Evidence is rapidly accruing to support the notion that these entities, particularly AMD, have a significant immune component.
Immunosuppressive therapy has been shown in a few patients to have a positive therapeutic effect in AMD.
Understanding these disorders furthers our knowledge of the multiple mechanisms that make up the downregulatory immune environment of the eye.
It seems fitting to include this chapter in a book on uveitis. Diseases that have not traditionally (at least by most eyecare specialists) been thought of as immune mediated are being considered to be within the immune sphere. We will deal with three very common ocular disorders: age-related macular degeneration (AMD), diabetic retinopathy, and glaucoma. It is clear that as we gather more information about glaucoma and AMD we are seeing striking parallels with other ‘degenerative’ disorders, such as atherosclerosis and Alzheimer’s disease. Diabetes has been long thought of an immune-mediated disease, but these concepts have not percolated to the eye particularly. It is clear that we will be able to support immune involvement in all three of the diseases. It is not whether the immune system is involved: rather, the more important question is just how central the immune mechanism is to the clinical manifestations we observe, and whether, with this knowledge, it can be manipulated in such a way as to arrive at a positive therapeutic response.
Age-related macular degeneration
Age related macular degeneration is the most common cause of irreversible blindness in the United States and the United Kingdom. It is estimated that 1.75 million people in the United States are affected with the more advanced form of the disease. Indeed, it is now considered the most common cause of irreversible blindness in the world. If there is no change in lifestyle or treatment strategies, it has been estimated that by 2020 this number will increase to 3 million Americans with the severe form of the disease. The disorder probably begins years before the patient is aware of visual alterations. It is characterized by the presence of large drusen, which can coalesce, disappear, and appear elsewhere. More about drusen later. Suffice to say that large drusen and the small drusen seen in many patients do not seem to be related. The risk increases with each risk factor in the eye. The two significant ones are large drusen and pigment changes. If you have four risk factors, that is, the two in each eye, the 5-year risk of developing advanced AMD in either eye is 50%.
In addition to the presence of large drusen, the disease will show loss of RPE (geographic atrophy) as well as hyperplasia of the RPE, and in a minority of patients in the United States, choroidal neovascularization (CNV).
Much work has centered on this condition. There have been several reviews, and the interested reader should consult them. , This chapter will be a short synopsis of what has become a large corpus of information.
Animal laser model
Punching a hole through Bruch’s membrane in a mouse or rat will result in a choroidal neovascular lesion which will appear 2–3 weeks after the insult ( Fig. 31-1 ). Using this model for CNV, several investigators have evaluated inflammatory responses as well as potential therapeutic approaches. Work using this model would suggest that macrophages limit the size of a CNV lesion, at least produced in this injury model. In addition, dendritic cells were found next to the induced CNV site, and immature dendritic cells appeared to enhance the CNV size. The same group identified genes in the mouse that appear to control the size of the angiogenic lesion after laser injury. Semkova et al. reported that overexpression of FasL in the RPE resulted in less CNV. Of interest was that the three most commonly used anti-VEGF agents in humans showed no effect in diminishing leakage from laser-induced CNV lesions in rats. The data from this model are sometimes contradictory, perhaps suggesting the complex interactions of subtypes of cells that may be involved.
Ccl2 and Ccr2 knockout model
Ccr2, a chemokine receptor, is present on macrophages. Its ligand is MCP1 (Ccl2) and it aids in the adhesion of macrophages to vasculature and helps them move into the tissues. Ambati et al. showed that by 9 months, these mice develop drusen and a thickening of Bruch’s membrane. By 16–18 months, they can demonstrate RPE atrophy and in some cases neovascularization. Impaired macrophage function leads to an accumulation of immune components such as C5 in the areas of AMD-like change. Bear in mind that Tsutsumi et al. showed that a reduced number of macrophages at a laser-induced lesion yielded less CNV.
Ccl2 and Cx3cr1 double knockout model
This model, developed at the Laboratory of Immunology of the NEI, , develops drusen-like lesions after 4–6 weeks ( Fig. 31-2 ).
There is increased complement activity with deposition at the drusen-like lesions, Bruch’s membrane, the RPE and the choroidal capillaries. These mice also demonstrated antiretinal antibodies as seen in AMD patients.
CEP induced AMD-like disease
Carboxyethylpyrrole (CEP), a unique oxidation fragment of docosahexaenoic acid, is found within the large drusen of patients with AMD. Hollyfield et al. immunized mice with this material, and over a period ranging from 3 to 9 months the animals developed posterior segment lesions that were AMD-like. Macrophages were noted near the RPE and there were areas of RPE loss ( Fig. 31-3 ). The significance of this model is that it was induced by immunization, and so a T-helper response, and that it was induced with material found in human eyes with this disease.
