The linkage disequilibrium plot of 5 LOXL1 single-nucleotide polymorphisms was calculated using Haploview. The D′ values are greater than 0.7 for all pairs.
Dubey SK, Hejtmancik JF, Krishnadas SR, Sharmila R, Haripriya A, Sundaresan P. Lysyl Oxidase–Like 1 Gene in the Reversal of Promoter Risk Allele in Pseudoexfoliation Syndrome. JAMA Ophthalmol. 2014;132(8):949-955. doi:10.1001/jamaophthalmol.2014.845
This study was necessary to establish the association between common genetic variants in the lysyl oxidase–like 1 (LOXL1) gene with pseudoexfoliation (PEX) syndrome and emphasize the reversal of promoter risk allele in a South Indian population.
To investigate the potential association of genetic variants across the LOXL1 gene in South Indian patients with PEX syndrome and glaucoma.
Design, Setting, and Participants
A case-control study of individuals from Madurai, India, with PEX syndrome and glaucoma as well as healthy people serving as controls. Three hundred unrelated people with PEX syndrome and 225 age- and ethnically matched controls were recruited for genetic analysis.
Main Outcomes and Measures
Four single-nucleotide polymorphisms in LOXL1 (rs16958477, rs1048661, rs3825942, and rs2165241) were genotyped by direct sequencing in all participants. Regulatory regions and 7 coding exons of LOXL1 were directly sequenced in 50 patients and 50 controls. A case-control association analysis was performed using the Golden Helix SVS suite.
An association between 4 LOXL1 single-nucleotide polymorphisms with PEX syndrome and glaucoma was observed (rs16958477, P = 4.77 × 10−6 [odds ratio, 0.50]; rs1048661, P = 4.28 × 10−5 [1.79]; rs3825942, P = 4.68 × 10−30 [9.19]; and rs2165241, P = 1.98 × 10−15 [2.88]). Sequencing of 7 exons and regulatory regions of LOXL1 identified 11 additional sequence variants; only rs41435250 showed an association (P = 3.80 × 10−5 [0.49]) with PEX syndrome and glaucoma.
Conclusions and Relevance
Genetic variants in LOXL1 are associated with PEX syndrome and glaucoma in the South Indian population. To our knowledge, this is the first study to demonstrate the association of rs41435250 with PEX as well as reversal of the promoter risk allele. Understanding the role of the LOXL1 gene in PEX pathogenesis will facilitate early detection in individuals at risk for this condition.
Pseudoexfoliation (PEX) syndrome is a generalized disorder of the extracellular matrix and an important risk factor for secondary open-angle glaucoma.1,2 The disease is characterized by deposition of abnormal microfibrillar aggregates on ocular tissues lining the aqueous-bathed surfaces of the anterior segment.2 The origin of PEX material is unclear, but it is believed to be multifocally produced by several ocular tissues, including the iris, lens epithelium, nonpigmented ciliary epithelium, trabecular meshwork, and corneal endothelium.3 In addition to the eye, extraocular tissues such as skin, lung, liver, cerebral meninges, kidney, blood vessels, and gallbladder exhibit the presence of the abnormal PEX material.4- 7 Pseudoexfoliation syndrome exhibits an age-related incidence and worldwide prevalence that varies markedly among different ethnic groups, from being virtually absent in Greenland Eskimos to affecting 20% to 25% of individuals in Iceland and Finland.8- 10 In a rural South Indian population older than 40 years, the prevalence of PEX syndrome is 6%.11 The cause of PEX syndrome and the molecular components of PEX materials are unclear, but several studies in recent years have shed light on the pathogenesis of PEX syndrome at genetic and protein levels. Thorleifsson et al12 performed a genome-wide association study in cases and controls from Icelandic and Swedish populations to identify the genetic factors associated with the increased risk of PEX syndrome. The study identified a strong association between PEX syndrome and glaucoma and 3 single-nucleotide polymorphisms (SNPs) in the lysyl oxidase–like 1 (LOXL1; OMIM:153456) gene. An extracellular copper-dependent monoamine oxidase, LOXL1 is involved in covalent cross-linking of collagen and elastin through oxidative deamination of lysine or hydroxylysine residues.13,14 Of the SNPs identified by Thorleifsson et al,12 2 mapped to exon1 (rs1048661 [R141L] and rs3825942 [G153D]), causing missense changes at the N-terminal part of the protein; the third SNP (rs2165241) mapped to the first intron of LOXL1. The study demonstrated that the 2 nonsynonymous SNPs of LOXL1 can account for more than 99% of all PEX cases (ie, syndrome and glaucoma).12 Subsequently, these findings were replicated in multiple populations globally, which established LOXL1 as a major genetic risk factor for PEX syndrome and glaucoma.15- 33
Although the LOXL1 SNPs showed consistent association among various ethnic groups, the risk-associated alleles and allele frequencies varied among different groups. The G allele of rs1048661 and T allele of rs2165241 increased risk in most studied populations, but Japanese, Chinese, and Korean individuals were exceptions, with the opposite alleles of both SNPs increasing the risk for PEX.23- 25,28,29 Similarly, the G allele of rs3825942 increased the risk in all studied ethnic groups except black South Africans, with the A allele conferring increased risk in that population.34,35 Thus, the relatively high prevalence of LOXL1 risk alleles in unaffected individuals, the variable association of specific alleles of these SNPs with PEX, and the lack of influence of 2 nonsynonymous variants (R141L and G153D) on amine oxidase activity of LOXL1 protein all suggest that these are not the causative variants.36 These variants are possibly in linkage disequilibrium with the functional variants or additional, as-yet unidentified, genetic and environmental factors might be involved in the development of PEX syndrome and glaucoma.
