The Aging Eye: Common Degenerative Mechanisms Between the Alzheimer's Brain and Retinal Disease

Abstract

Alzheimer's disease (AD) is a common, incurable, and progressive dementia, characterized by loss of learning and memory and the neuropathologic accumulation of amyloid plaques and neurofibrillary tangles in the brain. A number of similarities between AD pathology and several distinct retinal degenerations have been described, particularly with respect to either glaucoma or age-related macular degeneration (AMD), each a leading cause of vision loss and blindness worldwide. Although comparisons between these diseases may provide important new insights into their pathogenic mechanisms, glaucoma and AMD result in markedly different degenerations. Therefore, analyses of the differences and the similarities between these conditions may prove equally productive. Common mechanisms that appear to underlie all three diseases are explored here, as well as potential use of the retina as a biomarker for AD diagnosis and progression. Based on this comparison, past and current efforts to transfer therapeutic strategies between diseases are discussed.

Introduction

Alzheimer's disease (AD) is a common, progressive, and age-associated dementia that afflicts 67 in 1000 people over the age of 65, and more than 26 million people worldwide. ^1,2 In spite of intense research into the pathogenesis and treatment of AD, current approaches remain focused on minimizing the associated cognitive deficiencies and do not directly address pathogenic mechanisms. Diagnosis and progression of AD, especially early cases, are complicated by imprecise neuropsychological testing, sophisticated but expensive neuroimaging techniques, and/or invasive sampling of cerebrospinal fluid (CSF). ^3,4 In the case of positive test results, only a diagnosis of “probable Alzheimer's” is possible until postmortem analyses of the brain can confirm the presence of senile plaques and neurofibrillary tangles, leading to high error rates in diagnosis and clinical trial end points. ⁴ In short, there remains a great need for new research models that can be exploited to better understand mechanisms of AD pathogenesis, as well as simple and accessible diagnostic procedures to monitor suspected AD patients. In this regard, an increasing number of studies have explored sensory deficiencies in AD, particularly correlations between disease progression and visual dysfunction. ^5–7 Clarifying the etiology of these impairments has proven difficult, as defects have been identified in both retinal and cortical visual processing. ^5,7–9 However, careful study of retinal pathology and function in AD eyes has suggested that local tissue damage plays an important role, with relevant comparisons to common causes of vision loss and blindness. 

The retinal diseases that have been most commonly compared to AD pathology are glaucoma ^10,11 and age-related macular degeneration (AMD). ¹² Superficially, there are a number of striking similarities among the three disorders. All three are complex, multifactorial degenerations of central nervous system (CNS) tissue, in which age is a primary risk factor. Pathologically, they all feature progressive deposition of protein aggregates, including extracellular amyloid β (Aβ) plaques and intracellular microtubule inclusions containing hyperphosphorylated tau protein (pTau). Additional common features include prominent glial reactivity, neuroinflammation, and increased metabolic and oxidative stress. Importantly, despite many years of intense research, in each case the damage-triggering mechanism continues to be debated. However, glaucoma and AMD are very distinct conditions, so there may be much to learn from comparing them to each other as well. Common mechanisms that appear to underlie all three diseases are discussed here, along with potential use of the retina as a biomarker for AD diagnosis and progression. Based on this comparison, past and current attempts to transfer therapeutic strategies between diseases are described. 

The Case for an AD–Glaucoma Connection

Glaucoma continues to be a leading cause of vision loss and blindness, affecting an estimated 60 million people worldwide. ^13,14 The pathogenesis of glaucoma is complex and remains poorly understood; but progression of optic neuropathy is characterized by cupping of the optic nerve head (ONH), thinning of the nerve fiber layer (NFL), and chronic retinal gliosis, resulting in retinal ganglion cell (RGC) death and loss of visual function. ^15–17 In this sense, glaucoma can be considered a neurodegenerative process, to the extent that transynaptic degeneration has been observed in the lateral geniculate nucleus and visual cortex of patients. ^18,19 The largest single risk factor for glaucoma is age, ^20,21 and the most established clinical measure is increased intraocular pressure (IOP). ²² However, current treatments to reduce IOP do not directly address retinal and ONH damage, and most patients progress in development of optic neuropathy. ^15,23 Also, a substantial portion of primary open-angle glaucoma patients have IOPs that consistently remain in the normal range. ²⁴ Vascular dysregulation in normal-tension glaucoma is highlighted by associations with optic disc hemorrhages, changes in ocular blood flow, systemic hypotension, ^25,26 and increased activity of hypoxia-inducible factors (HIFs). ²⁷ These data suggest that a variety of local insults, including biomechanical strain and ischemia, combine to compromise the sensitive RGC homeostasis. 

