Evidence for an inflammatory process in age-related macular degeneration gains new support

Age-related macular degeneration (AMD) is a sight-threatening eye disease that affects large numbers of the human population over the age of 65. It is particularly prevalent in European countries and in the United States, which has a large population of European descent. However, its incidence is increasing in Japan, as well, for reasons that are not clear (1). Currently, it is estimated that 1.75 million individuals suffer from this disease in the United States, and 7 million are said to be “at risk” (2). The emotional and socioeconomic impact of AMD is large because of reading and driving impairment caused by this disease, which primarily affects central (straight on) rather than peripheral (side) vision. The disease involves the macula of the retina, a region ≈6 mm in diameter that lies within the central axis of vision. The macula has an abundance of densely packed, specialized neurons called photoreceptor cells (rods and cones) that receive the visual stimulus and initiate a complicated cascade of biochemical and ionic events (phototransduction) that begin the visual process. A stratum of cells called the retinal pigment epithelium (RPE), resting like a single layer of paving stones on a bed of extracellular matrix called Brüch's membrane, separates the photoreceptor cells from their blood supply in the choroid (middle layer) of the eye wall. It is within the RPE layer and Brüch's membrane that the mischief leading to AMD is thought to begin.

Risk factors for AMD are well established from epidemiologic studies (3). In addition to advanced age, the risk factors include ocular pigmentation, dietary factors, a positive family history for AMD, high blood pressure, and smoking. Insights into the etiology of AMD have been slow in development because of the late onset of the disease. However, recent access to the human genomic sequence has opened the door to more powerful analytical methods, including haplotype mapping and SNP analysis (4). In this issue of PNAS, Hageman et al. (5) report that a variation in the factor H gene (HF1/CFH) dramatically increases the likelihood of developing AMD as well as membranoproliferative glomerulonephritis type II (MPGN II). HF1 encodes a protein involved in the body's first line of immune defense (the innate system) against infection by bacteria and other microbes. This manuscript complements three separate studies (published during the review) linking the same gene to AMD (6-8). Hageman et al. (5) also provide information pertaining to the ocular distribution and expression of HF1. In addition, they present data regarding AMD-associated gene variations, protective and risk haplotype maps of the HF1 gene, a potential role of the risk haplotype in a second disease, and an intriguing and expanded hypothesis related to the potential role of infection and aberrant complement activation in AMD.

Hageman et al. (5) chose HF1 as an AMD candidate gene based on their work spanning the past 10 years and on functional and disease-related evidence. In previous studies, Hageman, Mullins, Anderson, and Johnson (9, 10) implicated the complement cascade, a pathway associated with the innate immune system, in the formation of drusen (Fig. 1), the hallmark lesions in Brüch's membrane that accompany AMD. Drusen include remnants of the RPE, dendritic cell processes, and a variety of immune-associated molecules including immunoglobulins, class II antigens, and a host of complement components, activators, and regulators (11, 12). One of these regulators, factor H, is a key component of the alternative pathway of complement activation. Collectively, these observations led the investigators to conclude that AMD, like other age-related diseases, such as Alzheimer's disease and atherosclerosis, could involve a major inflammatory component. The authors correctly reasoned that MPGN II might provide fresh insights into the pathophysiology of AMD. They noted that MPGN II, except for its early onset, has ocular manifestations that are indistin-guishable from AMD (13, 14). Additionally, it was noted that a point mutation in HF1 (I1166R) causes MPGN II in pigs (15) and that factor H-deficient mice develop severe glomerulonephritis (16). Moreover, affected individuals within a couple of extended families with MPGN III showed linkage to chromosome 1q31-32 (17), a locus close to the 1q25-31 region that previously had been associated with AMD in genome linkage scans (18). These collective observations led the investigators to consider factor H as a prime candidate gene for both AMD and MPGN II.

Fig. 1.

Immunocytochemistry of a druse (D) from the eye of an 85-year-old donor. The entire druse is stained with antibodies against complement factor H (green). In the center of the druse, factor H colocalizes with the C5b-9 membrane attack complex of complement (orange). The RPE, which is distorted by the druse, contains autofluorescent lipofuscin granules (blue). Factor H staining is also visible in the lumen (L) of the capillaries, which are separated from the RPE by Brüch's membrane (BM). Colocalization of factor H and C5b-9 is also observed in the capillary wall (orange). Image is courtesy of Patrick Johnson and Kellen Betts (University of California, Santa Barbara).

