Late stage age-related macular degeneration is classified as dry or wet according to characteristic features of these two pathogenic modes. The dry form, which in its end stage is termed geographic atrophy, is characterized by localized re­gions of atrophy of the retinal pigment epithelium with associated overlying regions of photo­receptor cell death. The wet form, also called neovascular AMD, involves the growth of new blood vessels from the choroid capillary network across the Bruch's membrane/RPE interface into the neural retina, which results in retinal detachment, subretinal and intraretinal edema, and scarring.

While these late stages show different pathologies, early stages of AMD show similarities in pathogenesis. RPE cells form the outer blood-retinal barrier and provide essential photoreceptor support functions. Dysfunction in the RPE layer is clinically manifested as altered pigmentation and lipofuscin accumulation within the macular region. Drusen accumulation between the RPE and Bruch's membrane and alterations in structural and biochemical properties of Bruch's membrane are similarly associated with aging and early-stage AMD. However, events responsible for the initial insult to the retina and progression to late-stage AMD are not currently un­derstood.^1,2

Within the last few years, there have been some exciting advancements in treating the wet form of AMD, including verteporfin pho­to­dy­nam­­ic therapy and development of anti-vascular en­dothelial growth factor (VEGF) compounds, such as pegaptanib sodium (Macugen, OSI/Eye­tech) and ranibizumab (Lucentis, Ge­nen­­tech).^3,4 Ef­fective treatment options for dry AMD are not currently available, however.

AMD has been classified as a complex disease with multiple genetic and environmental risk factors. Risk factors for AMD include increased sun­light exposure, Caucasian race, female gender, increased dietary fat and smoking. Mu­ta­tions in ApoE, ABCA4, FBLN5 and HEMICENTIN-1 genes have been reported to contribute, in part, to AMD progression.^5-12 Early evidence of a genetic component for AMD was reported in a classical study showing hard and soft drusen formation correlated with 57 percent increased AMD risk in monozygotic (identical) twins and 81 percent increased risk in dizygotic (maternal) twins. However, until recently, genetic factors pre­disposing a large population of pa­tients to AMD and pointing to a central mechanistic pathway in the pathophysiology of this disease have remained obscure. More recent genetic studies are closing this gap.

Figure.1. Histology from a 58-yr-old donor (A) showing normal Bruch's membrane, retinal pigment epithelium pigmentation, and photoreceptor layers (OS = outer segments, IS = inner segments, ONL = outer nuclear layer(photoreceptor nuclei). (B) An 89-year-old donor with drusen (asterisks), thickening of Bruch's membrane (circled), and RPE pigmentation abnormalities (arrow). Methyl green nuclear stain.

Molecular properties of the cho­roid/­ Bruch's membrane/RPE axis and subsequent perturbations occurring during aging have been extensively studied in order to identify the mechanistic pathways responsible for AMD pathogenesis. Alterations in pigmentation, accumulation of toxic lipofuscin pigments, and formation of drusen, which are extracellular deposits that accumulate between the RPE and Bruch's membrane, are all hallmark features of AMD (See Figure 1). 

The RPE plays a critical role in the support of photoreceptor cells. In addition to providing a selective cellular filtration component to the outer blood-retinal barrier, the RPE directly aids the photoreceptor by enzymatic re-isomerization of the visual chromophore and en­suring the daily turnover of photoreceptor shed disc membranes.¹³ These roles have been verified by identification of mutations in RPE genes affecting these functions, which result in photoreceptor degeneration. Lipofuscin in the RPE is a byproduct of incomplete outer segment digestion. These highly autofluorescent retinoids accumulate (See Figure 2) when the aging RPE is unable to completely digest the disc membranes.

Figure 2. A. Brightfield image of a human retina from an elderly donor. B. Autofluorescent image from the same section using an FITC filter set. Note that RPE cells, drusen (*) and Bruch's membrane (BM) are autofluorescent.