Clinical Associations between Glaucoma and AD

Several clinical studies have reported increased prevalence of glaucoma in patients with AD and dementia compared to control groups. ^28–30 For example, a 25.9% prevalence of glaucoma has been reported in AD patients, compared to 5.2% in a control group. ²⁸ Increased prevalence of cognitive impairment has likewise been shown to be associated with glaucoma patients, ³¹ though other studies have failed to find a positive correlation in this direction. ^32,33 Therefore, the epidemiologic relationship between glaucoma and AD remains unclear until larger prospective studies can be performed. 

Clinical data and experiments on human eye tissues have consistently supported common AD and glaucomatous retinal pathology (Table 1). Histopathologic analyses of enucleated glaucomatous eyes have shown evidence of substantial RGC loss compared to controls, ^34,35 while a higher proportion of patients with AD showed abnormalities in the retinal NFL. ³⁶ RGC damage has also been accompanied by increased glial cell numbers and reactivity in AD eyes. ³⁵ Interestingly, analyses of retinal hemodynamic parameters in patients with early AD showed significant vessel narrowing and reduced blood flow rate, accompanied by NFL thinning, compared to a control group. ³⁷  

Table 1. 

 

Similarities and Differences between AD, Glaucoma, and AMD

Characteristic	AD	Glaucoma	AMD
Location of retinal damage	RGC, NFL	RGC, NFL, ONH	RPE, photoreceptors
Aβ deposition	Main component of senile plaques	Reported in RGCs following ocular hypertension	Component of subretinal drusen
pTau	Component of NFTs	Found in inner retina, vitreous, and ONH	Not reported in association with drusen
Oxidative and metabolic stress	Associated with metal deficiencies and neuronal damage	Associated with metabolic strain on RGCs	Associated with visual cycle and drusen formation
Glial reactivity	Astrocytes and microglia around senile plaques	Astrocytes and microglia in NFL, peripapillary space, and ONH	Microglia in subretinal space and around drusen
Genetic linkage	No clear commonalities, though there may be a mild association between distinct APOE alleles for AD and AMD

Several studies have also looked for the presence of fibrillar proteins in retinas and aqueous humor from AD and glaucomatous eyes. Increased staining for pTau protein and isoforms has been demonstrated in the inner retina from patients with ocular hypertension, ³⁸ in optic nerve and peripapillary glia, ³⁹ and in retina and vitreous from aged and glaucoma patients. ^40,41 Tau is a microtubule-associated protein that when hyperphosphorylated can lead to formation of intracellular neurofibrillary tangles (NFTs) in the cortex, as well as disruption of neuronal cytoskeleton and axonal transport. In AD patients, increased pTau and certain tau isoforms have been commonly associated with NFT formation. ^4,42 The role of pTau in glaucomatous neuropathy is still unclear, but disruptions to retrograde and anterograde axonal transport have been implicated in the disease process. ⁴³  

A study of AD patient vitreous samples additionally reported a decrease in fibrillogenic Aβ₄₂ peptide, an isoform associated with AD senile plaques. ^44,45 Although counterintuitive, this result mirrors increased pTau and decreased Aβ₄₂ reported in CSF of AD patients. ⁴⁶ However, another study failed to find any change in aqueous Aβ₄₂ levels associated with glaucoma. ⁴⁷  

Similarities in Rodent Models of Glaucoma and AD

There are substantial similarities in retinal damage associated with rodent models of AD and glaucoma. Increased Aβ, and correspondingly decreased amyloid precursor protein (APP), were observed in RGCs in rat models of acute ocular hypertension. ^48,49 Aβs are small peptides that are processed from APP through the activities of β-APP cleaving enzyme (BACE1) and the γ-secretase complex containing presenilins 1 and 2 (PS1, PS2). Certain sizes of processed Aβ, particularly Aβ₄₂, tend to aggregate into plaques, which can induce the neurotoxic damage in AD, according to the amyloid cascade hypothesis. ^50,51 Guo et al. showed that targeting multiple aspects of Aβ production, clearance, and aggregation was effective at reducing IOP-induced RGC death. ⁴⁸  