Recently, Fan et al37 demonstrated a strong association between 2 LOXL1 promoter SNPs (rs16958477, located at –659 base pairs upstream of the transcription start site, and rs12914489, approximately 30 kilobases upstream of the transcription start site) and PEX syndrome and glaucoma in the white population of the United States. The A allele of SNP rs16958477, identified as a risk allele in this population, was shown37,38 to be associated with reduced LOXL1 expression. However, this SNP did not demonstrate34 an association with PEX syndrome in a black South African population. The purpose of the present study was to evaluate the association of LOXL1 gene variants (rs16958477, rs1048661, rs3825942, and rs2165241) with PEX syndrome and glaucoma in a South Indian population. In addition, this study investigated the potential association of other genetic variants by exploring the regulatory regions (promoter and untranslated regions) and all coding regions across LOXL1.
The study protocol was approved by the institutional review board of Aravind Eye Hospital and was in accordance with the tenets of the Declaration of Helsinki. Written informed consent was obtained from all study participants after the nature of the study was explained. Three hundred unrelated individuals (PEX syndrome, 150; PEX glaucoma, 150) and 225 age- and ethnically matched controls were enrolled for this study.
All participants underwent a comprehensive ophthalmic examination, including visual acuity, slitlamp biomicroscopy, Goldmann applanation tonometry, gonioscopy, and retinal examination. Cases of PEX syndrome were diagnosed using clinical evidence of PEX material on the pupillary margin and/or anterior lens capsule, intraocular pressure less than 19 mm Hg, and no glaucomatous optic atrophy and visual field defects. Cases of PEX glaucoma were defined as the presence of PEX material, open iridocorneal angles, intraocular pressure 22 mm Hg or greater in either eye, typical glaucomatous optic disc changes (defined as thinning or notching of the neuroretinal rim, vertical cup-disc ratio >0.7, or cup-disc ratio asymmetry ≥0.3 between the eyes), and typical visual field defects in computed perimetry. Participants serving as controls had no evidence of PEX material, intraocular pressure less than 19 mm Hg, no glaucomatous changes in the optic disc (no thinning or notching of neuroretinal rim, cup-disc ratio range 0.3 to 0.5, or cup-disc ratio asymmetry <0.2 between the eyes), normal visual fields, and no personal or family history of glaucoma.
