene investigation, disease progression, genetics, primary angle‐closure glaucoma
1. Introduction
Glaucoma is a leading cause of irreversible blindness worldwide [1]. Based on the anatomical appearance of the anterior chamber angle, glaucoma can be classified into open‐angle glaucoma and angle‐closure glaucoma. Primary angle‐closure glaucoma (PACG) reportedly affected approximately 17.14 million individuals globally in 2020 [2], and the number was projected to double by 2040 [3].
PACG is a multifactorial disease, with elevated intraocular pressure (IOP), usually resulting from a closed anterior chamber angle, being a major and modifiable risk factor. Other risk factors include older age, female sex, hyperopia, family history and ethnicity (e.g., African and East Asians) [1, 4].
Genetic factors also contribute to PACG. Genome‐wide association studies (GWAS) have identified 8 susceptible loci associated with PACG, namely EPDR1, CHAT, CLIS3, FERMT2, DPM2‐FAM102A, PLEKHA7, COL11A1 and PCMTD1‐ST18 [5, 6]. These associations were further investigated in different study cohorts, which revealed variable association profiles across populations [7]. In a recent systematic review and meta‐analysis of common single‐nucleotide polymorphisms (SNPs) and rare coding variants, 15 SNPs in 13 genes/loci were found to be associated with PACG and two additional SNPs in two genes/loci with primary angle‐closure disease (PACD) – a continuum spectrum of phenotypes including primary angle closure suspect (PACS), PAC and PACG [8]. Among these 17 SNPs, 13 of them had not been reported in PACG GWAS. From the systematic review, it was noted that most genetic investigations did not evaluate sex‐specific genetic effects in PACG [8]. Moreover, there was only one genetic study of PACG progression reporting negative results [9]. Therefore, whether these 15 PACG‐associated genes/loci also play a role in PACG progression remained to be investigated.
In this study, we studied the associations of the 17 SNPs in 15 genes/loci with PACG in a Hong Kong Chinese cohort, evaluated the sex‐stratified effects of the SNPs and explored the roles of these SNPs in PACG disease progression.
2. Methods
2.1. Patient Recruitment
The study protocol was approved by the Ethics Committee for Human Research at the Chinese University of Hong Kong. All procedures adhered to the principles of the Declaration of Helsinki and the International Council on Harmonisation–Good Clinical Practice (ICH‐GCP) guidelines. Written informed consent was obtained from each participant. We prospectively recruited patients from the eye clinics of the Chinese University of Hong Kong Eye Centre and Hong Kong Eye Hospital between September 2009 and October 2023 [10, 11, 12].
All recruited patients had PACG in at least one eye at baseline and they fulfilled the diagnostic criteria from the International Society of Geographical and Epidemiological Ophthalmology (ISGEO) [11, 13]. Any patients with angle closure and/or ocular hypertension from the following causes were excluded: uveitis, neovascularisation of iris/angle, iris/ciliary body cysts, posterior segment haemorrhage or tumour, Axenfeld‐Rieger syndrome, trauma, steroid‐induced and/or iatrogenic. Exfoliation syndrome/glaucoma had been ruled out in all patients.
Healthy control subjects were recruited from individuals attending eye clinics for unrelated eye conditions or eye check‐ups. They had undergone complete ocular assessments and were confirmed to have no glaucoma or other ocular diseases, except for dry eyes, mild cataracts and/or mild refractive errors. All control subjects had open anterior chamber angles (Shaffer grade 3 or 4 anterior chamber angle on gonioscopy) and IOP lower than 21 mmHg. Individuals with a family history of glaucoma were excluded from controls [11]. A total of 628 unrelated PACG patients and 564 healthy controls were included in the cross‐sectional genetic investigation on PACG risk. These subjects were independent of the patients included in the previous GWAS of PACG [6].
Among the 628 PACG patients, we had consent from 386 to participate in the longitudinal investigation. In longitudinal progression analysis, all patients received a complete ophthalmic examination at the time of recruitment, including visual acuity measurement, ocular tonometry for baseline intraocular pressure (IOP), slit‐lamp biomicroscopy, auto‐refraction, optical coherence tomography (OCT) for peripapillary retinal nerve fibre layer (RNFL) thickness and measurement of central corneal thickness (CCT) and axial length (AL). Serial visual field tests were performed every 6 months. A visual field result was considered unreliable and excluded when fixation losses > 33%, false positive rate > 25% and/or false negative rate > 25%. Only reliable tests were included in further analysis. A poor‐quality visual field result was also excluded if there was evidence of inattention fixation, eyelid or lens rim artefacts, fatigue effects and abnormal results caused by factors other than glaucoma [11]. We included those patients who had undergone standard glaucoma treatment and had been followed up by glaucoma specialists between September 2009 and October 2023. If a patient had both eyes affected by glaucoma at baseline, the worse eye, determined by baseline mean deviation (MD), was selected for the longitudinal analysis for PACG progression.
According to the preferred practice pattern of PACG, glaucoma progression included both visual function progression and structural progression [14]. The purpose of follow‐up examinations was to evaluate IOP level, visual field status, optic disc appearance, RNFL and macular imaging to determine if progressive damage has occurred [14]. Therefore, it was ideal if we could classify PACG patients into progressive or non‐progressive based on all these parameters [14]. However, instead of providing a catch‐all algorithm for PACG progression, we focused on longitudinal visual field progression to measure disease progression and analysed the genetic association. Referring to a previous genetic investigation on PACG severity [15], we divided the PACG patients into different severity subgroups, ranging from early (MD ≥ −6 dB), moderate (−6 dB > MD ≥ −12 dB), advanced (−12 dB > MD ≥ −20 dB), severe (MD < −20 dB) and end stage (the worse eye became unable to perform Humphrey visual field test due to glaucoma) [16, 17]. A patient was defined as having glaucoma progression if the worse eye had developed into a more advanced glaucomatous stage at any follow‐up visit compared to that at the first recruitment (baseline).