Transgenic mouse models of AD have also been studied to identify any resulting retinal damage that may develop. Mice harboring mutant APP, presenilins 1 and 2, and tau genes have been generated based on mutations associated with familial early-onset forms of AD. These strains exhibit varying degrees of AD brain pathology, including Aβ plaques and neurofibrillary tangles. ⁵² Correspondingly, these animals have generally been reported to develop overexpression of APP, deposits of Aβ and pTau, and neuronal cell death in their inner retinal layers (RGC, NFL, and INL). ^53–56 Liu et al. reported increased Aβ, pTau deposition, and neuroinflammation in the RGC and NFL of Tg2576 mouse retinas harboring a mutant APP gene, ⁵⁴ while Gasparini et al. show increased pTau filaments that inhibit stimulation of RGC axonal outgrowth in P301S tau mice. ⁵³ Therefore, mouse models of AD have generally exhibited glaucoma-like pathology, consistent with clinical and epidemiological data. 

The Role of Reactive Gliosis

Activation of retinal glia has been closely associated with pathogenesis of both glaucoma and AD. Astrocytes perform an essential role in maintaining neuronal homeostasis by mediating extracellular ion and neurotransmitter balance, regulating vascular flow and blood–brain barrier integrity, and secreting a host of growth factors and neurotrophic factors. ^57,58 In response to injury or disease, astrocytes become reactive, turning hypertrophic and migratory, characterized by increased networks of the intermediate filaments, glial fibrillary acidic protein (GFAP), and vimentin, eventually resulting in formation of a glial scar. ^58,59  

During pathogenesis of glaucoma, activated astrocytes of the ONH have been proposed to direct ONH cupping and NFL remodeling through secretion of matrix metalloproteases and extracellular matrix components, ^60–62 and signal through a variety of cytokines and growth factors, including epidermal growth factor, bFGF, TNF-α, and TGF-β1. ⁵⁸ Activation of Müller glia has also been described following chronic ocular hypertension. ^63,64 Similarly, reactive astrocytes have been consistently implicated in AD brains and animal models. ^65,66 A common hypothesis proposes that exposure to Aβ may induce a chronic neuroinflammatory response in the brain, characterized by secretion of cytokines, generation of nitrogen radicals, and modulation of phagocytosis. ^59,67,68 Activated microglia have also been observed in the retina and ONH of glaucomatous eyes ⁶⁹ and animal models of ocular hypertension. ^70,71 In AD brains, microglia appear on senile plaques and surrounding tissue. Once activated, microglia can fulfill supportive functions, such as the degradation and elimination of Aβ, but over time can also release damaging cytokines and chemokines, such as IL-6, TNF-α, and nitric oxide radicals. ^72,73  

Comparisons between AD and AMD

Age-related macular degeneration is another complex but common retinal disease, with a prevalence that increases exponentially with age. In the United States alone, AMD has been estimated in over 7 million individuals, with up to 50 million affected worldwide. ^74,75 The majority of patients suffer from the nonexudative “dry” form, characterized by thickening of Bruch's membrane; activation of the innate immune system; and progressive formation of drusen, toxic subretinal extracellular deposits. With time, the dry form can progress to the more severe exudative (wet) form, characterized by choroidal neovascularization (CNV) and retinal edema. ⁷⁶ Alternatively, drusen can continue to expand, eventually leading to degeneration of a large area of retinal pigmented epithelium (RPE) and associated photoreceptors in a process known as geographic atrophy. 

Comparison of Drusen and Senile Plaques

The formation and composition of drusen form the foundation for comparisons between AMD and AD (Table 1). Drusen are generally found between the RPE and the underlying Bruch's membrane. The precise mechanism of drusen formation is still unclear, but they are accompanied by RPE atrophy and may be generated through a buildup of toxic by-products of the phototransduction cycle. ⁷⁷ Resulting inflammation and metabolic and oxidative stress play an important role in the induction of a neovascular response from underlying choroidal vessels and progression to neovascular AMD. The density and size of drusen generally correlate with the severity of central vision loss, as well as risk for development of neovascular AMD. ^78–80 These deposits can be compared to the formation of senile plaques in the cortex and hippocampus of AD patient brains, with diffuse plaques occurring at earlier stages of the disease. ^12,81  

