Macular Degeneration, Age-Related, 4

A number sign (#) is used with this entry because of evidence that susceptibility to age-related macular degeneration-4 (ARMD4) is conferred by variation in the CFH (134370) on chromosome 1q31.

For a phenotypic description and a discussion of genetic heterogeneity of age-related macular degeneration, see ARMD1 (603075).

Molecular Genetics

Klein et al. (2005) performed a genomewide scan of 96 cases and 50 controls for polymorphisms associated with ARMD. They identified a risk allele (rs380390) with a p value of less than 10(-7) that increased the likelihood of ARMD by a factor of 7.4 in homozygous individuals (95% confidence interval 2.9 to 19). Resequencing revealed a polymorphism in linkage disequilibrium with the risk allele representing a tyr402-to-his substitution of complement factor H (Y402H; 134370.0008). This polymorphism is in a region of CFH that binds heparin and C-reactive protein (see 123260).

Haines et al. (2005) independently identified the CFH Y402H polymorphism in 2 independent data sets. They found that the Y402H variant significantly increased the risk for ARMD with odds ratios between 2.45 and 5.57. Haines et al. (2005) suggested that this common variant likely explains about 43% of ARMD. Edwards et al. (2005) focused on the 1q25-q31 region, the ARMD1 locus, using 2 independent case-control populations. They found significant association (p = 4.95 x 10(-10)) with the CFH Y402H polymorphism. In their study, possession of at least 1 histidine at amino acid position 402 increased the risk of ARMD 2.7-fold, which they suggested may account for 50% of the attributable risk of ARMD.

Hageman et al. (2005) found a significant association between 2 variants in the CFH gene, Y402H and I62V (134370.0009), and ARMD in 2 independent cohorts comprising over 900 cases of ARMD and 400 matched controls. One haplotype containing Y402H conferred an odds ratio of 2.46 and 3.51, when present in the heterozygous or homozygous state, respectively. Several protective haplotypes were also identified. Hageman et al. (2005) found that CFH and its ligand C3b/iC3b colocalized in amyloid-containing substructural elements within macular drusen from patients with age-related macular degeneration. Real-time quantitative PCR assays showed that CFH gene products were generated from human retinal epithelial cells from individuals with and without macular degeneration. The findings suggested that variations within the CFH gene may attenuate complement inhibitory function and put retinal pigmented epithelial and choroidal cells at risk for alternative pathway-mediated complement attack.

Conley et al. (2005) investigated 15 candidate genes for ARMD by performing family and case-control genetic association studies. The Y402H variant in exon 9 of the CFH gene was significantly associated with ARMD in the case-control allele (P less than 0.0001), case-control genotype (P less than 0.0001), and case-control family (P less than 0.0001) tests. They proposed a potential role for multiple pathways in the etiology of ARMD, including pathways involved with fatty acid biosynthesis and the complement system.

Gotoh et al. (2006) found no association between the Y402H polymorphism and exudative ARMD among 146 Japanese patients and 105 Japanese controls. There was also no association between ARMD and several CFH haplotypes.

Li et al. (2006) examined 84 polymorphisms in and around CFH in 726 affected individuals (including 544 unrelated individuals) and 268 unrelated controls. In this sample, 20 of these polymorphisms showed stronger association with disease susceptibility than did the Y402H variant (134370.0008). Further, no single polymorphism could account for the contribution of the CFH locus to disease susceptibility. Instead, multiple polymorphisms defined a set of 4 common haplotypes (of which 2 were associated with disease susceptibility and 2 seemed to be protective) and multiple rare haplotypes (associated with increased susceptibility in aggregate). The results suggested that there are multiple disease susceptibility alleles in the region and that noncoding CFH variants play a role in disease susceptibility.

In a case-control study drawn from a U.S.-based population of European descent, Maller et al. (2006) identified a previously unrecognized common, noncoding variant in CFH (rs1410996; 134370.0016) that substantially increases the influence of this locus on ARMD. Li et al. (2006) also identified this variant in their study as the SNP showing the second strongest association with ARMD. In addition, Maller et al. (2006) strongly replicated the associations of common alleles in the C2 (613927) and CFB (138470) genes.

Using a population-based study among Latinos, Tedeschi-Blok et al. (2007) found that the CFH Y402H polymorphism was not a major risk factor for overall early ARMD, but may play a role in susceptibility to bilateral early ARMD.