2.2. SNP Selection and Genotyping
Totally 17 SNPs in 15 genes were selected from the recent systematic review and meta‐analysis [8], including ABCA1 rs2422493, ABCC5 rs1401999, ATOH7 rs1900004, CALCRL rs9288141, CHAT rs1258267, COL11A1 rs3753841, FBN1 rs363830, HGF rs5745718 and rs3735520, IL6 rs1800796, LOXL1 rs3825942, MMP19 rs2291267, PCMTD1‐ST18 rs1015213, PLEKHA7 rs11024060 and rs11024102, VAV3 rs6689476 and ZNRF3 rs3178915. The SNPs selected from the same gene were not in strong linkage disequilibrium (LD; i.e., r 2 < 0.8) in the Chinese population [18].
Genomic DNA was extracted from whole blood using the Qiagen QIAamp DNA Blood Midi Kit (Qiagen, Hilden, Germany). The candidate SNPs were genotyped in all subjects using TaqMan genotyping assays (Applied Biosystems, Foster City, CA) on a Roche Light Cycler 480 Real‐Time PCR System (Roche, Basel, Switzerland) [19, 20, 21, 22]. Since there was no commercially available TaqMan assay for COL11A1 rs3753841, we selected COL11A1 rs1932350, which is in strong LD (r 2 = 0.86) with rs3753841 in Chinese, as a replacement [18].
2.3. Statistical Analysis
The Hardy–Weinberg equilibrium (HWE) of the 17 candidate SNPs was assessed in healthy control subjects using PLINK (version 1.9). PCMTD1‐ST18 rs1015213 deviated from HWE in controls and was excluded from subsequent association analysis. As the primary outcome, we first investigated the associations of 16 candidate SNPs with PACG in a case–control design, using logistic regression with age and sex adjusted. Odds ratio (OR) and 95% confidence interval (CI) were estimated. We adopted the Bonferroni correction for multiple testing. For the primary outcome, 16 SNPs were involved; therefore, a p value of < 0.0031 (0.05/16) was used to define a study‐wide significant genetic association.
Since female sex is a risk factor of PACG [2], as a secondary outcome, we stratified the study participants by sex and performed association analyses for PACG with age adjustment in logistic regression to explore potential sex‐specific genetic associations. PCMTD1‐ST18 rs1015213 deviated from HWE in female and male controls, ABCC5 rs1401999 deviated from HWE in female controls and HGF rs5745718 deviated from HWE in male controls. They were excluded from the association analyses in respective strata. Therefore, 30 comparisons were made in both males and females, and a p < 0.0017 (0.05/30) was adopted to define a study‐wide significant association in each sex group. Heterogeneity in the ORs between males and females was assessed by the Breslow‐Day test, where a p < 0.05 suggested a sex‐specific genetic association.
To explore the genetic associations with PACG progression, we first compared the SNP distribution between patients who progressed by the 3‐year/5‐year/10‐year cut‐offs and those who remained stable (non‐progressive) over the 10 years, using logistic regression with age and sex adjusted. Patients remaining stable for 10 years served as a stringent control group, increasing the detection power with a fixed cohort size [23]. p < 0.05 indicated a nominal association, while a p value of < 0.001 (0.05/16*3) indicated a study‐wide significant association.
3. Results
3.1. Demographics of Study Participants
This study involved 628 unrelated PACG patients, aged 31–94 years (median, 71; IQR, 14) and 564 healthy controls, aged 40–99 years (median, 71; IQR, 14; Table 1). There was no significant difference in age or sex between the two groups. Table 2 present demographics of 386 subjects recruited for the longitudinal investigation, including 69 mild PACG (aged 47–91 years, median, 69; IQR, 15.5), 100 moderate PACG (aged 36–87 years, median, 73; IQR, 13.5) and 217 severe PACG patients (aged 31–91 years, median, 72; IQR, 15). There was no significant difference in age or sex among the mild, moderate and severe PACG subgroups at study recruitment. By 31 October 2023, we had a collection of 350, 259 and 113 PACG patients who had completed the follow‐up visits by 3 years, 5 years and 10 years, respectively.
TABLE 1.
Demographics of subjects in the case–control analysis.
| PACG patients | Controls | |
|---|---|---|
| N (male/female) | 244/384 | 238/326 |
| Age* (median/IQR) | 71/14 | 71/14 |
| Age* | 31–94 | 40–99 |
Abbreviations: IQR, Interquartile range; PACG, Primary angle‐closure glaucoma.
Data collected at recruitment.
TABLE 2.
Demographics of subjects participating in longitudinal analysis.
| Mild glaucoma | Moderate glaucoma | Severe glaucoma | |
|---|---|---|---|
| N (male/female) | 21/48 | 39/61 | 90/127 |
| Age* (median/IQR) | 69/15.5 | 73/13.5 | 72/15 |
| Age* | 47–91 | 36–87 | 31–91 |
Abbreviation: IQR, Interquartile range.
Data collected at recruitment.
Table 3 shows the demographics and glaucoma‐related parameters of individuals analysed for glaucoma progression. By the 10‐year follow‐up visit, we had collected data from 113 PACG patients. Among them, the worse eye of 73 patients remained stable in 10 years. These patients were taken as stringent controls for the genetic association analysis of PACG progression. Among the 40 patients who had glaucoma progression, 34 manifested within the first 3 years, 3 within the fourth to fifth years and 3 after the fifth year.