Increased deposition of Aβ isoforms have been described on photoreceptor outer segments and along the RPE–Bruch's membrane interface in the aging human and mouse retina. ⁸² It is not clear, though, whether this increase plays a role in loss of RPE or photoreceptor function. Analyses of drusen components have shown deposition of Aβ within vesicles, particularly in eyes of advanced AMD patients. ^83,84 RPE cells have subsequently been demonstrated to express APP and associated processing enzymes, and to react to Aβ peptides with secretion of proinflammatory and proangiogenic factors. ⁸⁵ In addition to Aβ, proteomic analyses have implicated a host of other common proteins between drusen and senile plaques, including tau, basement membrane proteins, proinflammatory factors, and components of the complement cascade of the innate immune system. ¹²  

Genetic Links

Alleles of the lipid recycling apolipoprotein, APOE, have been associated with both AMD and AD. The APOE-e4 allele has been reported to be associated with lower risk of AMD, and the APOE-e2 allele with higher risk. ^86,87 The opposite association is found in AD: higher risk of developing late-onset disease with APOE-e4 ^88,89 and reduced risk with APOE-e2. ⁹⁰ Nevertheless, this common genetic connection would be consistent with involvement of APOE alleles in Aβ proteolysis and neuronal cell membrane renewal in the brain and macula. However, a follow-up study failed to find a link between APOE genotypes and AMD, ⁹¹ and the explanation for the opposite allelic effects is still unclear. 

Increasing evidence suggests involvement of the complement system in AMD pathogenesis, particularly through the alternative pathway. Multiple alleles of the alternative pathway inhibitor, factor H, have been strongly associated with increased risk of development of AMD, ^92,93 as well as other complement regulatory components, including factor B and C3. ⁹² From histologic and protein analyses, complement components, including active fragments of C3, C5, C6–9 (forming the membrane attack complex), and regulatory factors B, H, and I, have all been reported in both drusen ^94–96 and senile plaques. ^97–100 These findings suggest that common inflammatory mechanisms are involved in AMD and AD. However, support for a genetic linkage between complement factor H alleles and AD has so far been weak. ^101–103  

Oxidative and Metabolic Stress in AMD and AD

Due to their constant activity and turnover, photoreceptors in the macula are subjected to an extremely high metabolic and oxidative stress that increases with age. ^104,105 In this environment, any deficit in RPE function compromises an already delicate balance, resulting in acceleration of the degenerative process. RPE degeneration is further encouraged by accumulation of lipofuscin, cross-linked pigmentary deposits from photoreceptor membranes that RPE cells are unable to metabolize. Lipofuscin has been shown to have damaging oxidant properties and further compromise mitochondrial function. This cycle of metabolic and oxidative stress is exacerbated by the damaging effects of accumulation of Aβ and other protein aggregates. Many features of this mechanism are mirrored by established aspects of AD brains, including increased reactive oxygen species (ROS) and oxidative damage, lipofuscin formation, and mitochondrial dysfunction. ^42,106  

The Potential for Common Therapeutic Strategies

The Retina as an AD Biomarker

A major goal for ongoing clinical AD research is the early detection of plaques and tangles in the brain and their subsequent monitoring. Although advances in brain imaging have explored the use of magnetic resonance imaging (MRI) and positron emission tomography (PET), these approaches have so far been limited in sensitivity and resolution. ³ Likewise, CSF analysis is invasive, while neuropsychological testing can be imprecise. Based on the established visual deficits in AD patients, imaging of the eye could be a simple and accessible alternative. 

This potential biomarker strategy has been pursued using various retinal imaging approaches. Retinal photography, scanning laser ophthalmoscopy (SLO), and optical coherence tomography (OCT) have previously identified retinal NFL loss, retinal blood flow changes, and optic disc changes in AD patients ^{36,37,107,108} (Table 2). Multifocal ERG recordings have also shown subtle defects in electrical activity in the macula of AD patients. ¹⁰⁹ Koronyo-Hamaoui et al. have described increased Aβ plaques in the retinas of postmortem eyes from early-stage AD patients, and in vivo in double-transgenic APP_swe/PS1_ΔE9 mice, compared to controls. ¹¹⁰ In this case imaging was facilitated with prior injection of curcumin, a natural extract and fluorochrome that labels Aβ plaques. ^111,112 The group was able to demonstrate a correlation of retinal plaque number and size with brain pathology, and have proposed translation of this method into clinical use. ¹¹³ One trial, in progress, is using OCT to assess NFL thickness and other retinal and vascular indicators in early AD patients (ClinicalTrials.gov, ID NCT01555827). However, a challenge for these efforts will be to develop methods to differentiate AD pathology from other common retinal diseases. 