Peripheral blood samples (5 mL) were collected from each participant, and genomic DNA was isolated by the modified protocol of Miller et al.39 The entire coding region of LOXL1 along with the promoter region (848 base pairs upstream of transcription start site), 5′ and 3′ untranslated regions, and exon-intron boundaries were directly sequenced in 50 PEX (25 participants with PEX syndrome and 25 with PEX glaucoma) cases and 50 controls. Predesigned primers, as reported by Williams et al,34 were used for polymerase chain reaction (PCR) and sequencing. With the use of a gradient thermocycler (Astec), PCR was performed in a total volume of 20 μL containing 1× PCR buffer (10mM TRIS hydrochloride, pH 8.3; 50mM potassium chloride; 1.5mM magnesium chloride;, and 0.001% gelatin), 200 μM of deoxynucleotide triphosphate in solution (Medox Biotech India Pvt Ltd), 0.5 pmol of each primer, 100 ng of genomic DNA, and 1 U of Taq DNA polymerase (Sigma Aldrich). Thermal cycling conditions were 5 minutes at 95°C, followed by 35 cycles (30 seconds at 95°C, 30 seconds at the annealing temperature of the primers [53°C- 61°C], and 45 seconds at 72°C), and a final extension for 7 minutes at 72°C. The amplicons were gel purified using a spin-column DNA gel extraction kit (EZ-10; Bio Basic Inc). Direct sequencing was performed based on dideoxy termination chemistry (ABI 3130 genetic analyzer; Applied Biosystems). Sequence analysis was performed using Chromas software, version 2.33; Technelysium Pty Ltd; http://www.technelysium.com.au/chromas.html. Polymerase chain reaction amplicons harboring rs16958477, rs1048661, rs3825942, and rs41435250 were genotyped in all 525 study participants as described above, and genotyping of intronic SNP rs2165241 was performed by direct sequencing with a set of predesigned primers as reported earlier.40
Genotypic and allelic frequencies of PEX cases and controls were compared using χ2 and Fisher exact tests. Hardy-Weinberg equilibrium (HWE) for all SNPs was tested in the cases and controls using a Fisher exact test, all as implemented in Golden Helix SVS software, suite 7 (Golden Helix; http://www.goldenhelix.com). Odds ratios (ORs), 95% CIs, and call rates were calculated using the same program. Haploview41 was used to examine linkage disequilibrium, haplotype blocks, and HWE statistics. Haplotype association analysis was done using the same algorithms as integrated into the Golden Helix SVS suite and the expectation-maximization algorithm to estimate haplotype frequencies among cases and controls. Bonferroni correction was used for multiple testing, and P < .00045 was considered significant.
The mean (SD) age of the participants was 65.0 (7.0) years (range, 45-80 years) in PEX syndrome cases, 67.0 (7.6) years (range, 46-95 years) in PEX glaucoma cases, and 66.0 (6.2) years (range, 55-92 years) in controls. There were no significant differences detected among the mean ages of these 3 groups (P = .13, PEX syndrome vs control and P = .19, PEX glaucoma vs control).
The distribution of genotype and allele frequencies of 4 LOXL1 SNPs (rs16958477, rs1048661, rs3825942, and rs2165241) is reported in Table 1. The observed genotype frequencies were in HWE in control groups for all 4 SNPs and also in the cases in rs16958477, rs1048661, and rs2165241. However, there was a significant deviation from HWE in the observed genotype frequencies of rs3825942 in PEX glaucoma (P = 1.29 × 10−18) and PEX syndrome (P = 4.9 × 10−3) cases. This deviation might not be associated with the population structure because identical study participants were used for all 4 SNPs. Genotyping of rs3825942 by PCR-based restriction fragment-length polymorphism analysis using HinfI restriction enzyme showed no genotyping errors. Thus, deviation from the HWE of rs3825942 in affected individuals results from the low frequency of both homozygous and heterozygous A genotypes consistent with the strong association between this marker and PEX syndrome. This SNP also showed significant deviation from HWE (P = 2 × 10−5) in PEX cases in a previous study.35 Strong association of each SNP (rs16958477, rs1048661, rs3825942, and rs2165241) with both PEX syndrome and PEX glaucoma was observed in the South Indian cohort (Table 1). The G allele of rs1048661 (P = 4.28 × 10−5; OR, 1.79), G allele of rs3825942 (P = 4.68 × 10−30; OR, 9.19), and T allele of rs2165241 (P = 1.98 × 10−15; OR, 2.88) are the risk-associated alleles in this population (Table 1), which is similar to that found in the original study conducted in Icelandic and Swedish populations.12 The risk-associated allele of rs16958477 (promoter SNP) was different between the South Indian population and the white population of the United States. The A allele of rs16958477 was the risk allele in white individuals, whereas in Indian participants this allele showed a protective effect (P = 4.77 × 10−6; OR, 0.50).37
Sequencing of all 7 exons and regulatory regions of LOXL1 in 50 PEX cases and 50 controls identified 11 additional sequence variants (Table 2). Of these, only rs41435250 showed a significant difference in allele frequencies between the cases and controls (P < .001), whereas the other sequence variants were rare and not associated with PEX. The SNP rs41435250 was further genotyped with a 100% success rate in all 525 study participants and showed strong association (Table 1) with both PEX syndrome (P = 7.69 × 10−4; OR, 0.51) and PEX glaucoma (P = 1.45 × 10−4; OR, 0.48).
