The present knowledge on the association of single nucleotide polymorphisms (SNPs) of lysyl oxidase-like 1 (LOXL1) with pseudoexfoliation syndrome (PEXS) and pseudoexfoliation glaucoma (PEXG) is controversial and inconclusive. This meta-analysis sought to derive a more precise estimation of the effects of LOXL1 SNP loci (rs1048661, rs3825942, and rs2165241) on PEXS/PEXG. Literature searches were conducted on the PubMed, EMBASE, ISI Web of Science, and Cochrane Library databases through October 2013. Twelve studies describing 1810 cases and 1790 controls met the inclusion criteria. The strengths of the associations found through the meta-analysis were assessed with pooled odds ratios and their 95% confidence intervals (CI). A meta-regression analysis was also used to examine the influence of the study and population characteristics. The results indicated that rs1048661 TT carriers had 92.1% and 40.4% less risk of developing PEXS/PEXG than did the controls in the Caucasian and Asian populations, respectively. Carriers of rs3825942 AA or rs2165241 CC also had significantly less PEXS/PEXG susceptibility than did the non-carriers. Meta-regression showed that in Caucasians, the male proportion (slope: 0.272; 95% CI: 0.167–0.376; P = 0.0001) and mean age (slope: 0.796; 95% CI: 0.375–1.217; P = 0.0002) of the PEXS/PEXG subjects correlated positively with the effect of rs3825942 on PEXS/PEXG susceptibility. The meta-analysis suggested that LOXL1 rs1048661 TT, rs3825942 AA, and rs2165241 CC were associated with a reduced risk of developing PEXS/PEXG.
Citation: Tang J-Z, Wang X-Q, Ma F-F, Wang B, Wang P-F, Peng Z-X, et al. (2014) Association between Polymorphisms in Lysyl Oxidase-Like 1 and Susceptibility to Pseudoexfoliation Syndrome and Pseudoexfoliation Glaucoma. PLoS ONE 9(3): e90331. https://doi.org/10.1371/journal.pone.0090331
Editor: Emiliano Giardina, Tor Vergata University of Rome, Italy
Received: November 27, 2013; Accepted: January 28, 2014; Published: March 6, 2014
Copyright: © 2014 Tang et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: This work was supported by National Natural Science Foundation of China (81170858). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Pseudoexfoliation syndrome (PEXS) is an age-related systemic disorder that is the most common cause of secondary glaucoma worldwide and the most frequent cause of unilateral glaucoma , . In addition, PEXS progresses to pseudoexfoliation glaucoma (PEXG), which responds poorly to medical therapy in comparison to other types of glaucoma and can lead to the rapid progression of optic nerve damage .
Despite its worldwide distribution, there is a clear tendency for PEXS to cluster geographically and in certain racial or ethnic subgroups . Furthermore, PEXS has a strong familial association . The underlying causes of the different prevalence rates between age-matched geographical and ethnic populations remain unknown, but appear to be related to variation in genetic background , . Lysyl oxidase-like 1 (LOXL1) is a member of the lysyl oxidase gene family and is essential to the biogenesis of connective tissue ; it encodes an extracellular copper-dependent amine oxidase that catalyzes the first step in the formation of crosslinks in collagens and elastin , . A highly conserved amino acid sequence at the C-terminus appears to be sufficient for amine oxidase activity, suggesting that all family members may retain this function. A fibrillar, proteinaceous substance, is produced in abnormally high concentrations within the ocular tissues of patients with PEXG, and LOXL1 may be relevant to its formation , .
A genome-wide association study identified three common single nucleotide polymorphisms (SNPs) in the LOXL1 gene on chromosome 15q24.1 that were strongly associated with pseudoexfoliation syndrome . The LOXL1 polymorphisms included one intronic SNP, rs2165241, located within the first intron, and two non-synonymous coding SNPs, rs1048661 and rs3825942, located within the first exon. The association of LOXL1 with PEXS/PEXG has recently been confirmed in several different populations , , , , . However, the reported associations are controversial and inconclusive due to factors including the limited sample sizes,different ethnicities, and genotyping procedures , . Believing a meta-analysis would provide more credible evidence by systematically summarizing the existing data, we gathered eligible studies to investigate the association between the LOXL1 gene polymorphisms and susceptibility to PEXS/PEXG.
