Association of Genetic Polymorphisms and Age-Related Macular Degeneration in Chinese Population

Abstract

Purpose.: We explored associations between age-related macular degeneration (AMD) and genetic variants of 10 genes in a nationwide Chinese population.

Methods.: In this multicenter case-control study, 535 AMD patients and 469 controls were recruited from 16 centers that spread from the north to the south of China. All participants underwent comprehensive eye examinations, and 40 single nucleotide polymorphisms (SNPs) of 10 genes were selected. DNA samples were genotyped with the MassArray system. The effect of the genotypes and haplotypes on AMD was assessed with logistic regression analysis, adjusted for age, sex, long-term residence, and family origin.

Results.: In our study, 11 SNPs in complement H (CFH), 2 in age-related maculopathy susceptibility 2 (ARMS2), and 2 in high-temperature requirement factor A1 (HTRA1) were associated significantly with AMD. They were rs551397, rs800292, rs1329424, rs1061170, rs10801555, rs12124794, rs10733086, rs10737680, rs2274700, rs1410996, and rs380390 in CFH; rs10490924 and rs2736912 in ARMS2; and rs11200638 and rs3793917 in HTRA1. Three haplotypes in CFH, predisposed the patients significantly to AMD (P < 0.001, P = 0.001, and P < 0.001, respectively). With the sample size of our study, no relationship was found for AMD and the SNPs tested in complement 3 (C3); serpin peptidase inhibitor, clade G, member 1 (SERPING1); vascular endothelial growth factor (VEGF); cholesterol ester transfer protein (CETP); lipoprotein lipase (LPL); hepatic lipase (LIPC); and metallopeptidase inhibitor 3 (TIMP3) genes.

Conclusions.: Gene variants in CFH, ARMS2, and HTRA1 contribute to AMD in the Chinese population.

Introduction

Age-related macular degeneration (AMD) affects elder adults and there is no effective treatment against it. In the developed world, AMD accounts for approximately half of all vision impairment or blind registrations. ¹ In Asia, the pooled prevalence is 6.8% for early AMD and 0.56% for advanced AMD in populations between 40 and 79 years old. ² For Chinese, some regional studies indicated that the prevalence of early AMD ranges from 4.7%–9.2% and that of the advanced AMD from 0.2–1.9% in elderly people. ^3,4  

Despite the fact that AMD greatly impairs quality of life, its etiology remains unclear. The prevailing view on AMD etiology is that the genetic factor might have an important role. ^5,6 So far, studies about the genetic background of AMD have reported tremendous progress in identifying the polymorphisms in three genes to be highly associated with AMD in different ethnic groups: complement H (CFH), age-related maculopathy susceptibility 2 (ARMS2, also known as LOC387715), and high-temperature requirement factor A1 (HTRA1). ^7–13  

Several studies indicated that the genetic variants of other genes conferred disease risk in Caucasian populations, namely complement 3 (C3); serpin peptidase inhibitor, clade G, member 1 (SERPING1); vascular endothelial growth factor (VEGF); cholesterol ester transfer protein (CETP); lipoprotein lipase (LPL); hepatic lipase (LIPC), and metallopeptidase inhibitor 3 (TIMP3). ^14–22 However, a few studies in China and Japan reported that there were no significant contributions from C3 (rs2230199, rs1047286), SERPING1 (rs2511989, rs1005510), and VEGF (rs3025039) to AMD, ^7,23–26 whereas a study in Taiwan reported that VEGF (rs3025039) polymorphism was associated with wet AMD. ²⁷ So far, to our knowledge no more convincing data have been reported about the relationship between AMD and these genetic variants in the Chinese population. Therefore, we conducted the current study to investigate the association between multiple genetic polymorphisms and AMD in a nationwide Chinese population. 

Materials and Methods

AMD Patients and Control Subjects

The current study was part of the hospital-based multicenter project of Clinical Features and Interventions of AMD supported by the Ministry of Science and Technology of China (Grant No. 2006BAI02B05). People's Hospital, and the Department of Epidemiology and Biostatistics of Peking University led the study. To represent the Chinese population comprehensively, 16 local hospitals located sporadically in the northern and southern parts of China were selected as the research centers. All researchers in these local hospitals received training about the diagnosis of AMD, inclusion and exclusion criteria of the participants, and specimen collection, which was conducted by the ophthalmologists of People's Hospital, and experts of the Department of Epidemiology and Biostatistics. PASS software (Version 08.0.3; NCSS, LLC, Kaysville, UT) was used to calculated the sample size (α = 0.05, β = 0.20). According to the frequencies of the risk allele of single nucleotide polymorphisms (SNPs) that have been reported, 485 patients and 485 controls were needed. The hospitals with a large AMD patient capacity were planned to recruit 40 patients and 40 controls, and the hospitals with a small capacity were planned to recruit 20 patients and 20 controls. After strictly re-checking by the reading center of Peking University People's Hospital, 535 patients and 469 controls were eligible. The project was approved by the Ethics Committee of Peking University People's Hospital, and the procedures were consistent with the Declaration of Helsinki. All AMD patients and controls provided written informed consent. 