TABLE 3.
Demographics of patients with and without glaucoma progression in longitudinal follow‐ups.
| 3 years | 5 years | 10 years | ||||
|---|---|---|---|---|---|---|
| Progression | No progression | Progression | No progression | Progression | No progression | |
| N (male/female) | 15/19 | 120/196 | 16/21 | 79/143 | 18/22 | 15/58 |
| Age* (median/IQR) | 68.5/13 | 70/13 | 68/13.5 | 70/13 | 67/15.5 | 66/14 |
| Age range* | 35–87 | 31–91 | 35–87 | 31–91 | 35–87 | 31–82 |
| Vertical cup‐to‐disc ratio* | 0.62 ± 0.25 | 0.65 ± 0.23 | 0.61 ± 0.25 | 0.66 ± 0.22 | 0.60 ± 0.25 | 0.67 ± 0.18 |
| MD* (mean ± SD) | −9.32 ± 4.88 | −15.36 ± 9.63 | −9.13 ± 4.80 | −14.23 ± 9.22 | −8.70 ± 4.94 | −13.29 ± 8.87 |
Abbreviation: IQR: Interquartile range.
Data collected at study recruitment.
3.2. Genetic Association With PACG
The genotype call rate was greater than 99% for all SNPs. After adjusting for sex and age, rs3825942 in the LOXL1 gene showed a significant association with PACG, with the A allele conferring a protective effect (OR = 0.65, p = 0.0026). Another two SNPs, ABCA1 rs2422493 (A allele: OR = 0.81, p = 0.011) and ABCC5 rs1401999 (C allele: OR = 1.32, p = 0.023) showed nominal associations (Table 4).
TABLE 4.
Genetic association of primary angle‐closure glaucoma.
| Gene | SNP | A1/A2 a | OR (95% CI) | p b | Reference OR c |
|---|---|---|---|---|---|
| ABCA1 | rs2422493 | A/G | 0.81 (0.68, 0.95) | 0.011 | 0.93 |
| ABCC5 | rs1401999 | C/G | 1.32 (1.04, 1.68) | 0.023 | 1.08 |
| ATOH7 | rs1900004 | T/C | 1.08 (0.91, 1.28) | 0.39 | 1.10 |
| CALCRL | rs9288141 | G/A | 1.00 (0.81, 1.24) | 0.99 | 1.14 |
| CHAT | rs1258267 | G/A | 0.86 (0.71, 1.04) | 0.12 | 0.75 |
| COL11A1 | rs1932350 | C/T | 1.12 (0.94, 1.33) | 0.21 | 1.16 |
| FBN1 | rs363830 | T/C | 1.11 (0.81, 1.52) | 0.53 | 0.76 |
| HGF | rs3735520 | A/G | 0.86 (0.73, 1.01) | 0.072 | 1.07 |
| HGF | rs5745718 | T/C | 1.17 (0.94, 1.45) | 0.16 | 1.51 |
| IL6 | rs1800796 | G/C | 1.14 (0.94, 1.39) | 0.18 | 0.83 |
| LOXL1 | rs3825942 | A/G | 0.65 (0.49, 0.86) | 0.0026 | 0.88 |
| MMP19 | rs2291267 | A/G | 1.13 (0.88, 1.46) | 0.35 | 1.20 |
| PLEKHA7 | rs11024060 | T/G | 1.03 (0.87, 1.22) | 0.74 | 0.93 |
| PLEKHA7 | rs11024102 | C/T | 1.12 (0.95, 1.32) | 0.18 | 1.19 |
| VAV3 | rs6689476 | C/T | 1.31 (0.99, 1.71) | 0.057 | 1.22 |
| ZNRF3 | rs3178915 | A/G | 1.03 (0.87, 1.21) | 0.72 | 1.07 |
Abbreviations: ABCA1, ATP binding cassette subfamily A member 1; ABCC5, ATP binding cassette subfamily C member 5; ATOH7, atonal bHLH transcription factor 7; CALCRL, calcitonin receptor like receptor; CHAT, choline O‐acetyltransferase; COL11A1, collagen type XI alpha 1 chain; DPM2‐FAM102A, DPM2 stands for dolichyl‐phosphate mannosyltransferase subunit 2‐FAM102A, family with sequence similarity 102 member A, DPM2‐FAM102A indicates the SNP locates within these two genes; FBN1, fibrillin 1; IL6, interleukin 6; LOXL1, lysyl oxidase like 1; MMP19, matrix metallopeptidase 19; PCMTD1‐ST18, PCMTD1‐protein‐L‐isoaspartate (D‐aspartate) O‐methyltransferase domain containing 1; PCMTD1‐ST18 suggests the SNP located in the intergenic region of these genes; PLEKHA7, pleckstrin homology domain containing A7; ST18, ST18 C2H2C‐type zinc finger transcription factor; VAV3, vav guanine nucleotide exchange factor 3; ZNRF3, zinc and ring finger 3.
A1/A2, minor allele/major allele, minor allele is the effect allele for the odds ratio.
p < 0.05, indicating nominal association, p < 0.05/16 = 0.0031, indicating study‐wise significant association.
Reference OR, the effect size of the effect allele referring to the latest meta‐analysis [10].
In the stratification analysis by sex, ABCA1 rs2422493 showed a significant association with PACG in females (A allele: OR = 0.70, p = 0.0016), but not in males (OR = 0.99, p = 0.95; Table 5). The Breslow‐Day test confirmed the significant heterogeneity between the two ORs (p = 0.046; Table 5). The Breslow‐Day test also identified another SNP, CHAT rs1258267, showing marginal heterogeneity between females and males (p = 0.049). The G allele had a nominal association with PACG in females (OR = 0.70, p = 0.019) but not in males (OR = 1.09, p = 0.58; Table 5).