Table 2. 

 

Studies Testing the Retina as an AD Biomarker

Biomarker	End Point	Reference or Trial ID
OCT	NFL thickness, blood flow	Moschos et al., 2012; NCT01555827
SLO	ONH damage	Danesh-Meyer et al., 2006
Fundus photography	NFL damage	Hedges et al., 1996
Laser Doppler velocimetry	Retinal blood flow	Berisha et al., 2007
Multifocal ERG	Electrical activity in the macula	Muschos et al., 2012
Spectral imaging	Aβ plaques	Koronyo-Hamaoui et al., 2011

Neurotroprotection

Administration of small secreted neurotrophic factors has been tested with mixed results in various models of retinal degeneration and AD. Prominent factors include ciliary neurotrophic factor (CNTF), brain-derived neurotrophic factor (BDNF), glial cell line–derived neurotrophic factor (GDNF), and nerve growth factor (NGF). ¹¹⁴ In some cases, similar approaches have been positive in both retinal and AD models, reinforcing these common mechanisms and providing additional momentum toward clinical testing. For example, intravitreal injection of CNTF peptide or virus in a rat model of increased IOP significantly reduced RGC death. ^115,116 CNTF treatment also proved protective in multiple models of photoreceptor degeneration through injected, viral, and cell-based approaches. ^117–119 Recently, intraocular delivery of CNTF through an encapsulated cell implant (NT-501; Neurotech Pharmaceuticals, Cumberland, RI) has been reported to stabilize visual acuity and increase macular volume in a 12-month phase II clinical trial in patients with geographic atrophy AMD. ¹²⁰ A similar cell encapsulation approach has been used to provide long-term neuroprotective and cognitive benefits in multiple mouse AD models. ¹²¹  

Reduced BDNF levels have been reported in the cortex and hippocampus of AD brains. ^122,123 Viral delivery of BDNF to the entorhinal cortex in mouse, rat, and aged primates preserved synapses, reversed memory and spatial learning deficits, and improved cognitive performance. ¹²⁴ However, this treatment had no effect on plaque density. Similarly, administration of BDNF, and/or activation of its associated receptor TrkB, has been shown to be protective in various acute and chronic models of RGC and optic nerve damage. ^125–128 Massa et al. have recently reported discovery of a class of potent small-molecule BDNF mimetics that agonize TrkB. ¹²⁹  

Another common neuroprotective strategy has been the inhibition of excitotoxic neurodegeneration caused through excess glutamate, particularly via overstimulation of N-methyl-D-aspartate (NMDA) receptors. Glutamatergic excitotoxicity has been consistently implicated in AD pathology, primarily through inhibition of glial-mediated glutamate uptake and processing. ^130,131 Therefore, glutamate receptor antagonists have been used to dampen excessive neurotransmission. The NMDA receptor antagonist memantine HCl has been approved for use in AD patients to treat dementia, but its use has remained controversial as follow-up studies showed mixed results. ^132,133 Based on reports implicating similar excitotoxic damage in glaucoma, as well as positive results in an animal model of RGC injury, ^134,135 Allergan initiated a phase III clinical trial of oral memantine in over a thousand patients in a long-term trial from 1999 to 2008. ¹³⁶ An official report has not been published, but a press release indicated that although some benefits were observed, the results failed to significantly distinguish drug effects from placebo. This experience has dampened enthusiasm for further clinical testing of neuroprotective agents for glaucoma until improved trial designs and outcome measures are developed. 

