some primate lineages, it may encode a protein expressed mainly in placenta, but only weakly in the retina [59]. Using human retinal RNA, Kanda et al. expressed the purported LOC387715 protein in COS-1 cells. The expressed protein localized to the outer wall of mitochondria [66]. In comparison, Wang et al. used immunofluorescence and immunoblot analysis to show that the LOC387715 protein localized to the cytosol, not the mitochondria [67]. Further research will hopefully elucidate a possible functional mechanism by which this genetic variant confers risk for AMD.

HTRA1

The HTRA1 gene (also known as PRSS11) lies approximately 6.1 kb downstream of the LOC387715 gene. Its association to AMD has been established in a number of independent cohorts [68–75]. The SNP rs11200638 is found in the promoter region of HTRA1 and tagged a major disease haplotype. This variant’s A risk allele (G625A) alters a conserved binding element, AP2/SRF [70]. In a cohort of 581 AMD patients and 309 normal controls, it was found that those heterozygous or homozygous for the A allele were at a significantly increased risk for AMD (OR 1.83, 95% CI 1.25–2.68; OR 7.29, 95% CI 3.18–16.74, respectively) [74]. Similar numbers were obtained in a meta-anal- ysis of 14 studies (OR 2.13, 95% CI 1.9–2.39; OR 6.92, 95% CI 5.74–8.34, respectively) [76]. The population-attributable risk for HTRA1 ranges between 22% and 53% across studies [37].

The HTRA1 gene’s protein, HtrA serine protease 1, is an inhibitor of the angiogenesis regulator transforming growth factor-b (TGF-b). It also appears to increase the degradation of extracellular matrix (ECM) proteins by modifying the activity of matrix metalloproteinase enzymes [77]. It could be possible that overexpression of HTRA1 weakens Bruch’s membrane and promotes angiogenesis, creating the phenotypic CNV picture in advanced AMD. This hypothesis is strengthened by a study indicating its association specifically with CNV in a Chinese population [70] and data showing an increased

association with large CNV (³4 disc areas, OR 3.4, 95% CI 1.2–9.5) [57]. Researchers implicating HTRA1 in GA have shown a threefold increased expression of HTRA1 in the retinal pigmented epithelium of patients with the rs11200638 risk variant [74] and evidence of the HTRA1 protein in the drusen of AMD patients [70, 74]. Together, these studies add weight to the claim that HTRA1 has a significant role in AMD risk.

Although research disagrees about transcript expression levels of HTRA1 and LOC387715 in AMD patients, haplotype analysis has linked both genes. Yang et al. showed a synergistic relationship between the reported in/del variant of LOC387715 [64] and the A risk allele of rs11200638 in HTRA1. In-vitro experiments showed either was insufficient to generate increased expression of HTRA1. However, a two-fold increase in expression was exhibited when a disease haplotype was used, which included both risk variants (Fig. 1.4). Animal invivo experiments and mRNA expression levels in human placentas with risk haplotypes confirmed this association [65]. The relationship between LOC387715 and HTRA1 highlights the importance of haplotype examination in complex diseases such as AMD.

Pearl

HTRA1 is the most significant AMD susceptibility gene on chromosome 10q26.

Other Genetic Variants

Apolipoprotein E

Apolipoprotein E (ApoE) is a glycoprotein involved in lipid transport and lipid homeostasis in the central nervous system [79]. In the eye, it is found in the RPE, outer segments of photoreceptors, Bruch’s membrane, and drusen. Of the different isoforms of ApoE, the E4 allele appears to confer a protective effect, providing a twoto threefold decrease in AMD, while the E2 allele may be associated with AMD risk [80–84].

Fig. 1.4 Effects of the in/del variants and rs11200638 on HTRA1 expression in cultured human RPE cells and mouse RPE in vivo using luciferase activity. (a) Schematic diagram of constructs for luciferase reporter assays. L = long promoter construct. S = short promoter construct. A/G represents the allele at rs11200638. Black dot indicates in/del variant. (b) Luciferase activity in human

cultured RPE of different HTRA1 promoter sequences corresponding to risk (MT) and wild-type (WT) alleles of in/del and rs11200638 (WT(L-A), WT(L-G), (MT(L + in/ del), (WT(L + in/del), WT(S) and MT(S)) (c) Luciferase activities in mouse RPE in vivo corresponding to different HTRA1 promoter reporter constructs. Error bars represent the mean (6 SEM) (Reprinted from [78])

Although subsequent papers do not share similar results [85, 86], one meta -analysis found that having one E4 allele provided up to a 38% reduction in the risk of AMD [87].

Fibulin 5

Fibulin 5 is an extracellular matrix protein found in Bruch’s membrane. Patients with AMD have a faster degradation of Bruch’s membrane as they age, possibly increasing the risk for extension of blood vessels across this membrane and progression of CNV [88]. In a case-control Iowa cohort, Stone et al. found that missense variations in the gene encoding Fibulin 5 were statistically associated with AMD [89]. Lotery et al. showed that

reduced Fibulin 5 secretion was associated with AMD [90]. Together, this data points towards the hypothesis that variations in the Fibulin 5 protein may weaken Bruch’s membrane, thus contributing to the AMD phenotype. However, this association has not been consistently replicated in further research.

Hemicentin-1

Hemicentin 1 is another fibulin protein found in Bruch’s membrane. Family-association studies identified the gene encoding Hemicentin 1 as a candidate gene for AMD [25]. However, further studies in larger populations have not found this association [91, 92].