Materials and Methods
The PubMed, EMBASE, ISI Web of Science, and the Cochrane Library databases were electronically searched for case-control studies published through October 2013 that examined the association of the LOXL1 gene polymorphisms with the PEXS/PEXG susceptibility. The search strategy was based on a combination of “(lysyl oxidase-like 1 OR LOXL1) AND (gene OR variants OR polymorphism OR alleles OR mutation) AND (pseudoexfoliation syndrome OR pseudoexfoliation glaucoma)”. We also manually searched references in key articles. The language was limited to English.
The inclusion criteria were as follows: (a) evaluation of the association of the LOXL1 gene polymorphisms with the PEXS/PEXG risk; (b) case-control studies; (c) sufficient published data for estimating an odds ratio (OR) with a 95% confidence interval (CI); and (d) PEXS was diagnosed as the presence of pseudoexfoliative material on the anterior lens capsule after maximal pupil dilatation. PEXG was diagnosed if patients had typical features of PXFS and all of the following: an initial intraocular pressure of at least 22 mm Hg, glaucomatous optic disc changes, visual field defects consistentent with optic nerve damage, and no evidence of other conditions causing secondary glaucoma .The study authors were contacted for supplemental data if the information was not available in the publication. The studies with overlapping patient samples were excluded, and only the studies with the larger numbers of patients were included. Two of the authors independently identified and reviewed each relevant paper, and disagreements were reconciled through group discussion.
To show the relationship between the LOXL1 gene polymorphisms and PEXS/PEXG risk, the most strongly and independently associated SNPs were selected for the analysis (rs1048661, rs3825942, and rs2165241). Information regarding the following aspects was retrieved from each study according to a fixed protocol: study design; geographical location; population ethnicity, definition and numbers of cases and controls; DNA extraction and genotyping methods; frequency of the genotypes; mean age of the patients; and proportion of the patients who were male. When the studies included subjects of more than one ethnicity, the genotype data were extracted separately for each ethnic group. The genetic equilibrium of the LOXL1 gene for the control group of each study was evaluated by testing for Hardy-Weinberg equilibrium (HWE) using chi-square analyses . A state of disequilibrium was defined as a P value <0.05.
A summary OR with a 95% CI was calculated to assess the strength of the association of the LOXL1 gene polymorphisms with the PEXS/PEXG risk. The OR of each study was first calculated in a 2×2 table. Pooled ORs for the risk were then calculated for the allele frequency comparison and the additive, dominant, and recessive models. The between-study heterogeneity was evaluated with a Q statistic, and a P value <0.1 was considered statistically significant . If the P value was >0.1, a fixed-effect model was used for the meta-analysis; otherwise, a random-effect model was used. The significance of the pooled OR was determined with the Z-test, and P<0.05 was considered statistically significant.
To consider the possible sources of heterogeneity, we stratified the studies by ethnicity and repeated the analysis separately for each group. Furthermore, we performed a meta-regression analysis to assess the influence of the population characteristics . We also performed sensitivity analyses, serially excluding studies to determine the sources of heterogeneity and assess the stability of the results. The Begg funnel plot asymmetry wasassessed with Egger linear regression tests, a linear regression approach that measures funnel plot asymmetry on the natural logarithm scale of the OR. The significance of the intercept was determined by the t-test suggested by Egger, with P<0.05 considered to represent statistically significant publication bias .
All calculations were performed using the Comprehensive Meta-Analysis computer program version 2 (Biostat, Englewood, NJ, USA).
Characteristics of the articles in the analysis
The first search retrieved 83 articles. After eliminating the 60 studies not related to the topic, 23 potentially relevant studies were identified for further evaluation. In these studies, the titles or abstracts indicated that they evaluated the association between the LOXL1 gene polymorphisms and PEXS/PEXG susceptibility; however, some studies did not meet all of the study inclusion criteria. Ultimately, 11 studies were excluded because they were undesirable article types (review or letter; n = 8), had overlapping patient samples (n = 1) , insufficient data (n = 1) , or were not related to the three SNPs (n = 1) . Finally, 12 studies with a total of 1810 cases and 1790 controls were included in the analysis. Figure 1 shows the flowchart for the study selection.