Closer examination of the genotype and model-specific risks showed that, in general, LOXL1 region alleles that increase risk tend to have a codominant to recessive inheritance pattern (Table 3). The most dramatic example of this was rs2165241, for which the OR for CT heterozygotes was 2.22 (95% CI, 0.24-20.48), but for TT homozygotes the OR was 8.82 (4.69-16.57), with corresponding ORs and P values for the dominant (OR, 3.12; P = 5.02 × 10−2) and recessive (OR, 5.83; P = 2.67 × 10−10) models.
Haplotype analysis using the 5 associated markers across the LOXL1 region, rs16958477-rs1048661-rs3825942-rs41435250-rs2165241, confirmed and extended the results obtained with individual SNPs (Table 4). The entire group of markers was in moderate to strong linkage equilibrium, as seen by D′ values greater than 0.7 for all pairs (Figure). Haplotypes for the 5 markers were strongly associated with PEX (PEX syndrome and glaucoma combined), with the AGAGC haplotype being strongly protective (OR, 0.10; 95% CI, 0.06-0.16; P = 6.03 × 10−30) and the AGGGT haplotype showing increased risk (OR, 2.45; 1.69-3.54; P = 1.29 × 10−6). The AGGGC haplotype had an even higher risk (OR, 4.07; 1.37-12.09), but because of the small number of individuals with this haplotype, the finding did not reach significance after Bonferroni correction (P = 6.35 × 10−3).
The LOX family of proteins comprises 5 members: the prototypic LOX protein and the 4 LOX-like proteins (LOXL1-LOXL4). The LOXL1 protein makes covalent cross-links in the elastin monomers through oxidative deamination of lysine side chains to form mature elastin polymer fibers. Genes of all the LOX family members have 7 exons, of which exons 2 to 6 show high-sequence similarity and encode the C-terminal catalytic domain of proteins. Differences among LOX family members reside mainly in exon 1, which shows high variation in size and sequence and encodes the N-terminal propeptide necessary for enzyme activation, substrate recognition, and proper substrate binding to LOX. The 2 LOXL1 missense variants (R141L and G153D) identified by Thorleifsson et al12 are located in the N-terminal domain. These variants are hypothesized to affect either the catalytic activity of enzymes or binding of substrates, such as fibulin-5 and tropoelastin, to LOXL1. However, many recent studies23- 25,28,29,34,35 have reported the reversal of risk alleles of R141L and G153D in different ethnic populations. In addition, these 2 variants have no effect on the amine oxidase activity of LOXL1 protein, as reported recently.36 These discrepant genetic findings strongly suggest that R141L and G153D might not be causative but might rather be in linkage disequilibrium with the functional variants.
In this study, we have demonstrated the genetic association of LOXL1 with PEX syndrome and PEX glaucoma in a South Indian population. One SNP in the regulatory region (rs16958477 in the promoter), 3 SNPs in the coding region (rs1048661, rs3825942, and rs41435250), and 1 intronic SNP (rs2165241) showed an association with PEX syndrome in this population. The G allele of rs1048661, the G allele of rs3825942, and the T allele of rs2165241 are the risk alleles in the South Indian population, which is consistent with the original study conducted by Thorleifsson et al.12 We report a significant association of rs1048661 (P = 4.28 × 10−5; OR, 1.79) with PEX in the South Indian population, which contrasts with the findings of a previous study conducted by Ramprasad et al30 in a similar South Indian population. Ramprasad and colleagues30 did not find an association between rs1048661 and PEX (P = .15) in their study, most likely because of a small sample size.
