Purpose
To assess whether established and newly reported genetic variants, independent of known lifestyle factors, are associated with the risk of age-related macular degeneration (AMD) among women participating in the Women’s Health Initiative Sight Exam (WHI-SE) Genetic Ancillary Study.
Design
Multicenter case-control study.
Methods
One hundred and forty-six women with intermediate and late stages of AMD and 1269 subjects without AMD underwent ocular examinations and fundus photography to determine stage of AMD. Fourteen polymorphisms at or near 11 genes, including previously confirmed genes CFH , ARMS2/HTRA1 , C2 , C3 , and CFI ; recently reported AMD genes in the high-density lipoprotein cholesterol (HDL) pathway LIPC , ABCA1 , CETP , and LPL ; TIMP3/SYN3 , a known ocular gene recently linked with AMD; and APOE , were assessed using logistic regression analysis.
Results
After adjustment for demographic, behavioral, and other genetic factors, a protective effect was detected among TT carriers compared with non-carriers for the HDL pathway gene, LIPC rs493258, for intermediate and late AMD (OR [95% confidence interval]: 0.3 [0.2–0.7], P = .003). Variants in CFH rs1410996, ARMS2/HTRA1 A69S, and C3 R102G were significantly associated with an increased risk of AMD. Individuals with the homozygous CFI rs10033900 TT genotype had a 2.9 [1.2–7.2]-fold increased risk, and those with the CFH Y402H GG genotype had a 2.2 [1.0–4.8]-fold higher risk of developing AMD compared with non-carriers. APOE4 carriers may have a reduced risk of intermediate/late AMD (OR = 0.5 [0.3–0.9], P = .015. Suggestive associations were seen between AMD and the HDL pathway genes CETP and LPL.
Conclusion
In this unique national cohort of women, we found associations with established AMD-related genetic factors and the recently reported LIPC gene in the HDL pathway. These findings may help develop novel therapeutic targets to treat or delay the onset of the disease.
Age-related macular degeneration (AMD) is the leading cause of irreversible blindness among elderly individuals worldwide. The public health impact of this disease is rising since the elderly population is growing substantially. It has been forecast that cases of early AMD will increase from 9.1 million in 2010 to 17.8 million in 2050. Healthy lifestyles can reduce the prevalence and incidence of AMD. Modifiable risk factors linked to AMD over the past 2 decades include cigarette smoking, lack of physical activity, excessive weight, and insufficient nutritional antioxidant dietary and supplement intake.
There has also been compelling evidence for the role of genetic factors in AMD pathogenesis. Genetic variants associated with the prevalence of the disease have been identified in the complement pathway ( CFH , C2/CFB , C3 , and CFI ). There is also an association with a region containing several tightly linked genes on chromosome 10 ( LOC387715 , HTRA1 , ARMS2 ), although the function of those genes and variants is not fully understood. We have shown that a combination of these demographic, ocular, behavioral, and multi-loci genetic factors are highly predictive of AMD and its progression to the advanced stages, with a predictive ability of over 80% to discriminate progressors from non-progressors.
Recent genome-wide association (GWA) studies examining much larger numbers of individuals, and therefore with more statistical power, have confirmed all of the above AMD variants, and also have identified additional pathways associated with the disease, including one involved in the high-density lipoprotein cholesterol or reverse cholesterol transport pathway. Specifically, the hepatic lipase ( LIPC) gene was discovered to affect AMD susceptibility, which was corroborated by an independent GWA study. Both GWA studies implicate a number of other genes in high-density lipoprotein (HDL) metabolism. Furthermore, a susceptibility locus near the TIPM3 gene, a metalloproteinase involved in degradation of the extracellular matrix, previously shown to be mutated in cases of Sorsby fundus dystrophy, an early-onset maculopathy, has been implicated and corroborated. The apolipoprotein E ( APOE ) gene may also be associated with AMD but results are inconsistent, possibly because of the strong negative association of the E4 allele with longevity and increasing age.
