Two recent studies have provided conflicting evidence about the association between the gene encoding serpin peptidase inhibitor, clade G, member 1 (SERPING1) and age-related macular degeneration (AMD). An Article in The Lancet1 suggested that a variant (rs2511989) in intron 6 of SERPING1 conferred protection against the development of AMD in case–control studies from the UK and USA. A follow-up study2 showed the absence of any association in two US case–control studies, including the AREDS cohort.3 As a member of the classical complement pathway, SERPING1 is a plausible candidate gene for AMD since several genes encoding proteins involved in the innate immune response (CFH,4–7 C2/CFB,8 C39,10), are robustly associated with the development of this disorder. To test this hypothesis further, we genotyped rs2511989 in seven large case–control studies involving individuals of European descent (4881 patients with AMD and 2842 matched controls; see webappendix for details).
Samples from all cohorts in the seven centres were collected and genotyped independently to minimise data selection, analysis, or interpretation bias. Data from the seven studies are summarised in the table and figure. They show—both independently and collectively—that there is no significant association between the rs2511989 variant in the SERPING1 gene and AMD. In fact, data from no single study suggested even a trend towards protective association (p>0·3 in all cohorts; table); one cohort (AREDS) suggested an opposite, marginally significant (p=0·03), trend towards increased susceptibility. However, the study by Park and colleagues2 found no association after genotyping more AREDS samples, thereby suggesting that the marginal association detected in our study is due to a limited sample size. The minor allele frequency (MAF) for the rs2511989 single nucleotide polymorphism (SNP) varied substantially between studies, from 0·35 and 0·45 both in AMD patients (average 0·40) and in controls matched for age and ethnic group (average 0·40). However, there was almost no difference in the MAF between patients and controls within all, except for the AREDS, studies. When summarising the data across cohorts, the MAF of the rs2511989 SNP was practically identical in 2842 controls and in 4881 patients with AMD (0·40), clearly showing no association (p=0·99; odds ratio 1, table). The same lack of significance was seen for all AMD subphenotypes, early AMD, geographic atrophy, and choroidal neovascularisation, when analysed separately. Since the other SNPs (eg, rs2509879, rs2511988, etc) reported in the study by Ennis and colleagues are in high linkage disequilibrium with the rs2511989 variant,1 the MAFs for those confirmed the rs2511989 data (not shown),2 as did a combined analysis of all three studies by the Mantel-Haenszel fixed effects model (figure).
Table.
Genotype counts and association analysis of the SERPING1 rs2511989 variant in seven studies
| GG | GA | AA | MAF | HWE | p (allele) | p (genotype) | |
|---|---|---|---|---|---|---|---|
| Columbia | 0·62 | 0·39 | |||||
| AMD (n=1004) | 449 (44·7%) | 431 (42·9%) | 124 (12·4%) | 0·34 | 0·45 | ||
| Control (n=363) | 151 (41·6%) | 171 (47·1%) | 41 (12·3%) | 0·35 | 0·79 | ||
|
| |||||||
| Iowa | 0·57 | 0·68 | |||||
| AMD (n=368) | 116 (31·5%) | 178 (48·4%) | 74 (20·1%) | 0·44 | 0·91 | ||
| Control (n=115) | 37 (32·2%) | 59 (51·3%) | 19 (16·5%) | 0·42 | 0·89 | ||
|
| |||||||
| Amsterdam | 0·81 | 0·62 | |||||
| AMD (n=338) | 107 (31·7%) | 184 (54·4%) | 47 (13·9%) | 0·41 | 0·08 | ||
| Control (n=257) | 84 (32·7%) | 131 (51·0%) | 42 (16·3%) | 0·42 | 0·74 | ||
|
| |||||||
| Rotterdam | 0·85 | 0·53 | |||||
| AMD (n=1017) | 328 (32·3%) | 518 (50·9%) | 171 (16·8%) | 0·42 | 0·37 | ||
| Control (n=842) | 285 (33·8%) | 407 (48·3%) | 150 (17·9%) | 0·42 | 0·97 | ||
|
| |||||||
| AREDS | 0·03 | 0·04 | |||||
| AMD (n=415) | 133 (32·0%) | 188 (45·3%) | 94 (22·7%) | 0·45 | 0·20 | ||
| Control (n=213) | 90 (42·3%) | 81 (38·0%) | 42 (19·7%) | 0·39 | 0·02 | ||
|
| |||||||
| Australia | 0·28 | 0·43 | |||||
| AMD (n=741) | 251 (33·9%) | 367 (49·5%) | 123 (16·6%) | 0·41 | 0·83 | ||
| Control (n=327) | 105 (32·1%) | 157 (48·0%) | 65 (19·9%) | 0·44 | 0·90 | ||
|
| |||||||
| Germany | 0·71 | 0·80 | |||||
| AMD (n=998) | 377 (37·8%) | 485 (48·6%) | 136 (13·6%) | 0·38 | 0·58 | ||
| Control (n=725) | 284 (39·2%) | 341 (47·0%) | 100 (13·8%) | 0·37 | 0·99 | ||
|
| |||||||
| Total | 0·99 | 0·79 | |||||
| AMD (n=4881) | 1761 (36·1%) | 2351 (48·2%) | 769 (15·7%) | 0·40 | 0·94 | ||
| Control (n=2842) | 1036 (36·5%) | 1347 (47·4%) | 459 (16·1%) | 0·40 | 0·83 | ||
AMD=age-related macular degeneration. MAF=minor allele (A allele) frequency. HWE=Hardy-Weinberg equilibrium. Ages and AMD subtypes of all study participants are presented in the webappendix.
