The Genetics of Age-Related Macular Degeneration

Introduction

Age-related macular degeneration runs in families. Half of children who have 1 parent with wet AMD acquire the same condition, while only 12% of those who do not have an affected parent do. While this is an old observation, recent genetic research has identified the specific variations in DNA that are associated with an increased risk for age-related macular degeneration. Most variations are SNPs that occur in these 4 categories of genes:

• Complement factors (inflammatory)
• Cholesterol metabolizing enzymes
• Energy metabolism genes (oxidative stress)
• Extracellular matrix metabolic pathway

The genetics fit with a general model of AMD where oxidative damage triggers a heightened inflammatory response followed by vascular proliferation, bleeding, scarring and blindness.

Complement factors and AMD

The complement system is a primitive enzymatic cascade that has evolved to eliminate pathogens from the blood stream by disrupting their membranes through pore formation. It functions both in an antibody-directed fashion ("the Classical Pathway") and independent of specific antigenic recognition ("the Alternate Pathway"). Subtle functional differences in the inherited DNA sequence of complement components are found in some patients with AMD cause macular degeneration.

A. Complement factor H and AMD

The first genetic association of a complement pathway gene variation with AMD appeared in 2005 when the Complement Factor H (CFH) gene on chromosome 1q31 was linked to this disease. CFH is a natural inhibitor of C3 activation, which is key to activation of the entire system. It is synthesized in the retinal pigment epithelium (RPE) and is present within drusen. The CFH single nucleotide polymorphism (SNP) rs1061170 (T/C) was the first mutation discovered to be associated with disease. Further refinement of the association of the broader CFH locus with AMD has shown that this site may not itself be etiologically important but is linked to yet-to-be identified elements that affect the function of CFH or closely associated genes. Extended CFH haplotypes (multiple polymorphic markers found on the same strand of DNA) are independent of the rs1061170 polymorphism and are superior predictors of AMD risk. This presumably results from the improved ability of many markers to interrogate a larger part of this gene and similar adjacent genes. AMD Researchers have grouped CFH polymorphisms to specify 2 common risk haplotypes (H1 and H3), several rare risk haplotypes (H5-H8) and 2 that specify a reduced risk (H2 and H4). The odds ratio for having AMD for individuals having a risk haplotype on each copy of chromosome 1 compared to 2 low risk haplotypes is approximately 17.

B. Complement Component 3 (C3) and AMD

C3 is a convergence point of all complement pathways. Polymorphisms in the C3 gene predict increased risk of AMD development. A study of Caucasian subjects from the UK and other groups first identified the rs2230199 (C/G) polymorphism to confer risk. The odds ratio for individuals having the heterozygous form is 1.7, while for those having the risk allele in the homozygous form (G/G) it is 2.6.

C. Complement Factor I (CFI)

The CFI gene functions to cleave and inactivate the activated complement components, including activated C3. C3 inactivation by CFI is regulated by CFH. CFH acts as a cofactor for CFI-mediated cleavage of activated C3. A SNP, rs10033900, found on chromosome 4 within the CFI gene does not change the protein sequence of the gene, but is thought to regulate its function by affecting the amount or activity of the CFI enzyme.

D. Other complement components and AMD

Several polymorphisms in other complement factors have been identified. Polymorphisms on chromosome 6 within the adjacent complement factor B (CFB) and Complement Component 2 genes (C2) are associated with AMD in some reports. A polymorphism within the Serping 1 gene, a naturally occurring C1 inhibitor is reported by one group to occur more often in patients with AMD, a result which has not been confirmed by others.

Cholesterol Metabolizing Enzymes

Drusen contain cholesterol, partially contributing to their appearance. While a relationship between AMD and circulating levels of cholesterol has been difficult to demonstrate, SNPs in the enzymes that handle cholesterol metabolism are associated with AMD risk. These include the hepatic lipase gene (LIPC), the lipoprotein lipase gene (LPL), cholesterol ester transferase gene (CETP) and the ABC-binding cassette A1 (ABCA1) gene. When cholesterol levels fall, a synthetic process (anabolic) is initiated through production of lipoprotein lipase (LPL) and the ATP binding Cassette Transporter A1 (ABCA1) which mediates the efflux of cholesterol and phospholipids to lipid-poor apolipoproteins (apo-A1 and apoE), which in turn forms nascent high-density lipoproteins. Lipoprotein lipase has the dual function of triglyceride hydrolase and ligand/bridging factor for receptor-mediated lipoprotein uptake. This results in the presentation of unesterified colesterol (UC) and cellular phospholipid (PL) to Lecithin-cholesterol acyltransferase (LCAT, also called phosphatidylcholine-sterol O-acyltransferase) which generates water soluble cholesterol that becomes packaged into HDL lipoprotein particles. As cholesterol levels rise, catabolic mechanisms come into play involving the transfer of cholesterol ester into lower density lipoprotein particles enzymatically mediated by cholesterol ester transfer protein (CETP). LIPC has the dual functions of triglyceride hydrolase and ligand/bridging factor for receptor-mediated lipoprotein uptake leading to the conversion of IDL to LDL.