We evaluated the relative contribution of novel and established genetic factors toward AMD susceptibility in our Genetic Ancillary Study involving a cohort of women enrolled in the Women’s Health Initiative Sight Exam (WHI-SE) study. This is a study within a set of clinical trials and an observational study, which was designed to test the effects of postmenopausal hormone therapy, diet modification, and calcium and vitamin D supplements on heart disease, fractures, and breast and colorectal cancer. The findings of the initial study have shown that treatment with conjugated equine estrogens (CEE) alone and CEE with progestin (CEE+P) does not affect early- or late-stage AMD. The WHI-SE study cohort provides a unique opportunity to study AMD in a large population of women who were examined and uniformly assessed with regard to their AMD status. Detection of AMD-associated genetic markers expands the understanding of mechanisms related to AMD, can help identify individuals who are at high risk of developing AMD, and could lead to novel targets for preventive and therapeutic interventions.
Material and Methods
Study Population
In the WHI-SE, 4288 women 65 years and older underwent fundus photography for the determination of signs of AMD. Participants were recruited from April 2000 to June 2002 at 21 clinical sites. In the WHI Hormone Therapy Trial, women with a uterus were randomized to treatment with CEE+P or placebo. Women without a uterus were randomized to treatment with unopposed CEE or placebo. The present study is restricted to individuals of European descent. Non-white individuals were excluded because of the small sample size and because the distribution of advanced AMD in other populations differs considerably from that among white populations.
Phenotype Definitions
Fundus photography was performed using a standard protocol as described elsewhere. The sample of 1717 women was selected based on availability of adequate ocular photographs. AMD severity was defined as no AMD (N = 1269), minimal early AMD (N = 25), moderate early (N = 277) and severe early AMD (herein called intermediate; N = 98), and advanced AMD (geographic atrophy or dry, N = 13; and neovascular or wet, N = 35), as described elsewhere. Mild, moderate, and advanced cases were age-matched to controls originally, but mild cases were excluded from the analyses to evaluate more clinically and visually significant disease, so the final sample sets were not matched. We analyzed late AMD as 1 case group and the combined phenotypes of intermediate and late AMD as another group. Self-administered questionnaires were used to collect information on race, age, education, smoking status, and diabetes status, as well as lipid-lowering, antihypertensive, and diabetes medication history. At the first screening examination, body weight and height were measured to assess body mass index (BMI), calculated as the weight in kilograms divided by the square of the height in meters.
Single Nucleotide Polymorphism selection and Genotyping
We reviewed genome-wide association and candidate gene studies to select 14 single nucleotide polymorphisms (SNPs) at or near 11 genes previously associated with AMD in populations of European ancestry ( Table 1 ). Genotyping was performed with Taqman assays (Applied Biosystems; Supplemental Table , available at AJO.com ) on a 7900HT (Applied Biosystems, Carlsbad, California, USA) at Tufts Clinical and Translational Science Institute and Jean Mayer USDA Human Nutrition Research Center on Aging, Tufts University. The CFH polymorphism rs1061170 was genotyped using a custom assay (forward primer 59′-GGTCCTTAGGAAAATGTTATTTTCCTTATTTGG-39′, reverse primer 59′-GGCAGGCAACGTCTATAGATTTACC-39′, G allele probe (VIC): 59′-TTTCTTCCAT G ATTTTG-39′, and A allele probe (FAM): 59′-TTCTTCCAT A ATTTTG-39′). The minimum call rate was 96.9% and the average consensus rate from 20 duplicates was 100%. Seven samples were sequenced and the concordance with the initial genotype calls was 100%. All SNPs conformed to Hardy-Weinberg equilibrium. The allelic variants derived from the 2 SNPs within APOE , rs429358 and rs7412, and referred to as E2, E3, and E4, are differentiated on the basis of cysteine-arginine residue interchanges at positions 112 and 158 in the amino acid sequence and give rise to 6 bi-allelic haplotypes (E3/E3, E3/E4, E2/E3, E4/E4, E2/E4, and E2/E2).
| dbSNP Identifier a | Gene | Chromosomal Position | Nucleotide Substitution | Amino Acid Substitution |
|---|---|---|---|---|
| Established genetic variants | ||||
| rs1061170 | CFH | 1q32 | CT | Y402H |
| rs1410996 | CFH | 1q32 | CT | |
| rs10490924 | ARMS2/HTRA1 | 10q26.13 | GT | A69S |
| rs9332739 | C2 | 6p21.3 | GC | E318D |
| rs2230199 | C3 | 19p13.3 | CG | R102G |
| rs10033900 | CFI | 4q25 | CT | |
| Newly reported AMD variants | ||||
| rs493258 | LIPC | 15q21–q23 | CT | |
| rs10468017 | LIPC | 15q21–q23 | CT | |
| rs9621532 | TIMP3, SYN3 | 22q12.3 | AC | |
| Suggestive genetic variants | ||||
| rs429358 | APOE | 19q13.2 | CT | C130R |
| rs7412 | APOE | 19q13.2 | CT | R176C |
| Potential HDL pathway variants | ||||
| rs1883025 | ABCA1 | 9q31.1 | CT | |
| rs3764261 | CETP | 16q21 | CA | |
| rs12678919 | LPL | 8p22 | AG |
a Database of single nucleotide polymorphisms (SNPs): http://www.ncbi.nlm.nih.gov/projects/SNP/ .