SNPs, single nucleotide polymorphisms.
Table 1 shows the primary characteristics of the studies included in this meta-analysis. All of the studies used healthy control subjects, and the sample sizes ranged from 66 to 1133. The studies originated from Europe (n = 5) , , , , , Asia (n = 6) , , , , , , and Latin America (n = 1) . Two studies extracted only the allele frequencies , . The study by Park et al.  represented the data as (minor homozygosity + heterozygosity)/(major homozygosity). There were no significant differences between the case and control subjects with respect to age distribution.
Association of the rs1048661 T/G polymorphism with PEXS/PEXG.
Twelve studies evaluated the association between the rs1048661 T/G polymorphism and the risk of developing PEXS/PEXG. For the combined group data, significant associations between rs1048661 and susceptibility to PEXS/PEXG were observed for the additive, dominant, and recessive models, but not for the allele frequency comparison (Table 2). In a subgroup analysis performed by ethnicity, the pooled OR for the Caucasian population indicated a significantly decreased risk for PEXS/PEXG in the dominant (Figure 2), recessive, and TT versus GG additive models (Table 2). For the Asian population, the dominant (Figure 2) and two additive models showed a significant association between rs1048661 and a reduced susceptibility to PEXS/PEXG, while the allele frequency comparison and recessive model did not (Table 2). No significant associations were found in the Latin American population.
The first author of the study and year of publication are shown for each citation. The squares and horizontal lines correspond to the study specific odds ratio (OR) and 95% confidence interval (CI), respectively; the diamonds represent the pooled OR and 95% CI.
Association of the rs3825942 A/G polymorphism with PEXS/PEXG.
The twelve studies contained data regarding the association of the rs3825942 A/G polymorphism with the susceptibility to PEXS/PEXG. In comparison with the control group, the association between rs3825942 and decreased susceptibility to PEXS/PEXG was significant in all genetic models (Table 3). Furthermore, in the subgroup analysis by ethnicity, the genetic models showed a significant association between rs3825942 and a reduced susceptibility to PEXS/PEXG in the Caucasian and Asian populations (Table 3; Figure 3). In the Latin American population, the allele frequency comparison and the dominant model showed significant associations (Table 3) and a reduced risk (Figure 3).
The first author of the study and year of publication are shown for each citation. The squares and horizontal lines correspond to the study specific odds ratio (OR) and 95% confidence interval (CI); the diamonds represent the pooled OR and 95% CI.
Association of the rs2165241 C/T polymorphism and PEXS/PEXG.
Eight studies contained data for the rs216524 C/T polymorphism. The association between rs216524 and susceptibility to PEXS/PEXG was significant in all genetic models (Table 4). Furthermore, the subgroup analysis showed that all ethnicities had significant associations for all of the genetic models (Table 4) and reduced susceptibility to PEXS/PEXG (Figure 4).
The first author of the study and year of publication are shown for each citation. The squares and horizontal lines correspond to the study specific odds ratio (OR) and 95% confidence interval (CI); the diamonds represent the pooled OR and 95% CI.
Between-study heterogeneity analysis
The Q-test suggested significant between-study heterogeneity for several of the pooled models for each of the SNPs (Tables 2, 3, and 4). To examine the possible sources of heterogeneity, the studies were stratified by ethnicity, but the inconsistency index did not substantially decrease. We therefore performed a meta-regression analysis to assess the influence of the study characteristics on the effect estimates.
The male proportion and mean age of the subjects did not significantly affect the influences of rs1048661 and rs2165241 on PEXS/PEXG susceptibility. In contrast, for the effects of rs3825942 on PEXS/PEXG risk in Caucasians, a significant influence was detected for both the male proportion (Figure 5A; slope: 0.272; 95% CI: 0.167–0.376; P = 0.0001) and mean age (Figure 5B; slope: 0.796; 95% CI: 0.375–1.217; P = 0.0002) of the subjects.
Although the genotype distribution in four of the studies did not follow Hardy-Weinberg equilibrium , , , , the corresponding pooled ORs were not materially altered with or without including these studies (data not shown). In addition, the sensitivity analyses indicated that no single study had undue influence on the pooled OR results.