TABLE 5.
Stratification analysis of the genetic association of primary angle‐closure glaucoma by sex.
| Gene | SNP | A1/A2 a | Female | Male | Breslow‐day test (P) c | ||
|---|---|---|---|---|---|---|---|
| Adjusted OR | Adjusted p b | Adjusted OR | Adjusted p b | ||||
| ABCA1 | rs2422493 | A/G | 0.70 (0.60, 0.90) | 0.0016 | 0.99 (0.76, 1.29) | 0.95 | 0.046 |
| ABCC5 | rs1401999 | C/G | Deviated from HWE | 1.25 (0.85, 1.85) | 0.26 | 0.78 | |
| ATOH7 | rs1900004 | T/C | 1.20 (0.90, 1.50) | 0.16 | 0.94 (0.72, 1.23) | 0.65 | 0.22 |
| CALCRL | rs9288141 | G/A | 1.00 (0.80, 1.30) | 0.87 | 0.97 (0.70, 1.36) | 0.88 | 0.84 |
| CHAT | rs1258267 | G/A | 0.70 (0.60, 1.00) | 0.019 | 1.09 (0.80, 1.49) | 0.58 | 0.049 |
| COL11A1 | rs1932350 | C/T | 1.10 (0.90, 1.30) | 0.55 | 1.21 (0.92, 1.59) | 0.17 | 0.50 |
| FBN1 | rs363830 | T/C | 1.20 (0.80, 1.80) | 0.36 | 0.97 (0.57, 1.66) | 0.91 | 0.52 |
| HGF | rs3735520 | A/G | 0.90 (0.70, 1.10) | 0.15 | 0.88 (0.67, 1.15) | 0.34 | 0.76 |
| HGF | rs5745718 | T/C | 1.10 (0.80, 1.50) | 0.44 | Deviated from HWE | 0.81 | |
| IL6 | rs1800796 | G/C | 1.10 (0.80, 1.40) | 0.63 | 1.26 (0.94, 1.69) | 0.13 | 0.35 |
| LOXL1 | rs3825942 | A/G | 0.80 (0.60, 1.10) | 0.23 | 0.47 (0.29, 0.75) | 0.0014 | 0.076 |
| MMP19 | rs2291267 | A/G | 0.90 (0.70, 1.30) | 0.70 | 1.48 (1.00, 2.21) | 0.05 | 0.082 |
| PLEKHA7 | rs11024060 | T/G | 1.00 (0.80, 1.30) | 0.95 | 1.05 (0.80, 1.37) | 0.75 | 0.48 |
| PLEKHA7 | rs11024102 | C/T | 1.10 (0.90, 1.30) | 0.50 | 1.20 (0.93, 1.56) | 0.16 | 0.87 |
| VAV3 | rs6689476 | C/T | 1.20 (0.80, 1.70) | 0.38 | 1.51 (1.00, 2.27) | 0.049 | 0.35 |
| ZNRF3 | rs3178915 | A/G | 1.10 (0.90, 1.40) | 0.38 | 0.93 (0.72, 1.21) | 0.59 | 0.34 |
Abbreviations: ABCA1, ATP binding cassette subfamily A member 1; ABCC5, ATP binding cassette subfamily C member 5; ATOH7, atonal bHLH transcription factor 7; CALCRL, calcitonin receptor like receptor; CHAT, choline O‐acetyltransferase; COL11A1, collagen type XI alpha 1 chain; DPM2‐FAM102A, DPM2 stands for dolichyl‐phosphate mannosyltransferase subunit 2‐FAM102A stands for family with sequence similarity 102 member A, DPM2‐FAM102A indicates the SNP locates within these two genes; FBN1, fibrillin 1; IL6, interleukin 6; LOXL1, lysyl oxidase like 1; MMP19, matrix metallopeptidase 19; PCMTD1‐ST18, PCMTD1‐protein‐L‐isoaspartate (D‐aspartate) O‐methyltransferase domain containing 1; PCMTD1‐ST18 suggests the SNP located in the intergenic region of these genes; PLEKHA7, pleckstrin homology domain containing A7; ST18, ST18 C2H2C‐type zinc finger transcription factor; VAV3, vav guanine nucleotide exchange factor 3; ZNRF3, zinc and ring finger 3.
A1/A2, minor allele/major allele, minor allele is the effect allele for the odds ratio.
p < 0.05, indicating nominal association; p < 0.05/46 = 0.0011, indicating study‐wise significant association.
p < 0.05 in the Breslow‐Day test, suggesting a potential sex‐specific genetic effect.
3.3. Genetic Associations With PACG Progression
The VAV3 rs6689476 showed borderline associations with PACG progression (C allele: OR = 2.86, p = 0.045 at 3‐year; OR = 2.84, p = 0.037 at 5‐year; and OR = 2.74, p = 0.03 at 10‐year; Table 6). This SNP had a borderline association with PACG (OR = 1.31, p = 0.057; Table 4), with no sex‐specific effect, although it showed a nominal association with PACG in females (OR = 1.51, p = 0.049; Breslow‐Day test, p = 0.35; Table 5). PLEKHA7 rs11024102 also showed a borderline association with PACG progression at 3 years (OR = 0.50, p = 0.040; Table 6).
TABLE 6.