Statistical Analysis
Participants with intermediate and advanced AMD were individually compared to the control group of subjects with no AMD. First, univariate analysis was performed for each variant. For common variants, tests for trend for the number of risk alleles (0, 1, or 2) were calculated; for rare variants, carriers of 1 or 2 alleles were compared to non-carriers (dominant model). Then, multivariate logistic regression analysis was conducted to identify significant factors collectively associated with disease status, including age (continuous), education (high school diploma or less vs college or higher), cigarette smoking (never, past, current), treatment assignment in the clinical trial (CEE, CEE+P, or placebo for CEE or CEE+P), BMI (continuous), and 12 genetic variants. To account for potential survival effect of the APOE genotype an APOE- age interaction term was included in the fully adjusted model. A 1-tailed P < .05 was considered statistically significant. All analyses were performed using SAS/STAT and SAS/Genetics software version 9.1 (SAS Institute, Inc, Cary, North Carolina, USA).
Results
The mean ages of the subjects and controls were 74.5 ± 4.7 and 73.6 ± 4.7, respectively. Table 2 displays the unadjusted associations with intermediate and late AMD. Controls were more likely to have education beyond high school. There was a nonsignificant trend for cases to be older, to be former or current smokers, and to have higher BMI. Risk genotypes for C3 R102G, CFH rs1410996 and Y402H, CFI rs10033900, and ARMS2/HTRA1 A69S were significantly higher in cases, while the frequency of the homozygous TT genotype for LIPC rs493258 and rs10468017 (in strong linkage disequilibrium: D9′ = 0.95, r =0.63) was higher in controls. For APOE , the E4 allele tended to be higher in controls for the combined case group.
| Control, N (%) | Late AMD | Intermediate and Late AMD | |||||
|---|---|---|---|---|---|---|---|
| N (%) | OR (95% CI) | P | N (%) | OR (95% CI) | P | ||
| Total patients | 1269 | 48 | 146 | ||||
| Age (years) | |||||||
| <70 | 275 (21.7) | 6 (12.5) | 1.0 | 22 (15.1) | 1.0 | ||
| 70+ | 994 (78.3) | 42 (87.5) | 1.9 (0.8–4.6) | .13 | 124 (84.9) | 1.6 (0.97–2.5) | .07 |
| Education | |||||||
| ≤High school | 500 (39.6) | 26 (54.2) | 1.0 | 67 (46.2) | 1.0 | ||
| >High school | 762 (60.4) | 22 (45.8) | 0.6 (0.3–0.99) | .046 | 78 (53.8) | 0.8 (0.5–1.1) | .13 |
| Smoking | |||||||
| Never | 709 (56.5) | 22 (46.8) | 1.0 | 81 (56.6) | 1.0 | ||
| Past | 471 (37.5) | 22 (46.8) | 1.5 (0.8–2.8) | .18 | 53 (37.1) | 1.0 (0.7–1.4) | .94 |
| Current | 75 (6.0) | 3 (6.4) | 1.3 (0.4–4.4) | .69 | 9 (6.3) | 1.1 (0.5–2.2) | .89 |
| BMI | |||||||
| <25 | 348 (27.6) | 11 (22.9) | 1.0 | 39 (26.7) | 1.0 | ||
| 25–29 | 454 (36.0) | 17 (35.4) | 1.2 (0.5–2.6) | .62 | 52 (35.6) | 1.0 (0.7–1.6) | .83 |