There are other knockout models showing fundus changes that are reminiscent of AMD. These include knockouts of CFH, ABCR, and Cx3Cr1.
There is an increasing amount of information about AMD and the possible role of the immune system. Some of it dates back to the 1990s.
Penfold and colleagues reported the presence of retinal autoantibodies in patients with AMD. More recently, Patel and coworkers reported that patients with choroidal neovascularization and large drusen had autoantibodies to retinal elements far more often than did controls (82.2%, 93.75% and 8.69%, respectively).
Recently there has been a flurry of interest in evaluating gene variants (single nucleotide polymorphisms) with AMD. This has led to many different associations. Perhaps the finding that received that most interest was the association of the complement factor H variant Y402H with AMD. Although other variants have been associated with AMD, such as complement factor B (BF), complement component 2 (C2), C3, C7, and complement factor I (FI) and CX3CR1, the association of Y402H was seen to be particularly strong (χ 2 = 54.4 and P = 1.6 × 10 −13 ), and much speculation has centered around the implications of the association. CFH is encoded on chromosome 1. It is a major regulator of complement activity, blocking activation of the complement cascade. Complement is traditionally associated with the innate immune system, the oldest part of our immune system which reacts rapidly to invading organisms. It has been speculated that the Y402H variant does not inhibit complement activity as efficiently as the wildtype molecule, thus leading to the immune activation and AMD. However, to date no functional studies have definitively shown alterations in activity with the variant rather than the wild type of CFH.
Another variant that has been studied and associated with AMD is that of high temperature requirement serine protease (HTRA-1) (rs11200638). This variant, which is located on chromosome 10q26, is a major genetic risk factor for wet AMD. As with CFH HTRA-1, in initial studies in a population with a very low presence of the Y402H variant, i.e., a group of Eastern Asians in Hong Kong, the P value was as essentially as robust as that with CFH Of interest was the strong association of the same variant found in a Caucasian population of 581 patients and 309 normal controls in Utah. It was estimated that this variant conferred an attributable risk of 49.3%. In addition, Yang and coworkers, using anti-HTRA1 antibody to label drusen, demonstrated the presence of this variant in the centerpoint of AMD. In addition, Yoshida et al. found the same variant in a Japanese population.
What is special about HTRA-1? It appears to have several functions that may be important in the pathogenesis of AMD. It is part of the heat shock proteases that are induced with stress. In addition, it is a regulator of TGF-β activity. TGF-β is considered a downregulatory immune molecule, , one of the major molecules in maintaining a downregulatory immune environment in the eye. Further, a substrate of HTRA-1 is fibronectin. Fibronectin fragments stimulate proinflammatory and catabolic elements from murine RPE, including IL-6, MMP-3, MMP-9, and MCP-1. All of these products could contribute to inflammation, catabolism, and attraction of monocytes, which are all elements of AMD.
Are there others associations? Indeed, several others have been made, mostly with molecules that are somehow involved with the immune response. These include the CXCR1 variant V294I, complement factor B, complement components including C2, C3, and C7, the PLEKHA1 gene, IL-8, ARMS2, , and the toll-like receptor 3 variant L412F, which has been associated with protection from the atrophic form of AMD.
How do we then have so many molecules already considered as such dominant players in AMD? One possibility is that they are players in a larger downregulatory environment, as discussed below.
Macrophages and other cells
As mentioned both above and below, macrophages/monocytes appear to be a major player in the pathogenesis of this disease. The real question is, exactly what is it? Is it that they are not functionally efficient enough, or are they being overactive? The animal work has provided us with mixed information. Cousins et al. evaluated circulating macrophages from patients with AMD and from controls ( Table 31-1 ). In this study, ‘activated’ macrophages were defined as those making large quantities of the proinflammatory lymphokine TNF-α. This is a lymphokine discussed in several other chapters. Cousins et al. found that the macrophages of patients with the wet form of AMD produced statistically significantly greater amounts of TNF-α than did those from controls and the dry type of AMD.
|TNF-α mRNA Levels in Fresh Monocytes||Prevalence of Test Result, No. (%)||OR (95% CI)||P Value|
|Dry AMD (n = 22)||Neovascular AMD (n = 25)|
|Low||11 (50)||5 (20)||NA||NA|
|Medium||4 (18)||4 (16)||2.20 (1.99–2.38)||0.32|
|High||7 (32)||16 (64)||5.03 (4.99–5.06)||0.02|
It is important to bear in mind that macrophage subtypes may be very significant factors in this the induction of this disease. As discussed in the Fundamentals chapters, the M 1 subtype of macrophage can conceivably present antigen and produce proinflammatory substances. This is in contrast to the M 2 subtype, which may scavenge but may not present antigen, maintaining a downregulatory atmosphere.