Participants were enrolled from the ophthalmologic department of the 16 hospitals between 2008 and 2010. All of them received ophthalmologic examinations, including slit-lamp biomicroscopy and funduscopy. If they were revealed as having AMD by funduscopy, the patients would be examined further with fluorescence fundus angiography (FFA) (Topcon TRC-50EX; Topcon, Tokyo, Japan), indocyanine green angiography (ICGA) (Heidelberg Spectralis HRA; Heidelberg Engineering, Heidelberg, Germany), or optical coherence tomography (OCT) (Zeiss-Humphrey, Dublin, CA). When confirmed by any of the three tests, the AMD patients would be included if they were 50 years or older and had no signs of other retinal diseases. Patients were excluded if they had polypoidal choroidal vasculopathy (PCV), myopia of 6.00 diopters (D) or above, macular dystrophy, central serous chorioretinopathy, any kind of vein occlusion, diabetic retinopathy, uveitis, other diseases involving malfunction of retinal photoreceptor cells, or serious chronic diseases. Visitors who went to the ophthalmologic department were taken as controls if they were 50 years or older, confirmed negative for AMD with fundus examination, had no signs of other retinal diseases, or had no serious chronic diseases. In the end, we recruited 535 unrelated Chinese patients with AMD and 469 controls (Table 1). 

Table 1. 

 

Number of Patients and Controls of the 16 Hospitals

Hospital	City	Province/Municipality	Patients	Controls
The Second Affiliated Hospital of Harbin Med Univ.	Harbin	Heilongjiang	40	34
The First Affiliated Hospital of Harbin Med Univ.	Harbin	Heilongjiang	42	38
The First Hospital of China Med Univ.	Shenyang	Liaoning	40	32
Peking University First Hospital	Beijing	Beijing	19	21
Peking University People's Hospital	Beijing	Beijing	26	17
Peking Union Medical College Hospital	Beijing	Beijing	37	9
The General Hospital of The People's Liberation Army	Beijing	Beijing	19	18
Shandong Eye Institute	Qingdao	Shandong	35	21
Xijing Hospital	Xi'an	Shaanxi	39	40
Jiangsu Province Hospital	Nanjing	Jiangsu	19	20
Renmin Hospital of Wuhan Univ.	Wuhan	Hubei	31	39
Shanghai First People's Hospital	Shanghai	Shanghai	41	39
West China Medical School-West China Hospital	Chengdu	Sichuan	40	26
The Affiliated Eye Hospital of Wenzhou Medical College	Wenzhou	Zhejiang	37	37
The Second Xiangya Hospital of Central South Univ.	Changsha	Hunan	32	39
Zhongshan Ophthalmic Center	Guangzhou	Guangdong	38	39
Total
16	12	12	535	469

Classification of Age-Related Macular Degeneration

Fundus photographs taken at local hospitals were reviewed by the experts of People's Hospital. Stages of AMD were assigned based on the classification of the Age-Related Eye Disease Study (AREDS). ²⁸ Early AMD was considered drusen with or without pigmentary abnormalities and in the absence of signs of advanced AMD. Advanced AMD was considered atrophic or neovascular AMD. Atrophic AMD was identified when geographic atrophy forms by depigmentation of the RPE and exposure of underlying choroidal vessels, and when neovascular AMD is absent. Neovascular AMD was identified when a patient had choroidal neovascularization (CNV), sensory retinal detachment, RPE serous/hemorrhagic detachment, subretinal/sub-RPE fibrosis, or disciform scars. 

Genotyping

To analyze genetic polymorphisms of AMD, we selected 40 SNPs in 10 genes for their association with AMD (Table 2). Genomic DNA was extracted from the peripheral blood with a DNA extraction kit (DP319-01; Tiangen Biotech, Beijing, China). The PCR was performed on a thermal cycler (ABI GeneAmp 9700 384 Dual; ABI, Foster City, CA). Genotypes were analyzed with Typer 4.0 software (MassARRAY Compact System; Sequenom, San Diego, CA). All SNPs in this report had a genotyping success rate of >97%. 