Genetic association of primary angle‐closure glaucoma progression.
| Gene | SNP | A1/A2 a | 3 years | 5 years | 10 years | Disease risk b | |||
|---|---|---|---|---|---|---|---|---|---|
| OR (95% CI) | p c | OR (95% CI) | p c | OR (95% CI) | p c | OR (95% CI) | |||
| ABCA1 | rs2422493 | A/G | 1.18 (0.60, 2.34) | 0.63 | 1.23 (0.65, 2.34) | 0.53 | 1.10 (0.59, 2.06) | 0.76 | 0.81 (0.68, 0.95) |
| ABCC5 | rs1401999 | C/G | 0.65 (0.28, 1.49) | 0.31 | 0.72 (0.32, 1.61) | 0.42 | 0.77 (0.35, 1.67) | 0.50 | 1.32 (1.04, 1.68) |
| ATOH7 | rs1900004 | T/C | 0.90 (0.48, 1.69) | 0.73 | 0.95 (0.52, 1.77) | 0.88 | 1.01 (0.55, 1.86) | 0.96 | 1.08 (0.91, 1.28) |
| CALCRL | rs9288141 | G/A | 0.73 (0.31, 1.72) | 0.48 | 0.90 (0.40, 2.02) | 0.80 | 0.89 (0.40, 1.95) | 0.76 | 1.00 (0.81, 1.24) |
| CHAT | rs1258267 | G/A | 0.65 (0.25, 1.66) | 0.37 | 0.79 (0.34, 1.87) | 0.60 | 1.03 (0.48, 2.24) | 0.93 | 0.86 (0.71, 1.04) |
| COL11A1 | rs1932350 | C/T | 1.02 (0.52, 2.01) | 0.95 | 1.26 (0.66, 2.40) | 0.49 | 1.23 (0.65, 2.31) | 0.53 | 1.12 (0.94, 1.33) |
| FBN1 | rs363830 | T/C | 1.28 (0.43, 3.79) | 0.66 | 1.27 (0.43, 3.73) | 0.67 | 1.17 (0.40, 3.46) | 0.77 | 1.11 (0.81, 1.52) |
| HGF | rs3735520 | A/G | 0.84 (0.43, 1.65) | 0.61 | 0.80 (0.43, 1.49) | 0.48 | 0.81 (0.44, 1.49) | 0.50 | 0.86 (0.73, 1.01) |
| HGF | rs5745718 | T/C | 1.15 (0.54, 2.47) | 0.72 | 1.17 (0.55, 2.47) | 0.69 | 1.16 (0.55, 2.43) | 0.70 | 1.17 (0.94, 1.45) |
| IL6 | rs1800796 | G/C | 1.44 (0.67, 3.09) | 0.35 | 1.28 (0.61, 2.68) | 0.51 | 1.22 (0.59, 2.52) | 0.59 | 1.14 (0.94, 1.39) |
| LOXL1 | rs3825942 | A/G | 0.93 (0.31, 2.8) | 0.90 | 1.11 (0.39, 3.14) | 0.85 | 1.25 (0.46, 3.44) | 0.66 | 0.65 (0.49, 0.86) |
| MMP19 | rs2291267 | A/G | 1.09 (0.38, 3.16) | 0.87 | 0.99 (0.35, 2.84) | 0.99 | 1.04 (0.38, 2.84) | 0.95 | 1.13 (0.88, 1.46) |
| PLEKHA7 | rs11024060 | T/G | 1.28 (0.64, 2.56) | 0.49 | 1.39 (0.72, 2.70) | 0.32 | 1.51 (0.78, 2.89) | 0.22 | 1.03 (0.87, 1.22) |
| PLEKHA7 | rs11024102 | C/T | 0.50 (0.26, 0.97) | 0.040 | 0.58 (0.32, 1.08) | 0.088 | 0.68 (0.38, 1.21) | 0.19 | 1.12 (0.95, 1.32) |
| VAV3 | rs6689476 | C/T | 2.86 (1.02, 7.98) | 0.045 | 2.84 (1.07, 7.54) | 0.037 | 2.74 (1.08, 6.95) | 0.030 | 1.31 (0.99, 1.71) |
| ZNRF3 | rs3178915 | A/G | 1.05 (0.58, 1.91) | 0.87 | 1.20 (0.68, 2.12) | 0.53 | 1.21 (0.69, 2.13) | 0.50 | 1.03 (0.87, 1.21) |
Abbreviations: ABCA1, ATP binding cassette subfamily A member 1; ABCC5, ATP binding cassette subfamily C member 5; ATOH7, atonal bHLH transcription factor 7; CALCRL, calcitonin receptor like receptor; CHAT, choline O‐acetyltransferase; COL11A1, collagen type XI alpha 1 chain; DPM2‐FAM102A, DPM2 stands for dolichyl‐phosphate mannosyltransferase subunit 2, FAM102A stands for family with sequence similarity 102 member A, DPM2‐FAM102A indicates the SNP locates within these two genes; FBN1, fibrillin 1; IL6, interleukin 6; LOXL1, lysyl oxidase like 1; MMP19, matrix metallopeptidase 19; PCMTD1‐ST18, PCMTD1‐protein‐L‐isoaspartate (D‐aspartate) O‐methyltransferase domain containing 1; PCMTD1‐ST18 suggests the SNP located in the intergenic region of these genes; PLEKHA7, pleckstrin homology domain containing A7; ST18, ST18 C2H2C‐type zinc finger transcription factor; VAV3, vav guanine nucleotide exchange factor 3; ZNRF3, zinc and ring finger 3.
A1/A2, minor allele/major allele, minor allele is the effect allele for the odds ratio.
References from genetic associations study on PACG disease risk (Table 4).
p < 0.05, indicating nominal association.