Scott et al. (2007) examined the potential gene-environment interaction between cigarette smoking and the CFH Y402H polymorphism, 2 strong risk factors for ARMD. Effects of both Y402H genotype and cigarette smoking were stronger when comparing neovascular (grade 5) ARMD with grade 1 controls than when comparing all cases (grades 3-5) with grades 1-2 controls. Scott et al. (2007) concluded that cigarette smoking and Y402H are independent risk factors for ARMD and that both risk factors are associated more strongly with neovascular (wet) ARMD than with all other forms of ARMD combined.

Grassi et al. (2007) found a higher association of the Y402H polymorphism with the cuticular drusen phenotype of ARMD than with more typical ARMD cases.

In a case-control study with subjects originally recruited through the Cardiovascular Health Study (CHS) and the Age-Related Eye Disease Study (AREDS), Conley et al. (2006) found that CFH was significantly associated with ARMD in both cohorts (p less than 0.00001). A metaanalysis confirmed that the risk allele (Y402H) in the heterozygous or homozygous state (odds ratio = 2.4 and 6.2, respectively) conferred susceptibility. The LOC387715 gene (ARMS2; 611313) was also significantly associated with ARMD in both cohorts (p less than 0.00001) and a metaanalysis confirmed that the risk allele (A69S; 611313.0001) in the heterozygous and homozygous state (odds ratio = 2.5 and 7.3, respectively) conferred susceptibility. Both CFH and LOC387715 showed an allele dosage effect on the ARMD risk: individuals homozygous at either locus were at more than 2-fold risk compared to those heterozygous. Joint action of CFH and LOC387715 was best described by independent multiplicative effect without significant interaction in both cohorts. Interaction of both genes with cigarette smoking was insignificant in both cohorts.

Thompson et al. (2007) investigated the role of pigmentary abnormalities (PA) and geographic atrophy (GA) in ARMD. A previous genomewide scan was reanalyzed using the rate of change along the PA/GA scale. Evidence was found for linkage to 1q25, 5p13, 6q21-23, and 11q14 (p less than 0.01). The most significant peak was found on chromosome 1, near CFH (p = 6.20 x 10(-4)). Association analysis of CFH polymorphisms suggested that CFH might play a role in the development of pigmentary abnormalities and might modify the progression along the PA/GA scale.

Fritsche et al. (2013) identified association of the A allele of rs10737680 with increased risk of ARMD (OR 2.43; 95% CI 2.39-2.47; combined p = 1 x 10(-434)).

In an Amish ARMD family in which 3 affected sibs did not carry the common risk variants Y402H in CFH or A69S in ARMS2 Hoffman et al. (2014) performed exome sequencing and identified a missense mutation in the CFH gene, P503A (134370.0023), in the 3 affected sibs. The oldest sib was diagnosed with bilateral choroidal neovascularization at 75 years of age, whereas the other 2 affected sibs were diagnosed with large drusen in both eyes at ages 72 and 68 years, respectively. A sib who was initially considered to be unaffected at age 66 years but who was later diagnosed with ARMD at age 74 did not carry the P503A variant; however, he did carry the known Y402H risk variant. One of 2 remaining unaffected sibs, who carried both the P503A and A69S risk variants, presented with small drusen upon examination at age 72 years. Case-control analysis using self-reported affection status in an Amish sample population showed significant association of ARMD with P503A (p = 9.27 x 10(-13)). This sample consisted of 1,150 individuals connected into a single 13-generation pedigree, including 128 individuals with ARMD, 728 without ARMD, and 294 with no information. Results were consistent in the subanalysis of individuals with clinically confirmed ARMD (p = 5.21 x 10(-7)). Carriers of the variant could be traced back 4 generations to a shared common ancestor, suggesting a recent founder event. The P503A variant was not found in a non-Amish Caucasian dataset consisting of 1,456 cases and 791 controls.

Interaction of CFH with Mutation in Other Genes

In a cohort of 460 advanced ARMD cases and 269 age-matched controls and 57 archived ARMD cases and 18 age-matched non-ARMD controls, Tuo et al. (2006) found that a -6530C-G SNP in the ERCC6 gene (609413.0010) was associated with ARMD susceptibility, both independently and through interaction with an intronic G-C SNP in the CFH gene (134370.0008) previously reported by Klein et al. (2005). A disease odds ratio of 23 was conferred by homozygosity for risk alleles at both ERCC6 and CFH (G allele and C allele, respectively) compared to homozygosity for nonrisk alleles. Tuo et al. (2006) suggested that the strong ARMD predisposition conferred by the ERCC6 and CFH SNPs may result from biologic epistasis.