Table 2. 

 

The SNPs Selected of the 10 Genes

Gene	SNPs
CFH	rs3753394, rs7524776, rs551397, rs800292, rs1329424, rs1061170, rs10801555, rs12124794, rs6695321, rs10733086, rs10737680, rs2274700, rs3753396, rs1410996, rs380390
ARMS2	rs10490924, rs2736912
HTRA1	rs3793917, rs11200638
C3	rs11569562, rs7257062, rs408290, rs8112351, rs2230205, rs1047286, rs2230199, rs2250656
SERPING1	rs1005510, rs2511989
VEGF	rs833069, rs3025039
CETP	rs173539, rs17231506, rs9939224, rs1532625, rs2303789, rs289744
nearby LPL	rs12678919
nearby LIPC	rs10468017
nearby TIMP3	rs9621532

Statistical Analysis

We compared the genotypes of the AMD patients to those of the controls using the χ ² test. The adjusted odds ratios (ORs) and 95% confidence intervals (CIs) for the effect of genotype on the risk of AMD were estimated by the logistic regression analysis. Hardy-Weinberg equilibrium for each SNP polymorphism was tested by the χ ² test. All the above analyses were performed with SPSS software (version 15.0 for Windows; SPSS Inc., Chicago, IL). Permutation was used to correct for multiple testing. The linkage disequilibrium (LD), haplotypes (confidence intervals algorithm proposed by Gabriel et al. ²⁹ ) and the corrected probabilities estimated by the permutation test were examined with Haploview software (provided in the public domain, http://www.broadinstitute.org/haploview/haploview, version 4.1, April, 2008). The adjusted ORs of the haplotypes were calculated by logistic regression analysis via PLINK software (provided in the public domain, http://pngu.mgh.harvard.edu/∼purcell/plink, version 1.07, August 2009, accessed March 2011). P < 0.05 was considered statistically significant. 

Results

Of the 535 AMD patients 64 (12.0%) had early and 471 (88.0%) had advanced AMD. Among the patients with advanced AMD 464 (98.5%) had neovascular and 7 (1.5%) had atrophic AMD. There were more males in the patient group than in controls (P < 0.001) and the patients were significantly elder than controls (P < 0.001). There was no significant difference of the nationality between patients and controls (P = 0.307). The percentage of patients among the participants in the north was higher than that of the south (P = 0.025). There were more patients with family origin in the north than in the south (P = 0.043, Table 3). Because a significant difference was found in sex, age, long-term residence, and family origin in patients and controls, the following estimation of ORs were adjusted by logistic regression analysis. 

Table 3. 

 

Demographic Characteristics of AMD Patients and Controls

	Controls	Patients
		All AMD		Early AMD		Advanced AMD
	n	n	P*	n	P*	n	P*
Sex			<0.001		0.119		<0.001
Male	217	324		23		301
Female	252	211		41		170
Age			<0.001		0.139		<0.001
50∼	155	106		14		92
60∼	157	168		28		140
70∼	157	261		22		239
Nationality			0.307		0.403		0.205
Han	450	516		63		453
Minorities	5	10		0		10
Long-term residence			0.025		0.001		0.125
North of China	225	297		45		252
South of China	238	236		18		218
Family origin			0.043		0.001		0.202
North of China	227	295		46		249
South of China	236	237		18		219

*  χ²test.

All 40 SNPs but rs173539 showed no significant deviation from Hardy-Weinberg equilibrium in the control group (P > 0.05). The genotype frequencies for patients and controls are shown in Table 4. The risk allele frequencies of 11 SNPs tested in CFH, and 4 SNPs tested in ARMS2 and HTRA1 were significantly different between AMD patients and controls. The positive SNPs showed the same association with advanced and neovascular AMD. Other tested SNPs of C3, SERPING1, VEGF, CETP, LPL, LIPC, and TIMP3 did not reach the level of significance between AMD patients and controls. 

Table 4. 