Amyloid β Targeting

There are numerous current AD trials based on targeting Aβ plaques in the brain. Strategies for these treatments include secretase inhibitors (α, β, and γ), Aβ vaccines, aggregation blockers, and Aβ monoclonal antibodies, as reviewed in two previous reports. ^51,131 Based on current attempts in brain, Guo et al. comprehensively tested the role of Aβ in a rat ocular hypertension model. They targeted Aβ formation through administration of a β-secretase (BACE) inhibitor, enhanced Aβ clearance with an anti-Aβ antibody, and reduced aggregation with Congo red. Neuroprotective effects on RGCs were observed with all three approaches, though they were particularly strong with a combined strategy. ⁴⁸ Similarly, an antibody targeting Aβ was administered to a mouse model of drusen formation and showed reduced plaques, accompanied by a dose-dependent rescue of ERG readings. ¹³⁷ These results support the common amyloidopathic aspects of these diseases. However, in spite of the efforts to date, the clinical efficacy of this strategy in AD remains unclear. ⁵¹  

Anti-inflammatory and Immunosuppressive Approaches

The role of chronic inflammatory mechanisms in the eye and brain remains a subject of active research. Recent data point to the involvement of activated microglia and astrocytes in mediating proinflammatory signals and phagocytic processes in the eye and brain. ^17,58,59,138 However, the mechanisms regulating the activation process, as well as the positive and/or negative influences of these cells, are still under debate. The antibiotic minocycline mediates neuroprotective effects through inhibition of microglial activation. It has been tested with encouraging results in models of AD ^139,140,161 and RGC and optic nerve injury. ^141–143  

Another strategy involves administration of the synthetic copolymer glutiramer acetate (GA), a T-cell–modulating antigen. Butovsky et al. ¹⁴⁴ administered GA to AD double-transgenic APP_swe/PS1_Δe9 mice and observed reduced plaque formation, reduced cognitive decline, and increased neurogenesis. The mechanism proposed for this effect is through T-cell–mediated activation of resident microglia to buffer Aβ concentrations and secrete protective and neurotrophic factors. A small clinical trial has now tested this mechanism in a group of patients with dry AMD. After a course of 12 weekly subcutaneous injections of copolymer 1 (Cop-1), significantly reduced drusen area was observed. ¹⁴⁵ In a similar approach, Bakalash et al. had earlier shown that vaccination with GA produced neuroprotective effects in a rat model of chronic ocular hypertension. ¹⁴⁶ They screened this model and found that GA-induced neuroprotection correlated with signaling through the early growth response 1 (Egr-1) transcription factor. A similar correlation was observed in the brains of APP_swe/PS1_Δe9 transgenic AD mice. ¹⁴⁷ These experiments provide an interesting example of knowledge transfer between AD, glaucoma, and AMD in support of a common therapeutic strategy. 

The recent discoveries linking components of the complement cascade to AMD have produced several therapeutic initiatives. ^92,148 These include the C3 inhibitor POT-4 (Potentia Pharmaceuticals, Louisville, KY), the C5 antibody eculizimab (Alexion Pharmaceuticals, Cheshire, CT), the C5 aptamer ARC1905 (Ophthotech Corporation, Princeton, NJ), the C5 antibody LFG316 (Novartis AG, Basel, Switzerland), and the factor D antibody FCFD4514S (ClinicalTrials.gov, IDs NCT00935883, NCT01229215, NCT00473928, NCT01255462, NCT00950638). The impact of complement activation in AD pathogenesis and progression is still unclear, but research in this direction will directly benefit from the forthcoming AMD trial results. 

Oxidative Stress

The only current treatment for dry AMD is based on results of the Age-Related Eye Disease Study (AREDS) of over 3000 patients for an average of 6 years. ¹⁴⁹ The study showed that a mixture of micronutrients and antioxidants (vitamins C and E, β-carotene, and zinc) significantly reduced the risk of progression to advanced AMD, but not disease onset. The AREDS2 trial is currently in progress to further refine this formulation (ClinicalTrials.gov, ID NCT00345176). Supplementation of dietary antioxidants has so far proven of little benefit for patients with primary open-angle glaucoma. ^150,151 However, direct overexpression of antioxidant thioredoxins 1 and 2 has been shown to be protective in multiple models of RGC injury, ¹⁵² as has administration of the reducing chemical PB1. ¹⁵³ In AD, increased accumulation of iron has been well established as an early biomarker. ^154,155 Among other effects, iron and other heavy metals interact with ROS to produce damaging hydroxyl radicals. Excess iron and aluminum have been associated in AD brains with Aβ deposits, neurofibrillary tangles, and activated microglia. ^106,156 However, attempts to treat AD with iron chelators has so far met with limited clinical success. ^157,158  

Table 3. 