Hughes et al. (2006) found that a haplotype carrying deletion of the CFHR1 (134371) and CFHR3 (605336) genes was associated with decreased risk of ARMD, being present on 20% of chromosomes of controls and 8% of chromosomes of individuals with ARMD. The proteins encoded by these genes were absent in serum of homozygotes. The protective effect of the deletion haplotype could not be attributed to linkage disequilibrium with Y402H (134370.0008) and was replicated in an independent sample. The authors noted that the products of both the CFHR1 and CFHR3 genes are normally present in the circulation, where they have the potential to compete with CFH for C3 (120700) binding, and hypothesized that CFH produced from full-length transcript is beneficial and that other CFH-related proteins interfere with regulation of complement activity.

Schaumberg et al. (2007) studied the associations between the CFH Y402H and the LOC387715 A69S (611313.0008) variants with ARMD in a prospective, nested case-control study and investigated whether these variants interacted with modifiable risk factors. Participants with 1 or 2 copies of the Y402H variant were, respectively, 1.98 and 3.92 times more likely to develop ARMD, whereas the incidence rate ratios for 1 or 2 copies of A69S were 2.38 and 5.66, respectively. The fraction of ARMD cases attributable to these 2 variants was 63%. Subjects homozygous for both risk alleles had a 50-fold increased risk of ARMD, and cigarette smoking and obesity multiplied the risks associated with these variants.

Meyers et al. (2015) investigated whether the CFH Y402H allele or the ARMS2 A69S allele modified the risk for AMD in women with unhealthy lifestyles. They studied 1,663 women, aged 50 to 79 years, from the Carotenoids in Age-Related Eye Disease Study. Healthy lifestyle scores were assigned based on Healthy Eating Index scores, physical activity (metabolic equivalent of task hours/week), and smoking pack years. Odds of ARMD were 3.3 times greater in women with both low healthy lifestyle score (0-2/6) and high-risk CFH genotype (CC), relative to those with low genetic risk (TT) and high healthy lifestyle scores (4-6/6). There were no significant additive or multiplicative interactions for ARMS2 and lifestyle score. Meyers et al. (2015) concluded that having unhealthy lifestyles and 2 CFH risk alleles increased ARMD risk (primarily in the early stages), in an additive or greater (synergistic) manner. However, unhealthy lifestyles increased ARMD risk regardless of ARMD risk genotype.

In 711 individuals with ARMD and 1,041 controls, Raychaudhuri et al. (2010) reproduced associations at the CFH Y402H allele, using rs10801555 as a proxy, and CFH rs1410996, using rs10737680 as a proxy, but observed modest evidence for association with the CFHR1/CFHR3 deletion (p = 7.0 x 10(-21)). Logistic regression conditioned on rs10737680 resulted in substantially mitigated statistical strength for the protective effect of the CFHR1/CFHR3 deletion, suggesting that the CFHR1/CFHR3 deletion and rs10737680 were not entirely independent. Haplotype analysis demonstrated that both markers tagged a collection of low-risk haplotypes, but neither tagged all of them perfectly, suggesting that there could be 1 or more not-yet-identified variants that better explain disease risk. Raychaudhuri et al. (2010) favored the parsimonious explanations of a single functional allele in high correlation with rs10737680 acting on all protective haplotypes or of a risk variant acting on the intermediate risk haplotypes. In response, Hughes et al. (2010) noted that the finding of a lower statistical significance for the CFHR1/CFHR3 deletion than for rs10801555 or rs10737680 was a reflection of allele frequencies rather than effect size. The authors suggested that parsimonious explanations with the fewest functional elements are unnecessarily restrictive and noted that functional studies support a minimum of 3 factors.

To identify rare penetrant variants in the CFH locus, Raychaudhuri et al. (2011) phased genotypes for 20 common SNPs spanning the CFH-CFHR1-CFHR3 region and a common CFHR1-CFHR3 deletion in 711 individuals with advanced ARMD and 1,041 controls. They identified a rare high-risk haplotype ('H5') that lacked both the Y402H and rs10737680-rs1410996 risk alleles, but contained the R1210C substitution (134370.0017). Genotyping R1210C in 2,423 ARMD cases and 1,122 controls demonstrated high penetrance (present in 40 cases vs 1 control; p = 7.0 x 10(-6)) and an association with a 6-year-earlier onset of disease (p = 2.3 x 10(-6)). Because R1210C is known to cause familial renal disease (atypical hemolytic uremic syndrome; 235400), Raychaudhuri et al. (2011) assessed renal function in 17 unrelated R1210C heterozygotes with advanced ARMD but found no evidence of clinically significant renal dysfunction, although subclinical renal dysfunction was present (median calculated GFR of 62 mL/min). As ARMD patients in general have subclinical renal dysfunction, the authors compared renal function in the R1210C heterozygotes to that of 17 ARMD patients without R1210C who were matched for disease severity, age, and gender, but found no significant difference.