| ≥30 | 458 (36.4) | 20 (41.7) | 1.4 (0.7–3.0) | .36 | 55 (37.7) | 1.1 (0.7–1.7) | .67 |
| Treatment group b | |||||||
| Placebo for CEE + P | 378 (29.8) | 18 (37.5) | 1.0 | 46 (31.5) | 1.0 | ||
| CEE + P | 433 (34.1) | 11 (22.9) | 0.5 (0.2–1.1) | .11 | 38 (26.0) | 0.7 (0.5–1.1) | .16 |
| CEE | 216 (17.0) | 10 (20.8) | 1.0 (0.4–2.1) | .94 | 31 (21.2) | 1.2 (0.7–1.9) | .50 |
| Placebo for CEE | 242 (19.1) | 9 (18.8) | 0.8 (0.3–1.8) | .55 | 31 (21.2) | 1.1 (0.6–1.7) | .84 |
| ESTABLISHED GENETIC VARIANTS | |||||||
| CFH :rs1061170 (Y402H) | |||||||
| TT | 503 (40.0) | 8 (16.7) | 1.0 | 29 (19.9) | 1.0 | ||
| CT | 592 (47.1) | 28 (58.3) | 3.0 (1.3–6.6) | 70 (48.0) | 2.1 (1.3–3.2) | ||
| CC | 163 (13.0) | 12 (25.0) | 4.6 (1.9–11.5) | .0005 a | 47 (32.2) | 5.0 (3.0–8.2) | <.0001 a |
| CFH :rs1410996 | |||||||
| TT | 226 (18.0) | 3 (6.3) | 1.0 | 13 (8.9) | 1.0 | ||
| CT | 631 (50.2) | 14 (29.2) | 1.7 (0.5–5.9) | 39 (26.7) | 1.1 (0.6–2.1) | ||
| CC | 399 (31.8) | 31 (64.6) | 5.8 (1.8–19.4) | <.0001 a | 94 (64.4) | 4.1 (2.2–7.5) | <.0001 a |
| ARMS2/HTRA1 rs10490924 (A69S) | |||||||
| GG | 806 (63.9) | 17 (35.4) | 1.0 | 59 (40.4) | 1.0 | ||
| GT | 406 (32.2) | 23 (47.9) | 2.7 (1.4–5.1) | 74 (50.7) | 2.5 (1.7–3.6) | ||
| TT | 49 (3.9) | 8 (16.7) | 7.7 (3.2–18.9) | <.0001 a | 13 (8.9) | 3.6 (1.9–7.1) | <.0001 a |
| C2 :rs9332739 (E318D) | |||||||
| GG | 1163 (92.4) | 43 (89.6) | 1.0 | 140 (95.9) | 1.0 | ||
| CC+CG | 96 (7.6) | 5 (10.4) | 1.4 (0.5–3.6) | .48 | 6 (4.1) | 0.5 (0.2–1.2) | .13 |
| C3 :rs2230199 (R102G) | |||||||
| CC | 808 (64.1) | 26 (54.2) | 1.0 | 79 (54.1) | 1.0 | ||
| CG | 398 (31.6) | 14 (29.2) | 1.1 (1.6–2.1) | 52 (35.6) | 1.3 (0.9–1.9) | ||
| GG | 54 (4.3) | 8 (16.7) | 4.6 (2.0–10.7) | .01 a | 15 (10.3) | 2.8 (1.5–5.3) | .002 a |
| CFI :rs10033900 | |||||||
| CC | 348 (27.6) | 9 (18.8) | 1.0 | 34 (23.3) | 1.0 | ||
| CT | 623 (49.4) | 20 (41.7) | 1.2 (0.6–2.8) | 68 (46.6) | 1.1 (0.7–1.7) | ||
| TT | 289 (22.9) | 19 (39.6) | 2.5 (1.1–5.7) | .016 a | 44 (30.1) | 1.6 (1.0–2.5) | .06 a |
| NEWLY REPORTED AMD VARIANTS | |||||||
| LIPC :rs493258 | |||||||
| CC | 354 (28.8) | 19 (39.6) | 1.0 | 50 (35.0) | 1.0 | ||
| CT | 597 (48.6) | 26 (54.2) | 0.8 (0.4–1.5) | 80 (56.0) | 0.9 (0.7–1.4) | ||
| TT | 277 (22.6) | 3 (6.3) | 0.2 (0.1–0.7) | .01 a | 13 (9.1) | 0.3 (0.2–0.6) | .002 a |
| LIPC :rs10468017 | |||||||
| CC | 650 (52.7) | 34 (70.8) | 1.0 | 89 (63.1) | 1.0 | ||
| CT | 487 (39.5) | 11 (22.9) | 0.4 (0.2–0.9) | 46 (32.6) | 0.7 (0.5–1.0) | ||
| TT | 97 (7.9) | 3 (6.3) | 0.6 (0.2–2.0) | .04 a | 6 (4.3) | 0.5 (0.2–1.1) | .01 a |
| TIMP3/SYN3 :rs9621532 | |||||||
| AC+CC | 131 (10.5) | 5 (10.4) | 1.0 | 16 (11.1) | 1.0 | ||
| AA | 1114 (89.5) | 43 (89.6) | 1.0 (0.4–2.6) | .98 | 128 (88.9) | 0.9 (0.5–1.6) | .83 |
| SUGGESTIVE GENETIC VARIANT | |||||||
| APOE | |||||||
| E2E2+E2E3+E3E3 | 928 (73.7) | 39 (81.3) | 1.0 | 120 (82.2) | 1.0 | ||