 

Gene Variants Observed in AMD Patients and Controls

Gene	dbSNP ID	Risk Allele	Genotype Frequency
			Controls*	All AMD*	P corr†	Early AMD*	P†	Advanced AMD*	P corr†	Neovascular AMD*	P corr†
CFH	rs3753394	T	131/230/102	181/244/104	0.343	20/29/14	1.000	161/215/90	0.287	158/212/89	0.286
	rs7524776	C	1/46/416	3/50/480	1.000	0/7/57	1.000	3/43/423	1.000	3/42/417	1.000
	rs551397	G	157/224/75	248/213/55	<0.001	26/27/9	0.917	222/186/46	<0.001	218/183/46	<0.001
	rs800292	C	158/221/76	249/205/56	<0.001	26/26/9	0.901	223/179/47	<0.001	219/176/47	<0.001
	rs1329424	A	1/51/412	9/93/427	0.001	1/5/57	1.000	8/88/370	<0.001	8/86/365	<0.001
	rs1061170	C	1/49/395	9/89/391	0.001	1/6/55	1.000	8/83/336	<0.001	8/81/331	<0.001
	rs10801555	A	1/51/408	9/91/425	0.002	1/6/57	1.000	8/85/368	<0.001	8/83/363	<0.001
	rs12124794	A	188/216/56	264/213/49	0.031	26/27/10	0.999	238/186/39	0.008	233/184/39	0.012
	rs6695321	A	8/113/343	13/138/377	0.911	1/11/51	0.872	12/127/326	0.690	11/126/321	0.675
	rs10733086	A	1/65/397	10/113/410	0.001	1/7/56	1.000	9/106/354	<0.001	8/105/349	<0.001
	rs10737680	A	153/223/92	234/230/67	<0.001	23/27/13	1.000	211/203/54	<0.001	206/201/54	<0.001
	rs2274700	C	137/216/88	231/216/63	<0.001	24/26/12	0.955	207/190/51	<0.001	202/188/51	<0.001
	rs3753396	G	105/239/120	156/253/123	0.205	19/28/17	0.990	137/225/106	0.205	134/222/105	0.218
	rs1410996	C	148/224/92	235/228/67	<0.001	23/27/13	0.999	212/201/54	<0.001	207/199/54	<0.001
	rs380390	C	1/45/401	9/97/408	<0.001	1/6/55	0.998	8/91/353	<0.001	8/89/348	<0.001
ARMS2	rs10490924	T	97/223/147	251/197/82	<0.001	18/29/16	0.282	233/168/66	<0.001	228/166/66	<0.001
	rs2736912	C	358/91/11	460/68/3	0.001	51/13/0	0.7580	409/55/3	<0.001	402/55/3	<0.001
HTRA1	rs3793917	C	92/221/141	256/190/79	<0.001	20/27/16	0.111	236/163/63	<0.001	231/161/63	<0.001
	rs11200638	A	104/224/140	255/193/84	<0.001	19/28/16	0.309	236/165/68	<0.001	231/163/68	<0.001
C3	rs11569562	C	134/228/101	144/284/102	1.000	20/29/14	1.000	124/255/88	1.000	123/250/87	1.000
	rs7257062	T	250/177/37	289/210/32	0.997	34/23/6	1.000	255/187/26	0.989	251/184/26	0.992
	rs408290	C	323/127/15	390/127/12	0.675	52/10/1	0.285	338/117/11	0.906	333/115/11	0.905
	rs8112351	C	383/74/4	449/76/4	0.998	58/5/0	0.402	391/71/4	1.000	385/70/4	1.000
	rs2230205	A	107/221/137	130/246/156	1.000	16/31/17	0.999	114/215/139	1.000	113/210/138	1.000
	rs1047286	T	0/1/455	0/2/525	1.000	0/0/63	1.000	0/2/462	1.000	0/2/455	0.999
	rs2230199	G	0/2/463	0/4/528	0.995	0/0/63	1.000	0/4/465	0.982	0/4/458	0.969
	rs2250656	A	273/164/21	329/171/23	0.951	41/19/4	1.000	288/152/19	0.936	283/150/19	0.948
SERPING1	rs1005510	G	28/181/256	37/201/293	0.968	5/21/38	0.951	32/180/255	0.921	32/178/250	0.876
	rs2511989	A	7/86/371	13/96/422	0.837	3/9/51	0.866	10/87/371	0.856	10/86/365	0.848
VEGF	rs833069	A	155/227/84	190/253/89	0.629	27/28/8	0.209	163/225/81	0.863	161/222/79	0.835
	rs3025039	C	302/143/18	341/179/13	0.998	40/22/2	0.958	301/157/11	0.990	299/153/10	0.931
CETP	rs173539‡	T	14/4/397	19/6/465	0.984	2/0/59	0.982	17/6/406	0.959	17/6/400	0.943
	rs17231506	T	18/116/331	15/150/364	0.992	1/21/41	0.992	14/129/323	0.999	14/126/319	0.999
	rs9939224	T	5/84/374	9/94/427	1.000	1/8/53	0.982	8/86/374	0.996	8/86/367	0.989
	rs1532625	A	45/172/248	43/226/262	0.980	3/29/31	1.000	40/197/231	0.968	39/193/229	0.982
	rs2303789	T	320/131/14	388/130/14	0.707	46/16/2	0.999	342/114/12	0.700	338/112/11	0.613
	rs289744	A	211/193/37	260/210/49	1.000	32/27/5	0.999	228/183/44	1.000	225/179/44	1.000
LPL	rs12678919	A	386/76/5	438/86/5	0.939	51/11/1	0.746	387/75/4	0.876	380/75/4	0.936
LIPC	rs10468017	C	316/136/10	374/145/11	0.492	45/17/1	0.693	329/128/10	0.563	325/125/10	0.525
TIMP3	rs9621532	A	439/26/0	510/22/0	0.306	60/4/0	1.000	450/18/0	0.214	443/18/0	0.278