Figure. Forest plot of the combined analysis of ten case–control cohorts from the three studies.
Horizontal lines represent 95% CIs, and various-sized squares correspond to significance levels of p values. Test for heterogeneity was highly significant (p=3·1×10−15), suggesting that calculation of a combined odds ratio is not valid.
Excluding experimental errors discussed by Park and colleagues,2 the difference between our collective findings and those published by Ennis and colleagues could be explained by the two following scenarios:
Whereas most methods used in our study closely match those used by Ennis and colleagues (genotyping methods, selection criteria for AMD patients, etc), the selection and matching of controls in the seven centres in this study were more rigorous by age and, in some centres (Germany, the Netherlands), also by ethnicity (see webappendix). The significantly higher MAFs in controls in both cohorts used in Ennis and colleagues’ study (average 0·46 vs 0·40 in this study and the one by Park and colleagues2, p=8×10−6) could be explained by the fact that controls were younger than AMD patients.1
Other possible sources for the difference in results could be non-random sampling of study populations or population substructure (stratification). For example, although there was no significant difference in MAFs between any of the AMD and control cohorts within any one study, the same numbers varied significantly between some studies (MAF 0·35 in AMD cases from New York vs >0·41 in all other cohorts), probably originating from varied ethnic composition within populations of European descent. The New York (Columbia University) study population originates mainly from eastern Europe, whereas that from Iowa from western Europe and Scandinavia. At the same time it is of interest that one of the two cohorts in Ennis and colleagues’ study (248 AMD cases) was collected at exactly the same location, the University of Iowa, as was one of the seven cohorts in our study (368 AMD cases). However, whereas the frequencies of the minor allele in Iowa controls were very similar in the two studies (0·44 vs 0·42), they differed significantly in patients with AMD (0·35 vs 0·44; p=0·001).
In summary, analysis of the rs2511989 variant in SERPING1 in seven large, well characterised case–control studies did not confirm an association between this variant and AMD. Since there is also no evidence of a functional role for this variant, it is unlikely that the SERPING1 gene has a significant, if any, role in the aetiology of AMD.
Footnotes
Members of the International AMD Genetics Consortium: Joanna E Merriam, R Theodore Smith, Gaetano R Barile, Julie Sawitzke, Karen M Gehrs, Jill L Hageman, Norma J Miller, Mary L Howard, Robyn H Guymer, Andrea Richardson, Lintje Ho, Johannes R Vingerling, Andre G Uitterlinden, Paulus T de Jong, Dominique Baas, Lars G Fritsche, Claudia N Keilhauer. GSH receives grant support, equity, and salary from Optherion, which develops diagnostics and treatments for AMD. The other authors declare that they have no conflicts of interest.
References
- 1.Ennis S, Jomary C, Mullins R, et al. Association between the SERPING1 gene and age-related macular degeneration: a two-stage case-control study. Lancet. 2008;372:1828–34. doi: 10.1016/S0140-6736(08)61348-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Park KH, Ryu E, Tosakulwong N, Wu Y, Edwards AO. Common variation in the SERPING1 gene is not associated with age-related macular degeneration in two independent groups of subjects. Mol Vis. 2009;15:200–07. [PMC free article] [PubMed] [Google Scholar]
- 3.Age-Related Eye Disease Study Research Group. The Age-Related Eye Disease Study (AREDS): design implications. AREDS report no 1. Control Clin Trials. 1999;20:573–600. doi: 10.1016/s0197-2456(99)00031-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Edwards AO, Ritter R, 3rd, Abel KJ, Manning A, Panhuysen C, Farrer LA. Complement factor H polymorphism and age-related macular degeneration. Science. 2005;308:421–24. doi: 10.1126/science.1110189. [DOI] [PubMed] [Google Scholar]
- 5.Hageman GS, Anderson DH, Johnson LV, et al. A common haplotype in the complement regulatory gene factor H (HF1/CFH) predisposes individuals to age-related macular degeneration. Proc Natl Acad Sci USA. 2005;102:7227–32. doi: 10.1073/pnas.0501536102. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Haines JL, Hauser MA, Schmidt S, et al. Complement factor H variant increases the risk of age-related macular degeneration. Science. 2005;308:419–21. doi: 10.1126/science.1110359. [DOI] [PubMed] [Google Scholar]
- 7.Klein RJ, Zeiss C, Chew EY, et al. Complement factor H polymorphism in age-related macular degeneration. Science. 2005;308:385–89. doi: 10.1126/science.1109557. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Gold B, Merriam JE, Zernant J, et al. Variation in factor B (BF) and complement component 2 (C2) genes is associated with age-related macular degeneration. Nat Genet. 2006;38:458–62. doi: 10.1038/ng1750. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Maller JB, Fagerness JA, Reynolds RC, Neale BM, Daly MJ, Seddon JM. Variation in complement factor 3 is associated with risk of age-related macular degeneration. Nat Genet. 2007;39:1200–01. doi: 10.1038/ng2131. [DOI] [PubMed] [Google Scholar]
- 10.Yates JR, Sepp T, Matharu BK, et al. Complement C3 variant and the risk of age-related macular degeneration. N Engl J Med. 2007;357:553–61. doi: 10.1056/NEJMoa072618. [DOI] [PubMed] [Google Scholar]