| E2E4+E3E4+E4E4 | 331 (26.3) | 9 (18.8) | 0.6 (0.3–1.4) | .25 | 26 (17.8) | 0.6 (0.4–0.9) | .03 |
| POTENTIAL HDL PATHWAY VARIANTS | |||||||
| ABCA1 :rs1883025 | |||||||
| CC | 647 (53.0) | 29 (60.4) | 1.0 | 77 (55.0) | |||
| CT+TT | 573 (47.0) | 19 (39.6) | 0.7 (0.4–1.3) | .32 | 63 (45.0) | 0.9 (0.7–1.3) | .66 |
| CETP :rs3764261 | |||||||
| CC | 571 (45.8) | 20 (41.7) | 1.0 | 58 (40.6) | 1.0 | ||
| CA | 544 (43.6) | 21 (43.8) | 1.1 (0.6–2.1) | 67 (46.9) | 1.2 (0.8–1.8) | ||
| AA | 132 (10.6) | 7 (14.6) | 1.5 (0.6–3.7) | .41 a | 18 (12.6) | 1.3 (0.8–2.4) | .22 a |
| LPL :rs12678919 | |||||||
| AA | 1009 (81.3) | 42 (87.5) | 1.0 | 120 (84.5) | 1.0 | ||
| AG+GG | 232 (18.7) | 6 (12.5) | 0.6 (0.3–1.5) | .28 | 22 (15.5) | 0.8 (0.5–1.3) | .35 |
b Treatment randomization groups defined as: Placebo for CEE + P = placebo for estrogen replacement therapy (ERT) + progestin; CEE + P = ERT + progestin; CEE = ERT only; Placebo for CEE = placebo for ERT only.
In the multivariate analysis controlling for all demographic, environmental, and genetic factors among subjects with complete data for all variables ( Table 3 ), age and BMI were related to AMD, and subjects were more likely to be smokers (not significant). Genetic polymorphisms that remained significant after the adjustment for all demographic, lifestyle, and genetic factors included CFH rs1410996, ARMS2/HTRA1 A69S, and C3 R102G regardless of disease severity. In addition, a protective effect was seen for the minor allele (T) of LIPC rs493258 with the lower AMD risk seen for carriers of 2 minor alleles compared to non-carriers (odds ratio [OR], 95% confidence interval [CI]: 0.3, 0.2–0.7, P = .003 for individuals with intermediate and late AMD and OR, 95% CI: 0.3, 0.1–1.1, P = .06 for individuals with late wet and dry AMD). Also, CFI rs10033900 TT carriers had 2.9-fold higher odds of developing late AMD (95% CI 1.2–7.2; P = .02), whereas CFH Y402H GG carriers had a 2.2-fold higher odds of developing intermediate or late AMD (95% CI 1.0–4.8; P = .04) compared to non-carriers. In addition, CFH rs1410996 was significant in the multiple degree of freedom test for the overall effect ( P = .02 and P = .002 for advanced and intermediate and late AMD, respectively) independent of CFH Y402H. APOE E4 carriers were less likely to be diagnosed with intermediate and late AMD when compared to non-carriers (OR, 95% CI: 0.53, 0.32–0.89, P = .015), and the APOE E4-age interaction term was not significant. A suggestive association was observed between LPL rs12678919 and late AMD; however, this was not statistically significant (OR 0.5, 95% CI: 0.2–1.2, P = .10) for carriers of 1 or 2 G alleles compared to non-carriers. A nonsignificant association was suggested for the A allele of CETP (OR 2.3 for AA vs CC) comparing late AMD to controls.
