RPE lipofuscin granules accumulate in the central to basal portion of the cell as spherical membrane-bound granules of 1-2 μm diameter (Fig. 2.1) (7). In addition, in aged donors lipofuscin can be observed complexed with melanin-forming aggregates of heterogeneous composition and size, in which case they are termed melanolipofuscin. Ex vivo and in vivo analyses demonstrate an age-related increase in lipofuscin content, up until the age of about 70 years (6,8, 9, 10). Thereafter, lipofuscin levels plateau or even decline (9). The reason for this decline is not clear, but it may involve one or a combination of lipofuscin degradation routes, e.g., voiding into the extracellular space (as basal laminar deposits or precursor material to drusen) or association with melanosomes. Concomitant with the accumulation of lipofuscin granules are a decrease in melanosomes and an increase in melanolipfuscin granules (8). On the basis of early morphometric data, the accumulation of RPE lipofuscin was erroneously considered to be bimodal (6), but it is now apparent that such accumulation is linear—at least up until the later decades, when the effects of age-related pathology begin to have an impact (Fig. 2.2) (9). Lipofuscin demonstrates a regional distribution across the fundus (9, 10, 11) (see Chapter 3). The distribution of lipofuscin demonstrates a retinal topography similar to that of rod photoreceptors, suggesting that ingested POS are a major substrate for RPE lipofuscin (11). Alternatively, the highest density of photoreceptors will require the highest level of metabolic activity for the RPE, and thus mitochondrial turnover and autophagy will be much higher.

The cytoplasmic volume occupied by lipofuscin increases significantly with age (Fig. 2.1). Feeney-Burns and colleagues demonstrated that 8% of the cytoplasmic volume of the central RPE is occupied by lipofuscin granules by 40 years of age, and this increases to 19% by 80 years of age (8). It is likely that this would affect metabolic function, intracellular compartmentalization of organelles, and lysosomal activity, reducing the functional capacity of the RPE with increasing age. In old age, the cytoplasmic volume occupied by the equatorial and peripheral RPE is considerably less than that in the central retina.

In addition to the regional distribution of lipofuscin across the fundus, there is considerable cell-to-cell heterogeneity in lipofuscin density in any particular region of the retina (Fig. 2.1) (12). It is evident that two adjacent cells (presumably with the same functional and metabolic load) can have high, intermediate, or minimal levels of lipofuscin. This would support the proposal by Burke and Hjelmeland (12) that mosaicism of the RPE is a natural result of normal mechanisms regulating gene expression during development and postnatal aging. This heterogeneity may help explain the highly focal nature of the lesions associated with pathologies such as AMD.

Among the different tissue lipofuscins, RPE lipofuscin has a number of unique properties that determine its photophysical properties and its potential for contributing to RPE dysfunction, as will be discussed in more detail below. First, RPE lipofuscin is derived at least in part from POS and thus contains significant quantities of retinoids (4,13), including the diretinal conjugate A2E (4). Second, RPE lipofuscin is a broadband absorber that is constantly exposed to visible light and is located in a highoxygen environment, which makes it an ideal substrate for retinal photodamage (1).

Although RPE lipofuscin is believed to result from the incomplete degradation of POS and autophagy of spent or damaged intracellular organelles, the relative contributions of these pathways remain unknown. It is thought that prior oxidative damage to these substrates renders them undegradable by the lysosomal system, the universal intracellular degradation machinery that contains over 40 hydrolytic enzymes. However, prior oxidation does not account for the fluorescent properties of lipofuscin, since Eldred and Katz (14,15) showed that the autofluorescent products of lipid peroxidation differ from those of lipofuscin. Although we have gained an understanding of the initial substrates involved in lipofuscin genesis, we still have a limited understanding of the composition of the mature lipofuscin granules. Probably the first attempt at analysis was made by Eldred and Katz (13), who used Folch’s extraction to separate lipofuscin into three fractions: chloroform-soluble, methanol-soluble, and insoluble (Fig. 2.3A). Thin-layer chromotography analysis of the chloroform fraction revealed 10 discrete fluorophores: two green-emitting, which comigrated with retinol and retinyl palmitate; three emitting yellow/green; and one golden-yellow and four orange-red fluorophores, the most prominent of which was subsequently identified as A2E (Fig. 2.3) (16). Further studies by other researchers identified the fluorophores as retinoid derivatives (4,17). This was followed by many studies on the role of A2E, the only readily synthetic component of lipofuscin, and demonstrations of its phototoxic and lysosomotrophic properties. However, the content of the bulk (>70%) of lipofuscin material, which is located at the insoluble interface on Folch’s extraction, has largely been ignored (18).

It has long been considered that lipofuscin consists of cross-linked oxidatively modified lipids, proteins, and sugars, with the protein content ranging from 10% to 70% of the lipofuscin granule. The first proteome for RPE lipofuscin was reported by Schutt and colleagues (19) in 2002. They identified over 65 abundant cellular proteins, including cytoskeleton proteins, proteins of transduction, enzymes of metabolism, proteins of the mitochondrial respiratory chain, ion channel proteins, and chaperones, which supported the origin of lipofuscin as a combination of photoreceptor phagocytosis and autophagy. Furthermore, analysis of the samples revealed that many of
these proteins were damaged by the aberrant covalent modifications of MDA, 4-HNE and AGEs (20). In a subsequent proteome analysis of RPE lipofuscin, Warburton et al. (21) identified 41 cellular proteins; surprisingly, only 11% these proteins had been identified in the proteomic analysis of Schutt and colleagues (19). Although there are numerous explanations for this discrepancy, the likely problem lies in the purity of the starting material. Lipofuscin is generally isolated by repeated sucrose density gradient centrifugation (19,21,22); however, there is usually some residual contamination by
cellular debris (22). While investigating this further, we demonstrated that highly purified debris-free RPE lipofuscin granules contain little or no protein, but do contain significant amounts of modified material, such as carboxyethylpyrrole adducts (23). In conclusion, although our understanding of the composition of lipofuscin in the RPE remains limited, it does appear that protein makes only a very small contribution, if any, and that oxidized and modified lipids from outer-segment and mitochondrial membranes may be major contributors.