*  The number indicates homozygous risk genotype/heterozygous genotype/homozygous normal genotype.

†  The corrected probabilities of allelic association estimated by permutation test with Haploview (number of permutations = 10,000).

‡  This SNP violated the Hardy-Weinberg equilibrium.

The adjusted ORs, 95% CIs, and P values of the positive SNPs are shown in Table 5. The homozygosity of risk alleles for five SNPs in CFH (rs551397, rs800292, rs10737680, rs2274700, rs1410996) and one SNP in ARMS2 (rs2736912) conferred a 2.11-fold (95% CI, 1.41–3.26), 2.13-fold (95% CI, 1.41–3.23), 2.01-fold (95% CI, 1.37–2.97), 2.26-fold (95% CI, 1.52–3.38), 2.08-fold (95% CI, 1.41–3.07), and 4.83-fold (95% CI, 1.31–17.83) increased likelihood of AMD. Five SNPs in CFH (rs1329424, rs1061170, rs10801555, rs10733086, rs380390), one in ARMS2 (rs10490924) as well as the two in HTRA1 (rs11200638, rs3793917) showed significant association with AMD for heterozygosity and homozygosity of risk alleles. The homozygosity of risk alleles for rs12124794 in CFH conferred a 1.63-fold (95% CI, 1.02–2.61) risk of advanced AMD. 

Table 5. 

 

Estimated ORs, P Values of CFH, ARMS2, and HTRA1 Gene Variants

dbSNP ID	All AMD Patients				Advanced AMD Patients
	ORhetero (95% CI)*	P*	ORhomo (95% CI)†	P†	ORhetero (95% CI)*	P*	ORhomo (95% CI)†	P†
CFH
rs551397	1.18 (0.78–1.78)	0.427	2.11 (1.41–3.26)	<0.001	1.24 (0.80–1.91)	0.339	2.29 (1.48–3.56)	<0.001
rs800292	1.17 (0.77–1.75)	0.465	2.13 (1.41–3.23)	<0.001	1.21 (0.78–1.86)	0.387	2.28 (1.48–3.53)	<0.001
rs1329424	1.73 (1.19–2.53)	0.004	8.15 (1.02–65.44)	0.048	1.86 (1.27–2.73)	0.002	8.37 (1.03–68.39)	0.047
rs1061170	1.86 (1.27–2.74)	0.002	8.50 (1.06–68.38)	0.044	1.98 (1.34–2.95)	0.001	8.84 (1.08–72.39)	0.042
rs10801555	1.71 (1.17–2.50)	0.005	8.15 (1.02–65.42)	0.048	1.81 (1.23–2.66)	0.003	8.38 (1.03–68.40)	0.047
rs12124794	0.93 (0.60–1.45)	0.747	1.46 (0.94–2.28)	0.093	1.00 (0.62–1.60)	0.988	1.63 (1.02–2.61)	0.043
rs10733086	1.66 (1.18–2.34)	0.004	9.40 (1.19–74.59)	0.034	1.76 (1.24–2.50)	0.002	9.73 (1.21–78.26)	0.032
rs10737680	1.22 (0.84–1.78)	0.303	2.01 (1.37–2.97)	<0.001	1.29 (0.86–1.93)	0.217	2.21 (1.46–3.34)	<0.001
rs2274700	1.22 (0.83–1.80)	0.309	2.26 (1.52–3.38)	<0.001	1.29 (0.85–1.94)	0.234	2.45 (1.60–3.74)	<0.001
rs1410996	1.21 (0.83–1.77)	0.318	2.08 (1.41–3.07)	<0.001	1.28 (0.86–1.92)	0.229	2.28 (1.51–3.44)	<0.001
rs380390	2.07 (1.40–3.05)	<0.001	8.83 (1.10–70.72)	0.040	2.19 (1.48–3.26)	<0.001	9.04 (1.11–73.53)	0.040
ARMS2
rs10490924	1.51 (1.07–2.12)	0.018	4.25 (2.94–6.13)	<0.001	1.62 (1.13–2.34)	0.009	4.90 (3.33–7.21)	<0.001
rs2736912	2.64 (0.69–10.06)	0.156	4.83 (1.31–17.83)	0.018	2.10 (0.54–8.12)	0.282	4.35 (1.17–16.22)	0.028
HTRA1
rs3793917	1.56 (1.03–2.06)	0.033	4.56 (3.14–6.64)	<0.001	1.58 (1.09–2.30)	0.016	5.31 (3.58–7.88)	<0.001
rs11200638	1.41 (1.01–1.99)	0.048	3.82 (2.65–5.49)	<0.001	1.51 (1.05–2.18)	0.027	4.42 (3.01–6.48)	<0.001