4. Discussion
In this study, non‐overlapping genetic markers were identified for PACG and disease progression. The missense variant LOXL1 rs3825942 showed a significant association and an intronic variant rs1401999 in ABCC5 showed a nominal association with PACG in the Hong Kong Chinese population. ABCA1 rs2422493 and CHAT rs1258267 were associated with PACG in females but not in males, suggesting possible sex‐specific effects. In an exploratory investigation on PACG progression, no statistically significant association was detected, although VAV3 rs6689476 showed borderline associations with PACG progression over the 3‐year, 5‐year and 10‐year follow‐ups; further confirmation should be warranted.
The allele A of LOXL1 rs3825942 showed a protective effect for PACG (OR = 0.65, p = 0.0026; Table 4) in our cohort. Of note, this was the first report on the association of LOXL1 rs3825942 with PACG in Chinese, and its protective effect aligns with the pooled results (OR = 0.88; Table 4) from an admixed population across Asian (Indian) and European populations (UK Biobank, FinnGen) [8]. While the allele A in rs3825942 was well recognised as a protective factor for exfoliation syndrome (XFS) and exfoliation glaucoma (XFG) in multiple populations [24, 25], whether the same SNP was the main allele in LOXL1 having a protective effect on PACG still warranted exploration. Since this is the first candidate SNP investigation on the role of LOXL1 rs3825942 on PACG risk in the Chinese population, further haplotype‐tagging SNP investigation should be warranted to elucidate the role of LOXL1 in PACG architecture.
In this study, we have, for the first time, demonstrated the exclusive association of ABCA1 rs2422493 with PACG in females. ABCA1 is a key regulator of cholesterol efflux. Previous genetic studies have indicated sex‐specific associations of ABCA1 with lipoproteins metabolism [26]. The allele C of ABCA1 rs2275543 was associated with the increase of high‐density lipoprotein size in females, while in males it was associated with lower very‐low‐density lipoprotein, cholesterol and triglycerides [26]. In coronary heart disease with disrupted cholesterol homeostasis, the ABCA1 R219 allele independently conferred a risk of coronary heart disease in Turkish females, but not in males [27]. While these coding variants affected the protein properties and conferred disease risk, altering ABCA1 expression also affected retinal ganglion cell survival via regulating cholesterol homeostasis [28]. Though being located in the intronic region, rs2422493 was demonstrated to regulate the ABCA1 expression level in multiple tissues, including muscle and adipose (Figure S1) [29]. Moreover, ABCA1 was also associated with POAG [30], suggesting the involvement of altered cholesterol homeostasis in the pathogenesis of both POAG and PACG. Therefore, the association between ABCA1 and PACG, particularly its sex‐specific effect, should be further investigated in more populations. Meanwhile, our results also suggested a nominal sex‐specific effect of CHAT rs1258267 in Chinese. Of note, the heterogeneity of CHAT rs1258267 was marginal and awaited further validation in larger samples. Nonetheless, it underscored the necessity of considering sex‐stratification analysis when calculating the genetic risk score [15].
Apart from LOXL1 rs3825942, ABCA1 rs2422493 and CHAT rs1258267, the effect alleles of the other eight tested SNPs, namely ABCC5 rs1401999, ATOH7 rs1900004, COL11A1 rs1932350, HGF rs5745718, MMP19 rs2291267, PLEKHA7 rs11024102, VAV3 rs6689476 and ZNRF3 rs3178915, showed consistent directions of effect on PACG when compared with the combined effect reported in the recent systematic review and meta‐analysis [8]. The remaining five SNPs, FBN1 rs363830, CALCRL rs9288141, HGF rs3735520, IL6 rs1800796 and PLEKHA7 rs11024060, showed opposite directions of effect with PACG in the Hong Kong cohort. These inconsistencies may be attributed to population heterogeneities, as the associations between PACG and these SNPs reported in the meta‐analysis were observed in admixed populations dominated by European populations.
In the exploratory analysis on PACG progression, though some borderline associations were captured, they could not withstand the Bonferroni correction, leaving no statistically significant association for PACG progression. There are several possibilities for the explanation: (1) these candidate SNPs selected based on their association with PACG in meta‐analysis may not be associated with PACG progression, suggesting the non‐overlapping genetic components between PACG risk and progression; (2) the genetic effects might be masked by treatment effect, since the patients recruited in the longitudinal investigation were well‐treated by our glaucoma experts and had their IOP well‐controlled; (3) the limited statistical power to capture genetic association with PACG progression in view of the smaller sample size remaining after a long‐term follow‐up. Further longitudinal studies of PACG progression in a larger cohort should be warranted to investigate its genetic associations.
There are some limitations in this study. First, the candidate SNPs were selected from the associated SNPs for PACD in an admixed population reported in a systematic review and meta‐analysis. It was difficult to take discovery data from a cohort (results in the meta‐analysis) that was not made up of our ancestry of interest to validate (Hong Kong Chinese). This means that the set of 15 SNPs chosen could not completely represent the genetic factors for PACG in this ancestry group. Choosing Southern Chinese specific SNPs could be more applicable to investigate reported findings in this cohort, but it is beyond the scope of this study. Second, analysing the genetic components for disease progression in a well‐treated cohort limited the penetrance of genetic associations. Third, our progression analysis was based on the change of MD, which is a global parameter in the visual field test providing an overall picture of average glaucomatous damage. However, it could lack alignment with localised spatial information related to disease progression [31]. Despite these limitations, our study design offered some insights into PACG progression assessment.