How variations in genes mediating cholesterol metabolism contribute to AMD risk is not understood. Some studies have failed to identify a relationship between serum HDL levels and AMD while others have. An intriguing possibility is the identified role of HDL as the major lipoprotein transporter of the carotenoids; lutein, mesozeaxanthin and zeaxanthin. Reduced dietary intake of these is associated with an increased risk of AMD. Variation in the movement of these into the retina by HDL may cause AMD. Changes in HDL delivery of carotenoids may be a possible mechanism by which these polymorphisms affect AMD risk.

Oxidative Stress

Chronic unopposed oxidative stress may contribute to AMD-associated retinal damage. All cells maintain a reducing environment that is preserved by enzymes that require metabolic energy. Disturbances in this balance can cause toxic effects through the production of peroxides and free radicals that damage cellular components such as proteins, lipids and DNA. Most of these oxygenderived free radicals are constantly generated by normal aerobic metabolism and the damage they cause to cells must be continuously mitigated. Under extreme levels of oxidative stress, ATP depletion ensues resulting in cellular necrosis. For cells that are renewed slowly or not at all, such as neurons and retinal photoreceptors, these processes may lead to degenerative diseases such as AMD.

Predisposition to oxidative cell injury may result from the inheritance of functional variants of molecules that participate in the initiation or repair of oxidative stress. The AMD-associated locus on chromosome 10q26 contains the Age Related Maculopathy Susceptibility 2 (ARMS2) gene, which is expressed in retinal mitochondria. The rs10490924 G/T polymorphism in this site tags a large linkage block associating with AMD. It is in absolute linkage disequilibrium with a small insertion/deletion polymorphism. This variant form (indel) results from the deletion of 448 bases of DNA and the insertion of 54 different bases in its place. This deletion includes a very important piece of the gene, namely the poly adenylation signal of the ARMS2 gene. The poly adenylation signal marks the end of a gene and causes it to be more stable. The deletion associated with AMD results in absent protein because of gene instability. The del443ins54 variant confers a 2.9 fold increase in AMD risk when found in the heterozygous (one "risk" and one "good" gene copy) form and 8.1 in the homozygous form (two "risk" genes). The ARMS2 protein is usually found in the mitochondria and localizes to the ellipsoid region of the photoreceptors suggesting a role in energy metabolism and indirectly, oxidative stress.

A variation in the DNA found within the mitochondria DNA itself is also associated with AMD. The mitochondria are key to retinal oxidative stress through their role in energy metabolism. Each cell has thousands of mitochondria with identical 16,500 bp haploid sequences. The human mitochondrion codes for 37 proteins that function largely in aerobic cellular respiration. Within the mitochondrial genome, 144-catalogued sites of genetic variation exist with frequency greater than 1%. A mitochondrial SNP at position 4917 (A → G) is observed more frequently in patients with vision loss from AMD than in controls as measured in a study population of 1547. The G polymorphism is associated with a ratio of risk of 2.16 compared to the A SNP. Additional risk associated polymorphisms within mitochondria DNA contribute to disease risk supporting the role of energy metabolism in the etiology of AMD.

Extra Cellular Matrix

Variations in tissue inhibitor of metalloproteinase 3 (TIMP3) gene predict for the development of wet AMD. TIMP3 belongs to a group of peptidases involved in degradation of the extracellular matrix (ECM). The extracellular matrix is present in the intercellular spaces and consists of gels of polysaccharides and fibrous proteins, which act as a buffer against shear and compression forces. Mutations in the TIMP3 gene have been linked to Sorsby fundus dystrophy (SFD), an autosomal dominant inherited retinal degenerative disease that leads to blindness. SFD is clinically similar to AMD as it commonly manifests as a hemorrhagic maculopathy secondary to choroidal neovascularization. The age of onset of maculopathy is much earlier than AMD, with many affected individuals presenting in the third decade of life. How TIMP3 affects intracellular matrix remodelling and vascular neogenesis remains to be elucidated. It is known that metalloproteinases are powerful inducers of vascular growth, so potentially lack of inhibition may result in the clinical manifestations of hemorrhagic maculopathy.

Use of Genetics to Predict AMD Progression Risk

A method of interpreting genetic risk for AMD must allow for multiple genetic risk factors that vary in strength and may be present in 1 or 2 copies in an individual. Conceptually a risk assessment tool would include the all these features to evaluate the risk of an individual. This is illustrated below:

This simplified example shows risk alleles at 3 genes for 2 individuals. There are 2 copies of each risk allele (the same color) and can either be the form that is associated with AMD or not. Some gene alleles are powerful predictors (such as the "square" gene risk allele) and others less so (such as the "triangle" gene risk allele). Having the non-risk version of any gene does not contribute to risk. Since the gene alleles function independent of each other, the total amount of risk can be determined as the product of the risk given by individual alleles. For the "grey" person the total risk score is 1,200 while for the "blue" person has a much lower risk score (180).

Macula Risk measures over a dozen risk markers and has developed a risk prediction scoring system similar conceptually to the example provided. Because it measures all the known risk markers we believe that the scoring system is the most accurate way of predicting disease risk.

Understanding the risk for AMD progression in an individual is the most important step in designing a personalized approach to disease surveillance, early intervention and sight preservation.