A Begg funnel plot analysis and Egger tests were used to access the publication bias of the literature. The shapes of the funnel plots of dominant models for the SNPs did not reveal evidence of obvious asymmetry (Figure S1).The Egger test results suggested that no publication bias was found in any of the comparison models (Tables 2, 3, and 4).
PEXS/PEXG is a disorder characterized by the accumulation of abnormal fibrillar deposits in the anterior segment of the eye. LOXL1, which serves both as a cross-linking enzyme and an element of the scaffold to ensure spatially defined deposition of elastin and collagen substrates , was recently identified by genetic linkage studies as associated with a susceptibility to PEXS/PEXG , , . A large number of high-frequency SNPs have been identified for this gene , . After screening these SNPs with the selection criteria, the rs1048661 T/G, rs3825942 A/G, and rs2165241 C/T polymorphisms were chosen to examine for their association with PEXS/PEXG susceptibility in this meta-analysis.
Our results indicated that rs1048661 TT carriers had 92.1% and 40.4% less risk of developing PEXS/PEXG than did the controls in Caucasian and Asian populations, respectively, but had no influence on the susceptibility in the Latin American population. Carriers of rs3825942 AA or rs2165241 CC also had significantly less risk of developing PEXS/PEXG than did the non-carriers. Despite the ethnic heterogeneity in the LOXL1 genotypes and the consequent variable susceptibility to PEXS/PEXG, the rs1048661, rs3825942 and rs2165241 SNPs may provide a powerful diagnostic tool to identify subjects who are more likely to develop PEXS/PEXG.
The analyses for heterozygosity of the variants are probably due to chance; the studies with small sample sizes for the minor homozygous alleles in PEXS/PEXG subjects would have insufficient statistical power to detect slight effects, or they may have generated a fluctuated risk estimate. Given this situation, our evaluation of the effects associated with heterozygosity for the variants in our analysis should be interpreted with caution.
In meta-analysis studies, heterogeneity could potentially restrict the interpretation of the pooled estimates, and ethnicity could play a role in determining the heterogeneity among studies. The different allele frequencies among the different ethnicities were a strong cause of the heterogeneity, leading us to do a subgroup analysis stratified by ethnicity; however, this resulted in no substantial decrease in the heterogeneity. Although applying a random effect model allowed us to estimate the effects of the different studies, we also performed a meta-regression to assess the influence of the population characteristics on the effect estimates. The results showed that the male proportion and mean age of the PEXS/PEXG patients positively correlated with the effect of SNP rs3825942 on the PEXS/PEXG susceptibility in the Caucasian population . Although our study cannot explain how LOXL1 interacts with gender and mean age, the correlations between them in the Caucasian population could be the primary cause of the heterogeneity. Considering the LOXL1 SNP genotypes together with these clinical predictors (gender and mean age) may allow for greater accuracy in predicting the probability of developing PEXS/PEXG.
To our knowledge, this study is the first meta-analysis to assess the association of the LOXL1 polymorphisms with PEXS/PEXG. This statistical method increased the power to detect and quantify an effect, and it provided a control for population differences that could lead to spurious associations if there are differences in gene frequency among the groups. Furthermore, this method allowed us to confirm the reliability and stability of the meta-analysis by performing publication bias and sensitivity analyses.
Some limitations of this study should be taken into consideration. First, the study populations were primarily Caucasian and Asian. The subgroup meta-analysis for ethnicity had little or no information for other ethnic groups. Thus, strengthening the statistical power will require more data from other ethnic groups. Second, although we were able to discern a significant association between the LOXL1 polymorphisms and susceptibility to PEXS/PEXG in the Caucasian and Asian populations, the sample size after pooling the existing studies was still relatively small. Third, the lack of available data prevented an adjustment for subgroup factors such as age, gender, and other variables that can interact with genetic factors to influence the marginal association estimates between the SNPs and PEXS/PEXG susceptibility. Should such data become available, a more precise analysis allowing for the adjustment of other covariates such as age, family history, environmental factors, and lifestyle would be feasible.
In summary, despite its limitations, this meta-analysis identified an association between the LOXL1 rs1048661, rs3825942, and rs2165241 polymorphisms and the risk of developing PEXS/PEXG. Future studies with larger sample sizes and additional ethnic groups are required to further clarify the association of the polymorphisms with the incidence of PEXS/PEXG, and these studies should consider gene-gene and gene-environment interactions in the analysis. Such studies should lead to a more comprehensive understanding of the influence of the LOXL1 gene polymorphisms on susceptibility to PEXS/PEXG.