Bear in mind also the possible role of microglia, resident immune cells, as well as dendritic cells. All of these cells have been shown to potentially interact with immune cells when they enter the eye.
The attempt here is not to discuss the pathologic changes classically described as associated with AMD, but whether there are implications of an immune response. Penfold et al., in 1985, reported the presence of macrophage, fibroblasts, lymphocytes and mast cells not only at the level of the neovascular lesion, but also in the atrophy of the RPE and the breakdown of Bruch’s membrane. It is quite possible that the inflammatory process occurs early on in the disease, at a time when there are minimal or no visual alterations. What we see as the disease is the consequence of something that occurred quite some time earlier. Large drusen, the most reliable harbinger of AMD, have been extensively evaluated. Using liquid chromatography tandem mass spectrometry analyses of drusen, Crabbe et al. identified 129 different proteins in both small and hard and large and soft drusen. Crystallins, heat shock proteins, and products of oxidative reactions were noted. Anderson et al. evaluated the drusen and surrounding tissue in more than 400 AMD eyes. They noted RPE cells undergoing cell death with the accumulation of C5 in their cytoplasm. Drusen in AMD were also noted to harbor C3 and C5 activation fragments, in addition to HTRA1 noted above.
The downregulatory immune environment
Can this be the underlying link? We have noted evidence of autoantibodies implicating T- and B-cell activation. Macrophages seem somehow central to the underlying disease. We know that ocular resident cells, particularly the RPE, can interact readily with the immune system. We have seen that drusen contain fragments of molecules involved with the immune response, both innate and acquired. Finally, the genetic associations constantly increase, and almost always involve genes associated with immune regulation. All these genetic associations are suggested as the underlying cause of the disease. How is that possible? Another possible explanation is that all of these associations are really affecting a major characteristic of the ocular environment, which we have termed the downregulatory immune environment (DIE). With this concept it really does not matter what the genetic alterations are, and we are sure to find others. It is the disruption of DIE that leads to the disorder, and it can occur through multiple mechanisms. We know that the multiple molecules in the eye create an environment that suppresses immune responses. It would not be in the organism’s best interest to have inflammatory responses in the eye except when absolutely necessary. ACAID is one of the better-described ways that this occurs. Table 31-2 lists many of the ways in which this downregulatory environment is maintained, involving many of the cells found in the back of the eye as well. One could explain the possible scenario leading to AMD as follows:
The RPE produces debris which is excreted into the environment. This includes immunogenic antigens such as β crystallins and retinal antigens such as arrestin. The multiple mechanisms associated with DIE are in play. This would include the presence of TGF-β and CFH.
Macrophages: we could hypothesize that the M 2 type – those that clear debris but do not necessarily induce an inflammatory response – come in small numbers to help clean up the debris. However, if there is a defect in macrophage recruitment then the debris will collect in larger amounts. Perhaps this explains some of the animal models. This clearing of debris is a lifelong process.
With aging, DIE is not as an effective immune downregulatory mechanism. More macrophages and a small number of inflammatory cells now gather. There are also Th cells present, and macrophages are producing the proinflammatory cytokine IL-17. Activation with proinflammatory cytokine production and autoantibodies occurs.
Over a long a period of time the proinflammatory cytokines and autoantibodies are toxic to the RPE and could cause apoptosis of these cells. TLR3 variants may protect this from happening.
For those with either the CFH or HTRA1 SNPs (and others) that put individuals at higher risk for AMD, DIE’s control is further diminished. The presence of drusen is the body’s attempt to wall off the active components of the immune system to prevent a heightened immune response. The drusen are thus a sign of a chronic inflammatory response that has overcome DIE’s control.
A more rigorous inflammatory response will lead to further loss of RPE (because of toxic lymphokines) and the development of CNV with the recruitment of bone marrow cells ( Fig. 31-4 ).
|Cellular||Cell Surface and Soluble Factors||Physical|
|NKT cells||Transforming growth factor-β||Light|
|F4/80 macrophages||CD95 ligand|
|IL-1 receptor antagonist|
|CD55, CD59, CD46|
|Soluble inhibitors of C1q and C3 convertase|
|Vasoactive intestinal peptide|
|Calcitonin gene-related peptide|