*  The OR_hetero and P values were calculated by comparing the heterozygosity for the risk allele with the homozygosity for the normal allele, and adjusted for age (continuous variable), sex, long-term residence, and family origin by logistic regression analysis.

†  The OR_homo and P values were calculated by comparing the homozygosity for the risk allele with the homozygosity for the normal allele, and adjusted for age (continuous variable), sex, long-term residence, and family origin by logistic regression analysis.

With the exception of rs3753394, all SNPs genotyped in CFH were included within three haplotype blocks (see Fig.). The frequencies, P values, and ORs for haplotypes with a frequency greater than 5% are shown in Table 6. There were three SNPs within the first block: rs7524776 (intron), rs551397 (intron), and rs800292 (exon, Val62Ile). The haplotypes TGC and TAT showed significant association with AMD (TGC OR = 1.52, P _corr < 0.001 for all AMD, and OR = 1.57, P _corr < 0.001 for advanced AMD; TAT OR = 0.64, P _corr < 0.001 for all AMD, and OR = 0.62, P _corr < 0.001 for advanced AMD). After permutation, these two haplotypes remained significant. Block 2 contained seven SNPs: rs1329424 (intron), rs1061170 (exon, Tyr402His), rs10801555 (intron), rs12124794 (intron), rs6695321 (intron), rs10733086 (intron), and rs10737680 (intron). Only the haplotype ACAAAAA predisposed significantly to AMD after permutation (OR = 1.92, P _corr = 0.001 for all AMD, and OR = 2.05, P _corr < 0.001 for advanced AMD). Block 3 contained four SNPs: rs2274700 (exon, Ala473Ala), rs3753396 (exon, Gln672Gln), rs1410996 (intron), and rs380390 (intron). The haplotypes TATG and CACC were associated significantly with AMD (TATG OR = 0.63, P _corr < 0.001 for all AMD, and OR = 0.60, P _corr < 0.001 for advanced AMD; CACC OR = 2.27, P _corr < 0.001 for all AMD, and OR = 2.39, P _corr < 0.001 for advanced AMD). 

Figure. 

 

LD between the polymorphism in the complement factor H gene. The D between any two single nucleotide polymorphisms is listed in the cross cells.

Table 6.  

 

Haplotype Analysis of the CFH Gene

Haplotype	Controls	All AMD Patients			Advanced AMD Patients
	Freq.	Freq.	P/P corr*	OR†	Freq.	P/P corr*	OR†
Block1: rs7524776, rs551397, rs800292
TGC	0.59	0.68	<0.001/<0.001	1.52	0.69	<0.001/<0.001	1.57
TAT	0.36	0.27	<0.001/<0.001	0.64	0.26	<0.001/<0.001	0.62
CAT	0.05	0.05	0.88/1.00	NS	0.05	0.90/1.00	NS
Block2: rs1329424, rs1061170, rs10801555, rs12124794, rs6695321, rs10733086, rs10737680
CTGAGTA	0.49	0.53	0.09/0.57	NS	0.54	0.051/0.43	NS
CTGTGTC	0.35	0.30	0.01/0.06	0.78	0.28	0.002/0.02	0.74
ACAAAAA	0.06	0.11	<0.001/ 0.001	1.92	0.11	<0.001/<0.001	2.05
Block3: rs2274700, rs3753396, rs1410996, rs380390
CGCG	0.48	0.53	0.03/0.22	1.22	0.53	0.03/0.23	1.23
TATG	0.44	0.34	<0.001/<0.001	0.63	0.33	<0.001/<0.001	0.60
CACC	0.06	0.11	<0.001/<0.001	2.27	0.11	<0.001/<0.001	2.39