Funnel plot analysis for publication bias. (A) rs3825942 TT+TG vs. GG; (B) rs3825942 AA+AG vs. GG; (C) rs2165241 CC+CT vs. TT.
Conceived and designed the experiments: XYZ JZT. Performed the experiments: JZT BW XQW FFM ZXP. Analyzed the data: JZT PFW XQW XYZ. Contributed reagents/materials/analysis tools: BW PFW. Wrote the paper: JZT XQW XYZ.
- 1. Elhawy E, Kamthan G, Dong CQ, Danias J (2012) Pseudoexfoliation syndrome, a systemic disorder with ocular manifestations. Hum Genomics 6: 22.
- 2. Schlotzer-Schrehardt U, Naumann GO (2006) Ocular and systemic pseudoexfoliation syndrome. Am J Ophthalmol 141: 921–937.
- 3. Bengtsson B, Heijl A (2005) A long-term prospective study of risk factors for glaucomatous visual field loss in patients with ocular hypertension. J Glaucoma 14: 135–138.
- 4. Shakya S, Dulal S, Maharjan IM (2008) Pseudoexfoliation syndrome in various ethnic population of Nepal. Nepal Med Coll J 10: 147–150.
- 5. Challa P (2009) Genetics of pseudoexfoliation syndrome. Curr Opin Ophthalmol 20: 88–91.
- 6. Schlotzer-Schrehardt U (2011) Genetics and genomics of pseudoexfoliation syndrome/glaucoma. Middle East Afr J Ophthalmol 18: 30–36.
- 7. Liu X, Zhao Y, Gao J, Pawlyk B, Starcher B, et al. (2004) Elastic fiber homeostasis requires lysyl oxidase-like 1 protein. Nat Genet 36: 178–182.
- 8. Wagenseil JE, Mecham RP (2007) New insights into elastic fiber assembly. Birth Defects Res C Embryo Today 81: 229–240.
- 9. Schlotzer-Schrehardt U, Hammer CM, Krysta AW, Hofmann-Rummelt C, Pasutto F, et al. (2012) LOXL1 deficiency in the lamina cribrosa as candidate susceptibility factor for a pseudoexfoliation-specific risk of glaucoma. Ophthalmology 119: 1832–1843.
- 10. Zenkel M, Krysta A, Pasutto F, Juenemann A, Kruse FE, et al. (2011) Regulation of lysyl oxidase-like 1 (LOXL1) and elastin-related genes by pathogenic factors associated with pseudoexfoliation syndrome. Invest Ophthalmol Vis Sci 52: 8488–8495.
- 11. Thorleifsson G, Magnusson KP, Sulem P, Walters GB, Gudbjartsson DF, et al. (2007) Common sequence variants in the LOXL1 gene confer susceptibility to exfoliation glaucoma. Science 317: 1397–1400.
- 12. Challa P, Schmidt S, Liu Y, Qin X, Vann RR, et al. (2008) Analysis of LOXL1 polymorphisms in a United States population with pseudoexfoliation glaucoma. Mol Vis 14: 146–149.
- 13. Abu-Amero KK, Osman EA, Dewedar AS, Schmidt S, Allingham RR, et al. (2010) Analysis of LOXL1 polymorphisms in a Saudi Arabian population with pseudoexfoliation glaucoma. Mol Vis 16: 2805–2810.
- 14. Malukiewicz G, Lesiewska-Junk H, Linkowska K, Mielnik M, Grzybowski T, et al. (2011) Analysis of LOXL1 single nucleotide polymorphisms in Polish population with pseudoexfoliation syndrome. Acta Ophthalmol 89: e64–66.
- 15. Micheal S, Khan MI, Akhtar F, Ali M, Ahmed A, et al. (2012) Role of Lysyl oxidase-like 1 gene polymorphisms in Pakistani patients with pseudoexfoliative glaucoma. Mol Vis 18: 1040–1044.