In summary, this study highlighted LOXL1 rs3825942 as a genetic marker for PACG in the Chinese population and revealed a sex‐specific association between ABCA1 rs2422493 and PACG in females. No statistically significant genetic association of PACG progression was captured, although VAV3 rs6689476 showed borderline associations. These findings enriched the growing understanding of the genetic architecture of PACG and provided potentially useful genetic markers for monitoring PACG development and progression. Further investigations, especially in different ethnic groups with long‐term follow‐up of patients, should be warranted.
Conflicts of Interest
The authors declare no conflicts of interest.
Supporting information
Figure S1. Cis‐eQTLs associated with ABCA1 rs2422493 generated by FIVEx.1
Liang Y. J., Ling A., Chan P. P., et al., “Genetic Association of Primary Angle‐Closure Glaucoma and Disease Progression,” Clinical & Experimental Ophthalmology 53, no. 6 (2025): 660–667, 10.1111/ceo.14539.
Funding: This work was supported by the research grants from the Health and Medical Research Fund Hong Kong 07180256 (L.J.C), General Research Fund, Hong Kong 14102122 and 14105320 (C.C.T.); 14105916 and 2141022 (C.P.P.); and the Endowment Fund for Lim Por‐Yen Eye Genetics Research Centre, Hong Kong.
Contributor Information
Clement C. Tham, Email: clemtham@cuhk.edu.hk.
Li Jia Chen, Email: lijia_chen@cuhk.edu.hk.
Data Availability Statement
The data that support the findings of this study are available from the corresponding author upon reasonable request.
References
- 1. Jonas J. B., Aung T., Bourne R. R., Bron A. M., Ritch R., and Panda‐Jonas S., “Glaucoma,” Lancet 390, no. 10108 (November 2017): 2183–2193. [DOI] [PubMed] [Google Scholar]
- 2. Zhang N., Wang J., Chen B., Li Y., and Jiang B., “Prevalence of Primary Angle Closure Glaucoma in the Last 20 Years: A Meta‐Analysis and Systematic Review,” Frontiers in Medicine 7 (2020): 624179. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Tham Y. C., Li X., Wong T. Y., Quigley H. A., Aung T., and Cheng C. Y., “Global Prevalence of Glaucoma and Projections of Glaucoma Burden Through 2040: A Systematic Review and Meta‐Analysis,” Ophthalmology 121, no. 11 (2014): 2081–2090. [DOI] [PubMed] [Google Scholar]
- 4. Ahram D. F., Alward W. L., and Kuehn M. H., “The Genetic Mechanisms of Primary Angle Closure Glaucoma,” Eye (London, England) 29, no. 10 (2015): 1251–1259. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Vithana E. N., Khor C. C., Qiao C., et al., “Genome‐Wide Association Analyses Identify Three New Susceptibility Loci for Primary Angle Closure Glaucoma,” Nature Genetics 44, no. 10 (October 2012): 1142–1146. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Khor C. C., Do T., Jia H., et al., “Genome‐Wide Association Study Identifies Five New Susceptibility Loci for Primary Angle Closure Glaucoma,” Nature Genetics May 2016; 48, no. 5 (2016): 556–562, 10.1038/ng.3540. [DOI] [PubMed] [Google Scholar]
- 7. Rong S. S., Tang F. Y., Chu W. K., et al., “Genetic Associations of Primary Angle‐Closure Disease: A Systematic Review and Meta‐Analysis,” Ophthalmology 123, no. 6 (2016): 1211–1221, 10.1016/j.ophtha.2015.12.027. [DOI] [PubMed] [Google Scholar]
- 8. Liang Y. J., Wang Y. Y., Rong S. S., et al., “Genetic Associations of Primary Angle‐Closure Disease: A Systematic Review and Meta‐Analysis,” JAMA Ophthalmology 142, no. 5 (May 2024): 437–444. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Thangavelu L., Che Mat Nor S. M., Abd Aziz D., Sulong S., Tin A., and Ahmad Tajudin L. S., “Genetic Markers PLEKHA7, ABCC5, and KALRN Are Not Associated With the Progression of Primary Angle Closure Glaucoma (PACG) in Malays,” Cureus 13, no. 10 (2021): e18823. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Tan S., Yu M., Baig N., et al., “Association of Ultra‐Short‐Term Intraocular Pressure Fluctuation With Disease Progression in Primary Angle Closure Glaucoma: The CUPAL Study,” Journal of Glaucoma 31, no. 11 (November 2022): 874–880. [DOI] [PubMed] [Google Scholar]
- 11. Tang F. Y., Ma L., Tam P. O. S., Pang C. P., Tham C. C., and Chen L. J., “Genetic Association of the PARL‐ABCC5‐HTR3D‐HTR3C Locus With Primary Angle‐Closure Glaucoma in Chinese,” Investigative Ophthalmology and Visual Science 58, no. 10 (August 2017): 4384–4389. [DOI] [PubMed] [Google Scholar]
- 12. Tan S., Yu M., Baig N., Chan P. P., Tang F. Y., and Tham C. C., “Circadian Intraocular Pressure Fluctuation and Disease Progression in Primary Angle Closure Glaucoma,” Investigative Ophthalmology and Visual Science 56, no. 8 (2015): 4994–5005. [DOI] [PubMed] [Google Scholar]
- 13. Foster P. J., Buhrmann R., Quigley H. A., and Johnson G. J., “The Definition and Classification of Glaucoma in Prevalence Surveys,” British Journal of Ophthalmology 86, no. 2 (2002): 238–242. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Gedde S. J., Chen P. P., Muir K. W., et al., “Primary Angle‐Closure Disease Preferred Practice Pattern,” Ophthalmology 128, no. 1 (2021): P30–P70. [DOI] [PubMed] [Google Scholar]