 Only haplotypes with a frequency greater than 5% are shown. NS, not significant.

*   P _corr association analysis results from permutation test (number of permutations = 10,000).

†  Age, sex, long-term residence, and family origin adjusted ORs calculated by logistic regression analysis with PLINK software.

Discussion

Our study described the genotype characteristics of AMD in a Chinese population, which revealed a significant association between AMD and CFH, ARMS2, and HTRA1 gene variants. With the current sample size, we did not find a significant association between AMD and the genetic variants of C3, SERPING1, VEGF, CETP, LPL, LIPC, and TIMP3 genes. 

CFH has been identified by various studies as a major AMD susceptible gene in the Chinese, Japanese, and Caucasian populations, ^7,8,11,12,26 which is consistent with our study. The frequency of the risk allele of rs1061170, which has been identified as an important genetic variant in CFH for developing AMD in Caucasians, is lower in the Chinese population (10.9% for C allele in our study) compared to Caucasians (34.9% for C allele). ³⁰ This may explain why some studies, especially those with small sample size, failed to find the association between this SNP and AMD in Chinese populations. ^8,31 In our study, rs1061170 was associated significantly with AMD, though not as strong as in the Caucasian population. We also found out that other genetic variants in CFH, including rs551397, rs800292, rs1410996, rs2274700, rs380390, rs1329424, rs10801555, and rs10737680, were associated with AMD, and this is consistent with some previous studies. ^{7,8,23,26,32–34} Moreover, our study found that rs12124794 and rs10733086 in CFH were associated significantly with AMD. Both of these SNPs are intron and their molecular effects still are not clear. After adjusting for rs2274700, rs12124794 did not reach the significant level (P = 0.075). Meanwhile, after adjusting for rs1061170, rs10733086 did not reach the significant level (P = 0.723). Therefore, these two SNPs might not associate with AMD independently and they might just reflect the effect of the known SNPs that are in linkage disequilibrium with them. Furthermore, we found three haplotype blocks. Each of the first two haplotypes contained one SNP that encoded for a missense amino acid change, whereas 2 SNPs in block 3 encoded for synonymous changes. The haplotypes composed by the detrimental alleles of SNPs that had significant association with AMD in the single association analysis conferred a 1.52 (TGC), 1.92 (ACAAAAA), and 2.27 (CACC)-fold increased risk of AMD, whereas the haplotypes that contained the non-detrimental alleles had evident protective effect. Our results regarding the block containing rs551397 and rs800292, and another block containing rs2274700 and rs3753396 are consistent with the previous reports. However, there was a difference in haplotypes between Caucasian and Asian populations. In Caucasians, rs1061170 was in the same block with rs551397 and rs800292, while studies in an Asian population, which were consistent with our results, found that rs1061170 did not occur in the same block with rs551397 and rs80029. ^8,33  

In agreement with previous studies, we found that rs10490924 in ARMS2, and rs11200638 and rs3793917 in HTRA1 conferred strong susceptibility for AMD. The OR_homo values in our study are similar to those of studies in a Caucasian population and some Chinese populations. ^13,31,35 However, some other studies in Chinese populations revealed much higher ORs. ^9,10,36,37 DeWan et al reported that individuals with the risk-associated genotype were estimated to have a likelihood of wet AMD developing around 10 times (OR_homo = 11.14 for rs10490924, OR_homo = 10.00 for rs11200638) that of individuals with the wild-type genotype. ⁹ Yang et al found that individuals with the homo risk genotype were around 10-fold more likely to have AMD (OR_homo = 8.99 for rs10490924, OR_homo = 12.93 for rs11200638) than individuals without the risk genotype in a Han Chinese cohort. ¹⁰ This difference in the estimation of ORs might have resulted from the sample size or study design. In a word, no matter how great the OR, ARMS2 and HTRA1 conferred susceptibility for AMD. 