- 16. Jaimes M, Rivera-Parra D, Miranda-Duarte A, Valdes G, Zenteno JC (2012) Prevalence of high-risk alleles in the LOXL1 gene and its association with pseudoexfoliation syndrome and exfoliation glaucoma in a Latin American population. Ophthalmic Genet 33: 12–17.
- 17. Ozaki M, Lee KY, Vithana EN, Yong VH, Thalamuthu A, et al. (2008) Association of LOXL1 gene polymorphisms with pseudoexfoliation in the Japanese. Invest Ophthalmol Vis Sci 49: 3976–3980.
- 18. Park DY, Won HH, Cho HK, Kee C (2013) Evaluation of lysyl oxidase-like 1 gene polymorphisms in pseudoexfoliation syndrome in a Korean population. Mol Vis 19: 448–453.
- 19. Sowka J (2004) Pseudoexfoliation syndrome and pseudoexfoliative glaucoma. Optometry 75: 245–250.
- 20. Wigginton JE, Cutler DJ, Abecasis GR (2005) A note on exact tests of Hardy-Weinberg equilibrium. Am J Hum Genet 76: 887–893.
- 21. Higgins JP, Thompson SG, Deeks JJ, Altman DG (2003) Measuring inconsistency in meta-analyses. BMJ 327: 557–560.
- 22. Kavvoura FK, Ioannidis JP (2008) Methods for meta-analysis in genetic association studies: a review of their potential and pitfalls. Hum Genet 123: 1–14.
- 23. Song F, Gilbody S (1998) Bias in meta-analysis detected by a simple, graphical test. Increase in studies of publication bias coincided with increasing use of meta-analysis. BMJ 316: 471.
- 24. Fan BJ, Chen T, Grosskreutz C, Pasquale L, Rhee D, et al. (2008) Lack of association of polymorphisms in homocysteine metabolism genes with pseudoexfoliation syndrome and glaucoma. Mol Vis 14: 2484–2491.
- 25. Hewitt AW, Sharma S, Burdon KP, Wang JJ, Baird PN, et al. (2008) Ancestral LOXL1 variants are associated with pseudoexfoliation in Caucasian Australians but with markedly lower penetrance than in Nordic people. Hum Mol Genet 17: 710–716.
- 26. Guadarrama-Vallejo D, Miranda-Duarte A, Zenteno JC (2013) The T allele of lysyl oxidase-like 1 rs41435250 is a novel risk factor for pseudoexfoliation syndrome and pseudoexfoliation glaucoma independently and through intragenic epistatic interaction. Mol Vis 19: 1937–1944.
- 27. Fan BJ, Pasquale L, Grosskreutz CL, Rhee D, Chen T, et al. (2008) DNA sequence variants in the LOXL1 gene are associated with pseudoexfoliation glaucoma in a U.S. clinic-based population with broad ethnic diversity. BMC Med Genet 9: 5.
- 28. Pasutto F, Krumbiegel M, Mardin CY, Paoli D, Lammer R, et al. (2008) Association of LOXL1 common sequence variants in German and Italian patients with pseudoexfoliation syndrome and pseudoexfoliation glaucoma. Invest Ophthalmol Vis Sci 49: 1459–1463.
- 29. Metaxaki I, Constantoulakis P, Papadimitropoulos M, Filiou E, Georgopoulos G, et al. (2013) Association of lysyl oxidase-like 1 gene common sequence variants in Greek patients with pseudoexfoliation syndrome and pseudoexfoliation glaucoma. Mol Vis 19: 1446–1452.
- 30. Ramprasad VL, George R, Soumittra N, Sharmila F, Vijaya L, et al. (2008) Association of non-synonymous single nucleotide polymorphisms in the LOXL1 gene with pseudoexfoliation syndrome in India. Mol Vis 14: 318–322.
- 31. Lee KY, Ho SL, Thalamuthu A, Venkatraman A, Venkataraman D, et al. (2009) Association of LOXL1 polymorphisms with pseudoexfoliation in the Chinese. Mol Vis 15: 1120–1126.
- 32. Topouzis F, Wilson MR, Harris A, Founti P, Yu F, et al. (2011) Risk factors for primary open-angle glaucoma and pseudoexfoliative glaucoma in the Thessaloniki eye study. Am J Ophthalmol 152: 219–228 e211.