The production of one such lipofuscin fluorophore, A2E, is a marker for RPE dysfunction and is significantly associated with AMD pathogenesis.¹⁴ A2E is a down­stream condensation product of all-trans-retinol with phospha­ti­dyl­eth­a­no­lamine (PE). A fraction of all-trans-retinol is conjugated to PE within the disc membranes for transfer across the disc membrane by the photoreceptor-specific retinal re­cept­or/trans­porter en­coded by the ABCA4 gene and subsequent conversion to all-trans-retinol by all-trans-retinol dehydrogenase.¹³ Con­ditions that favor increased levels of all-trans-retinol and de­creased conversion to all-trans-retinol lead to accumulation of A2E. A2E accumulation in RPE cells destabilizes membranes, in­duces mitochondrial release of pro­apoptotic factors, inhibits proper degradation of ingested pho­toreceptor outer segments, and sensitizes cells to blue-light damage, all of which result in RPE damage and cell death.¹⁵ Mutation in gene ABCA4 is associated with Star­gardt's macular de­generation. Mice with genetic dis­ruption of ABCA4 develop Stargardt-like retinal degeneration, in­cluding RPE ac­cumulation of lipofuscin and A2E, and RPE and photoreceptor atrophy.

Drusen, composed of cellular debris and modified proteins and lipids from the RPE and Bruch's membrane, are associated with aging and believed to predispose the retina to disease progression. Drusen are subcategorized into both hard and soft forms, with the soft form conferring a significantly higher risk for AMD progression. Studies show compositional similarities between dru­sen and amyloid deposits, atherosclerotic plaques and dense deposits in atherosclerosis, amyloidosis, elastosis, and dense deposit disease associated with mem­branoproliferative glomerulo­neph­ritis type II.^16-18 Studies show that molecular and cellular components of drusen are consistent with local cellular (RPE) dysfunction and atrophy.

Accelerated lipofuscin and A2E accumulation occurs in patients with AMD and Stargardt macular dystrophy, a juvenile form of macular degeneration sharing many pathological features of AMD, and in mouse models of these retinal dis­orders. Accumulation of A2E has been associated with RPE dysfunction and atrophy and photoreceptor loss. As such, recent efforts have centered on tar­geting A2E formation in RPE cells as a therapeutic strategy for AMD and AMD-like retinal disorders. In early studies, it was shown that reducing dietary levels of vitamin A (all-trans-retinol) inhibits lipofuscin and autofluorescence accumulation in the RPE.¹⁹ While im­practical as a therapy, these results led to developing an alternative ap­proach for inhibiting vitamin A delivery specifically to the retina.

Figure.3. The alternative complement activation pathway.

Treatment of ABCA4-/- mice with Ac­cutane (isotretinoin), which inhibits 11-cis-retinol dehydrogenase (the final step in regeneration of the visual chromophore) and perhaps other steps in the visual cycle, has been shown to re­duce all-trans-retinal and subsequent formation of light-sensitive oxidation A2E products, A2E oxiranes.^14,20 While doses necessary to achieve these results in humans would be toxic, these results nonetheless showed that reducing all-trans-retinol can inhibit retinal A2E products in vivo.

An alternative strategy for reducing retinol levels specifically in the retina was examined.²⁰ N-(4-hydroxyphenyl) re­tinamide (HPR) (fenretinide) is a synthetic amide of all-trans-retinoic acid that has been used for decades, most commonly in cancer treatments, to re­move serum retinol bound to carrier protein, retinol-binding protein (RBP).²⁰ While this treatment has relatively few side ef­fects, treated patients reported ef­fects on night vision. It was subsequently shown that while additional retinol-binding proteins can compensate for re­ti­nol delivery to other tissues, the eye relies primarily on RBP for its retinol supply.²¹ Therefore, the selective effect of HPR treatments on retinol delivery to the eye was the basis for testing the efficacy of this drug on A2E formation in the ABCA4-/- Stargardt mouse model. HPR treaents inhibited the formation of RPE lipofuscin and A2E in these mice. However, prolonged treatment studies are needed, and efficacy in inhibiting the degeneration of RPE and photoreceptors has not yet been ascertained. More potent inhibitors of the retinoid (visual) cycle with prolonged drug activities have been identified.²² These compounds will need to be further developed before human trials can begin.