- 15. Liu C., Nongpiur M. E., Cheng C. Y., et al., “Evaluation of Primary Angle‐Closure Glaucoma Susceptibility Loci for Estimating Angle Closure Disease Severity,” Ophthalmology 128, no. 3 (2021): 403–409, 10.1016/j.ophtha.2020.07.027. [DOI] [PubMed] [Google Scholar]
- 16. Brusini P. and Johnson C. A., “Staging Functional Damage in Glaucoma: Review of Different Classification Methods,” Survey of Ophthalmology 52, no. 2 (2007): 156–179. [DOI] [PubMed] [Google Scholar]
- 17. Mills R. P., Budenz D. L., Lee P. P., et al., “Categorizing the Stage of Glaucoma From Pre‐Diagnosis to End‐Stage Disease,” American Journal of Ophthalmology 141, no. 1 (2006): 24–30, 10.1016/j.ajo.2005.07.044. [DOI] [PubMed] [Google Scholar]
- 18. Auton A., Abecasis G. R., Altshuler D. M., et al., “A Global Reference for Human Genetic Variation,” Nature 526, no. 7571 (2015): 68–74. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Lu S. Y., Rong S. S., Wu Z., et al., “Association of the CAV1‐CAV2 Locus With Normal‐Tension Glaucoma in Chinese and Japanese,” Clinical and Experimental Ophthalmology 48, no. 5 (2020): 658–665, 10.1111/ceo.13744. [DOI] [PubMed] [Google Scholar]
- 20. Chen Z. J., Ng D. S. C., Cen L. P., et al., “Multi‐Polymorphism Analysis Reveals Joint Effects in Males With Chronic Central Serous Chorioretinopathy,” Investigative Ophthalmology and Visual Science 64, no. 4 (April 2023): 19. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Chen Z. J., Ma L., Brelen M. E., et al., “Identification of TIE2 as a Susceptibility Gene for Neovascular Age‐Related Macular Degeneration and Polypoidal Choroidal Vasculopathy,” British Journal of Ophthalmology 105, no. 7 (2021): 1035–1040, 10.1136/bjophthalmol-2019-315746. [DOI] [PubMed] [Google Scholar]
- 22. Chen Z. J., Ng D. S., Ho M., et al., “Genetic Associations of Central Serous Chorioretinopathy Subtypes, Neovascular Age‐Related Macular Degeneration, and Polypoidal Choroidal Vasculopathy,” Asia‐Pacific Journal of Ophthalmology (Philadelphia, Pa.) 13, no. 1 (2024): 100003, 10.1016/j.apjo.2023.100003. [DOI] [PubMed] [Google Scholar]
- 23. Garagnani P., Giuliani C., Pirazzini C., et al., “Centenarians as Super‐Controls to Assess the Biological Relevance of Genetic Risk Factors for Common Age‐Related Diseases: A Proof of Principle on Type 2 Diabetes,” Aging (Albany NY) 5, no. 5 (2013): 373–385, 10.18632/aging.100562. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Li X., He J., and Sun J., “LOXL1 Gene Polymorphisms Are Associated With Exfoliation Syndrome/Exfoliation Glaucoma Risk: An Updated Meta‐Analysis,” PLoS One 16, no. 4 (2021): e0250772. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Aung T., Ozaki M., Mizoguchi T., et al., “A Common Variant Mapping to CACNA1A Is Associated With Susceptibility to Exfoliation Syndrome,” Nature Genetics 47, no. 4 (2015): 387–392, 10.1038/ng.3226. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Zago V. H. S., Scherrer D. Z., Parra E. S., Vieira I. C., Marson F. A. L., and de Faria E. C., “Effects of SNVs in ABCA1, ABCG1, ABCG5, ABCG8, and SCARB1 Genes on Plasma Lipids, Lipoproteins, and Adiposity Markers in a Brazilian Population,” Biochemical Genetics 60, no. 2 (2022): 822–841. [DOI] [PubMed] [Google Scholar]
- 27. Çoban N., Onat A., Kömürcü Bayrak E., Güleç Ç., Can G., and Erginel Ünaltuna N., “Gender Specific Association of ABCA1 Gene R219K Variant in Coronary Disease Risk Through Interactions With Serum Triglyceride Elevation in Turkish Adults,” Anadolu Kardiyoloji Dergisi 14, no. 1 (2014): 18–25. [DOI] [PubMed] [Google Scholar]
- 28. Yang J., Chen Y., Zou T., et al., “Cholesterol Homeostasis Regulated by ABCA1 Is Critical for Retinal Ganglion Cell Survival,” Science China. Life Sciences 66, no. 2 (2023): 211–225, 10.1007/s11427-021-2126-2. [DOI] [PubMed] [Google Scholar]
- 29. Kwong A., Boughton A. P., Wang M., et al., “FIVEx: An Interactive eQTL Browser Across Public Datasets,” Bioinformatics 38, no. 2 (January 2022): 559–561. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30. Chen Y., Lin Y., Vithana E. N., et al., “Common Variants Near ABCA1 and in PMM2 Are Associated With Primary Open‐Angle Glaucoma,” Nature Genetics 46, no. 10 (2014): 1115–1119, 10.1038/ng.3078. [DOI] [PubMed] [Google Scholar]
- 31. Asman P., Heijl A., Olsson J., and Rootzén H., “Spatial Analyses of Glaucomatous Visual Fields; a Comparison With Traditional Visual Field Indices,” Acta Ophthalmologica 70, no. 5 (1992): 679–686. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Figure S1. Cis‐eQTLs associated with ABCA1 rs2422493 generated by FIVEx.1
Data Availability Statement
The data that support the findings of this study are available from the corresponding author upon reasonable request.