The G allele of rs3825942 is the most strongly risk-associated allele (P = 4.68 × 10−30; OR, 9.19) in the South Indian population. To date, all replication studies15- 33 in multiple ethnic populations have demonstrated a strong association of LOXL1 with PEX syndrome and the G allele of rs3825942 has been considered the high-risk allele in most of the populations. However, in some populations, different alleles of LOXL1 SNPs are associated with an increased risk of PEX. For instance, the T allele of rs1048661 in Chinese, Japanese, and Korean individuals23- 25,28,29; the G allele of rs1048661 in Turkish people42; the T allele of rs2165241 in the Latin American population43; and the A allele of rs3825942 in black South Africans34,35 are the high-risk alleles. Our study has also established the association of the promoter SNP rs16958477 with PEX syndrome and PEX glaucoma in the South Indian population. This SNP showed a significant association with PEX in the white population of the United States but not in black South Africans.34,37 The A allele of SNP rs16958477 that was identified as a risk allele (P = 6.4 × 10−7; OR, 2.05) in white individuals was also associated with reduced LOXL1 promoter activity, as reported previously.38 An uncommon phenomenon observed in the present study was that the A allele of rs16958477 showed a protective effect (P = 4.77 × 10−6; OR, 0.50) in the South Indian cohort. This result is consistent with the recent genetic finding23- 25,28,29,34,35 on the reversal of risk alleles of 3 LOXL1 SNPs (rs1048661, rs3825942, and rs2165241). This unique observation suggests that rs16958477 might not be causative in PEX abnormalities and that the true functional variants are yet to be identified.
Sequencing of the coding and regulatory regions of LOXL1 in 50 cases and 50 controls identified 4 known SNPs (rs41435250, rs74026313, rs8818, and rs3588) and 7 novel variants (g.62,425C>T, g.62,544C>T, V111E, P333P, E398A, g.79,399G>C, and D456D). Among these changes, only rs41435250 showed a significant association (P < .001) with PEX; the other sequence variants were rare and did not show any association (Table 2). When genotyped in all 525 study participants, rs41435250, a synonymous (A320A) SNP, showed a strong association with PEX (P = 3.80 × 10−5; OR, 0.49), and the G allele of this SNP showed a protective effect. The frequency of the risk allele T of rs41435250 was relatively uncommon in individuals serving as controls (12.4%) in comparison with cases (22.3%). To our knowledge, this is the first study showing an association between rs41435250 and PEX syndrome. The SNP rs41435250 extends the list of genetic markers that may be useful in identification of individuals predisposed to this condition. However, elucidating the possible functional role of this SNP in the development of PEX will be challenging, and additional replication studies are warranted to identify the risk allele of this SNP in a variety of populations.
The LOXL1 SNPs are highly associated with PEX syndrome in the South Indian population. To our knowledge, this is the first study to demonstrate the association of rs41435250 with PEX as well as the reversal of the promoter risk allele in our population. Replication studies of LOXL1 SNPs in multiple populations, as well as the strength of the association of these SNPs with PEX, have convincingly shown that LOXL1 is a major genetic risk factor for the development of PEX syndrome and glaucoma. These LOXL1 SNPs hold promise as potential markers for identification of individuals who are at increased risk of developing PEX syndrome and glaucoma. However, high frequencies of risk alleles in the general population, reduced disease penetrance in some populations regardless of the high frequency of risk alleles, reversal of the risk alleles of associated SNPs, and a lack of demonstrated effect of the R141L and G153D changes on amine oxidase activity of the LOXL1 protein suggest that other, as yet unknown, factors might be involved in the development of PEX syndrome. Resequencing of the entire LOXL1 gene region, including regulatory and intronic regions, is essential to understand the involvement of this gene in the pathogenesis of PEX syndrome.
Submitted for Publication: October 14, 2013; final revision received January 9, 2014; accepted January 17, 2014.
Corresponding Author: Periasamy Sundaresan, PhD, Department of Genetics, Aravind Medical Research Foundation, Dr G. Venkataswamy Eye Research Institute, Aravind Eye Hospital, 1, Anna Nagar, Madurai 625020, Tamil Nadu, India (email@example.com).
Published Online: May 8, 2014. doi:10.1001/jamaophthalmol.2014.845.
Author Contributions: Dr Sundaresan had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: Dubey, Hejtmancik, Sharmila, Haripriya, Sundaresan.
Acquisition, analysis, or interpretation of data: Dubey, Hejtmancik, Krishnadas, Haripriya.
Drafting of the manuscript: Dubey, Sharmila, Sundaresan.
Critical revision of the manuscript for important intellectual content: Dubey, Hejtmancik, Krishnadas, Haripriya.
Statistical analysis: Dubey, Hejtmancik, Haripriya.
Obtained funding: Sundaresan.
Administrative, technical, or material support: Dubey, Sharmila, Sundaresan.
Study supervision: Krishnadas, Sundaresan.
Conflict of Interest Disclosures: None reported.
Funding/Support: This study was supported by a research grant from Alcon-Aravind Eye Care System, India.
Role of the Sponsor: The Alcon-Aravind Eye Care System had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.
Additional Contributions: We thank all the patients and healthy individuals for participating in this study.