Alternatively, death of photoreceptor and RPE cells could result from the ac­cumulation of visual pigments uncoupled from the chromophore.²³ Thus, sup­plementation of the chromophore for visual pigment regeneration might be required.²⁴ These two individual strategies, or the combination of both, will re­quire direct testing in humans to determine optimal results.

 

When RPE cells die, debris serves as a "nucleation" site for drusen formation and subsequent tissue inflammatory re­sponses.²⁵ Proteomic analysis of drusen from normal vs. AMD patients shows ev­idence of oxidative protein modifications that may contribute to drusen formation and AMD pathogenesis.²⁶ Dru­sen formation in the aging eye has been compared to other aging disorders, such as Alzheimer's disease and atherosclerosis, where plaque formation and deposits induce chronic in­flam­matory stimuli and contribute to dis­ease progression. Complement C3-, C5- and C5-9 activation components have also been identified in drusen.^27,28

The alternative complement pathway is initiated by the spontaneous hydrolysis of plasma C3 to C3a and C3b (See Figure 3).²⁹ Unregulated C3b binds to cell surfaces, thereby pro­viding a mechanism for complement activation that will proceed unless downregulated by complement regulating proteins. This mechanism is the basis for the alternative pathway of complement activation. Com­ple­ment factor H (CFH), an essential member of this regulatory pathway, is a plasma glycoprotein that binds to cell surface sialic acids or neutral or anionic polysaccharides, as well as to C3b. This interaction accelerates C3b de­gradation and inhibits C3 convertase ac­tivation. C3b not neutralized by CFH is available for binding to factor B, resulting in activation of the alternative pathway. In brief, C3b stabilization results in C3-convertase activation, converting C3 to C3a and C3b, with subsequent convertase activation to produce C5a (a macrophage che­mo­attractant) and C5b. C5b, combines with C6, C7, C8, and C9 to form a membrane attack complex (MAC), with C5b-9 in the terminal complement pathway. Therefore, production of C5a and C5b-9 activates, promotes and amplifies inflammatory re­sponses.

 

Considerable advances have been made in the last year linking the complement system to AMD. Several laboratories reported a significant association between AMD risk and mutations in the gene for CFH, the important regulator of the alternative complement pathway (See Figure 3).^30-33 A tyrosine residue at position 402 re­placed by a histidine residue (a T to C substitution at nu­cleotide 1277 in exon 9) shows allele frequency be­tween 0.61 to 0.94 in AMD cases and 0.34 to 0.46 in age-matched controls (collectively) among these studies. Further, this variant confers a strong risk for soft drusen formation and both dry and wet forms of AMD.³⁴

In a recent study, haplotype analysis of patients carrying the Y402H variant who do not develop AMD revealed factor B and complement component 2 gene variants that confer a protective effect, seemingly counterbalancing effects from the CFH variant.³⁵ These results strongly support a central role for immune system regulation in the pathogenesis of AMD.

In support of these results, Ccl-2 (mo­no­cyte chemoattractant protein, MCP-1) and Ccr-2 (MCP-1 receptor) knockout mice develop AMD-like path­ology including drusen formation, thickening of Bruch's membrane, li­po­fuscin and A2E accumulation in RPE cells, and photoreceptor degeneration. Normally, choroidal mac­ro­phages are attracted to C5a and IgG deposits in drusen for subsequent de­gradation of these immune complexes. Therefore, loss of MCP-1 and its re­ceptor by genetic ablation in this mouse model was proposed to im­pair the recruitment of macrophages, re­sulting in chronic complement activation and tissue pathogenesis in these mice.³⁶

The recent genetic evidence correlating complement factors with AMD risk strongly supports a major inflammatory component for AMD pathogenesis. While previous identification of environmental and genetic risk factors have contributed to our knowledge of retinal aging disorders, recent identification of immune system in­volve­ment fills an im­portant gap in our understanding of the pathogenesis of AMD. In addition, the current re­search to control the formation of pho­tosensitizing fluorophores such as A2E during aging represents an equal­ly exciting approach for future therapies. Future research will focus on these pathogenic mechanisms in the aging eye. In addition, new strategies can now be developed that specifically address these injurious pathways, and may finally provide a therapeutic option for dry AMD.