Our study also revealed that the tested SNPs of C3, SERPING1, VEGF, CETP, LPL, LIPC, and TIMP3 were not associated significantly with AMD in a Chinese population, although they were associated with AMD in previous genome-wide association studies (GWAS) or candidate studies in Caucasian populations. ^{14,17,19–22,38,39} In a Chinese population, Liu et al indicated that none of the SNPs they tested in C3 showed significant association with AMD, ⁷ while Pei et al showed that the G allele of C3 (rs2250656) might be a significant protective factor for neovascular AMD. ⁴⁰ Most studies in Chinese and Japanese populations did not provide evidence for the role of VEGF to AMD, ^25,26 and this is consistent with our results. However, Lin et al reported that VEGF-rs3025039 polymorphisms were associated dependently with wet AMD in the Taiwan Chinese population. ²⁷ This disagreement with our study possibly is due to the difference in inclusion and exclusion criteria of the two studies. In addition, our study in a Chinese population concerned the relationship about AMD, and genetic variants of CETP, LPL, LIPC, and TIMP3, though no significant association was disclosed. This disagrees with results of studies in Caucasians, ^20–22 which might be due to the fact that we only included a few SNPs of the above genes, possibly missing some other SNPs that could have an effect. Besides, with our sample size, the powers of the negative associations were less than 0.5, which was not adequate to state the negative association. Based on the above situations, further study with more SNPs and larger sample size may be needed. 

Strengths of our study include the large, nationwide, well-characterized Chinese patients, classification of AMD by standardized ophthalmologic examinations, and the reading center. Besides, to our knowledge, this is the first study to examine the association on the genetic variants of rs12124794, rs10733086, rs6695321, and rs7524776 in CFH; rs2736912 in ARMS2; rs408290, rs8112351, rs11569562, and rs7257062 in C3; and rs2303789, rs17231506, rs1532625, rs289744, and rs9939224 in CETP to AMD, as well as the first study to examine the association on the genetic variants of rs12678919 nearby LPL, rs10468017 nearby LIPC, and rs9621532 nearby TIMP3 to AMD in the Chinese population. 

In conclusion, we confirmed that CFH, ARMS2, and HTRA1 genetic variants were related to risk of AMD in the Chinese population. Further research is needed to explore the interaction of heredity and environment, so that the pathogenic mechanisms of AMD can be featured and intervened. 

Acknowledgments

The following collaborators recruited participants: Shaomin Peng (Department of Ophthalmology, The Second Affiliated Hospital of Harbin Medical University), Hao Cui (Department of Ophthalmology, The First Affiliated Hospital of Harbin Medical University), Zheli Liu (Department of Ophthalmology, The First Hospital of China Medical University), Liu Yang (Department of Ophthalmology, Peking University First Hospital), Youxin Chen (Department of Ophthalmology, Peking Union Medical College Hospital), Maonian Zhang (Department of Ophthalmology, The General Hospital of The People's Liberation Army), Xiaoguang Dong (Department of Ophthalmology, Shandong Eye Institute), Yusheng Wang (Department of Ophthalmology, Xijing Hospital), Qinghuai Liu (Department of Ophthalmology, Jiangsu Province Hospital), Yiqiao Xing (Department of Ophthalmology, Renmin Hospital of Wuhan University), Xiaodong Sun (Department of Ophthalmology, Shanghai First People's Hospital), Junjun Zhang (Department of Ophthalmology, West China Medical School-West China Hospital), Xiaoling Liu (Department of Ophthalmology, The Affiliated Eye Hospital of Wenzhou Medical College), Xiaohua Zhu (Department of Ophthalmology, The Second Xiangya Hospital of Central South University), and Xiaoling Liang (Department of Ophthalmology, Zhongshan Ophthalmic Center). Zhang Jian edited the manuscript. 

Footnotes

 Supported by the Mega-projects of Science Research for the 11th Five-Year Plan from The Ministry of Science and Technology of the People's Republic of China (2006BAI02B05), and the National Basic Research Program of China (973 Program, 2011CB510200). The funding organization had no role in the design or conduct of this research.

Footnotes

  ² These authors contributed equally to the work presented here and should therefore be considered equivalent authors.

Footnotes

 Disclosure: J. Tian, None; W. Yu, None; X. Qin, None; K. Fang, None; Q. Chen, None; J. Hou, None; J. Li, None; D. Chen, None; Y. Hu, None; X. Li, None