The authors analyzed HF1 in 900 AMD patients and 400 matched controls in two cohorts from two geographic locations for genetic variation associated with disease (characterized independently at the University of Iowa, Iowa City, and Columbia University, New York). Hageman et al. (5) identify several common SNPs in HF1 as risk factors associated with AMD. Interestingly, the investigators also defined both AMD-associated and protective haplotypes of the HF1 gene. The other recently published studies have assessed broader haplotypes across the regulator of complement activation (RCA) gene cluster on chromosome 1q. The most frequent at-risk haplotype was associated with nearly half of the individuals with AMD, compared with ≈29% of controls with a P value of 10^-13, again providing strong evidence that HF1 likely plays a major role in susceptibility to AMD. Importantly, there was remarkable concordance when the University of Iowa and Columbia University groups were compared. This percentage is by far the most impressive “guilt-by-association” value obtained thus far for putative AMD-associated genes, which, to date, have typically hovered below 2%. The magnitude of the observed association is striking when compared with genetic abnormalities previously attributed to AMD (ABCA4, FBNL5, FBNL6, and APOE genes).

As the body's first line of defense against microorganisms and other foreign particles, the complement system is poised to recognize, attack, and kill invading microorganisms by creating holes in their cell membranes. In some cases, however, sustained complement activation can lead to chronic inflammation, aggravate local tissue damage, and contribute significantly to disease progression, such as that which occurs in Alzheimer's disease and atherosclerosis (19). To prevent such damage, a number of proteins, including factor H, the major soluble inhibitor of the alternative pathway of complement activation, keep the system under tight control. Because most of the identified variations occupy important functional sites, called short consensus repeat domains, of the HF1 protein, the researchers suggest that these AMD-associated risk haplotypes may alter the behavior of HF1 protein and hinder its role in regulating the immune system's complement pathway. One might speculate, therefore, that individuals with AMD and MPGN II share a functional defect in factor H protein that affects the function of the complement system. SNP variations might change binding of HF1 to complement fragment C3b or C reactive protein, sialic acid, or heparin, for example. Similarly, these variations may alter binding interactions between HF1 and microbes, perhaps making local tissues, such as the RPE, more susceptible to infection. This concept is consistent with the fact that the major “triggers” of the alternative pathway are common components found on many bacterial and viral surfaces. Whatever the exact triggering mechanism involved, this paper proposes the concept that dysfunction of the complement system results in local tissue damage, particularly at vulnerable locations such as the renal glomerulus and the retinal macula. It is relevant to note here that Hageman and colleagues (20), in an earlier publication this year, have shown that the elastic layer of Brüch's membrane, which normally keeps blood vessels of the choroid at bay, is preferentially thinner in the macular region (20). Disruption of this thinned layer by ensuing inflammation may help to explain the predilection of the macula toward lesion formation, including the formation of choroidal neovascular membranes.

The data presented in the paper by Hageman et al. (5) also show that the HF1 risk haplotype is associated with a broad range of AMD phenotypes, the possible exception being geographic atrophy, which is seen in a relatively small subset of patients. This finding would suggest that there is no distinct relationship between disease-associated genotype and clinical phenotype. The results of this study are also consistent with some of the established epidemiological risk factors for AMD. Smoking, as stated earlier, is known to inhibit factor H activity, and this habit increases the risk of AMD by as much as 4- to 5-fold.

The data presented in the study by Hageman et al. (5) should be of broad interest to both the genetics and ophthalmic communities. The study is important for many reasons, and the results provide compelling evidence that the early prediction that complement plays a key role in AMD was correct (9). Although much remains to be done, it appears likely that the inheritance of the at-risk HF1 haplotype, in combination with an infectious agent or other atypical activators of the alternative pathway, such as immune complexes, nephritic factors, amyloid-β peptide, or cholesterol, may substantially increase one's susceptibility to AMD and MPGN II. The way is now paved for the development of new diagnostic assays, novel bioassays, and new animal models that faithfully mimic AMD pathogenesis. It is also clear that molecules involved in complement activation and its regulation will now move to the forefront as prime targets for the development of early diagnostic tests and therapeutic treatments for AMD and perhaps for other inflammation-based diseases as well. Most importantly, these observations should provide a degree of comfort to the many individuals who are afflicted with this devastating condition.