 

Recent Tests of Clinical and Laboratory Agents Overlapping for AD, Glaucoma, and AMD

Target	Agent	Mechanism	Disease	Reference
CNTFR	CNTF	Neurotrophic factor	AD	Garcia et al., 2010
			Glaucoma	Ji et al., 2004; Pease et al., 2009
			AMD	Zhang et al., 2011
TrkB	BDNF	Neurotrophic factor	AD	Nagahara et al., 2009
			Glaucoma	Cheng et al., 2002; Di Polo et al., 1998; Hu et al., 2010; Mansour-Robaey et al., 1994
NMDAR	Memantine HCl	Glutamate excitotoxicity	AD	Bordji et al., 2011; Iliffe, 2007; Mount & Downton, 2006
			Glaucoma	Hare et al., 2001; Seki & Lipton, 2008; Quigley, 2012
Amyloid β	Various: antibodies, vaccines, aggr blockers, secretase inhibitors	Plaque/drusen formation	AD	Lukiw, 2012; Karran et al., 2011
			Glaucoma	Guo et al., 2007
			AMD	Ding et al., 2011
T-cell–mediated gliosis	Cop-1	Neuroinflammation	AD	Butovsky et al., 2006; Bakalash et al., 2011
			Glaucoma	Bakalash et al., 2005; Bakalash et al., 2011
			AMD	Landa et al., 2008
Microglial activation	Minocycline	Neuroinflammation	AD	Song et al., 2006; Parachikova et al., 2010
			Glaucoma	Baptiste et al., 2005; Shimazawa et al., 2005; Bosco et al., 2008

 aggr, aggregation; CNTFR, ciliary neurotrophic factor receptor; NMDAR, N-methyl-D-aspartate receptor.

Discussion

Besides the visual dysfunction associated with AD itself, several common disease features appear in retinas from AD cases and animal models of glaucoma and AMD. Although the overlap is not complete, it is clear that these three age-related CNS diseases display components of amyloid and pTau aggregation, neuroinflammation, and oxidative damage. In the retina, such stresses are particularly relevant to the exposed and metabolically sensitive RGC and photoreceptor cells. It is not surprising, therefore, that these two retinal layers are the focus of damage in glaucoma and AMD, respectively. However, it is still unclear whether these similarities represent common initiating neurotoxic mechanisms or the downstream consequences of distinct insults (Figure). This discussion is mirrored and compounded within the research community for each disease as well. 

Figure. 

 

Two models to integrate common pathology associated with AD, glaucoma, and AMD. (A) Deposition of Aβ and pTau, combined with other common factors such as oxidative and metabolic stress, directly contributes to neuroinflammatory responses and cell death. (B) Insults specific to each tissue and disease result in common effects that combine to compromise neuronal viability.

Although there are parallels between the visual deficits and pathology associated with glaucoma and AD, the relative sequence of brain and retinal damage needs to be clarified. Also, the correlation between cases of glaucoma and cognitive impairment remains poor, and the initial insult appears to differ; the biggest discrepancy in this regard is the association between glaucoma and biomechanical stress through increased intraocular pressure. Finally, no clear genetic link has yet been established between the diseases. However, consistent evidence of amyloid and pTau aggregation in the inner retina, common RGC and NFL damage, and chronic gliosis suggest overlapping degenerative processes (Table 1). 

In the cases of AD and AMD, both feature plaque formation, inflammation, and oxidative stress. However, there are also many points where these diseases differ. First among these is the lack of clinical correlation: A study of 2088 individuals with early AMD was unable to find a significant association with AD, ¹⁵⁹ and AD eyes do not tend to show RPE deposits or increased drusen. Consistent with these observations, mouse models of AD have generally produced effects in the inner retina and optic nerve, more reminiscent of glaucoma. However, mice deficient in the amyloid metabolizing metalloprotease, neprilysin, have shown Aβ deposition below the RPE and associated cell death. ⁸⁵ The impact of a recently uncovered rd8 (Crb1) mutation in the C57BL/6N mouse substrain on previous AMD-related research remains unclear. ¹⁶⁰ Therefore, genetic or epidemiologic data supporting a direct connection between AD and AMD has been limited, and these conditions likely represent two distinct but related amyloidopathies that may benefit from common targeted therapeutic approaches. 

Based on the similarities that have been described here, a growing number of opportunities exist to adapt therapeutic agents and strategies for testing between AD and the retina. Some successful examples involving neuroprotection, plaque formation, inflammation, and oxidative stress have been discussed (Table 3). Therefore, with efforts effectively pooled, a clinical breakthrough in the retina will hopefully promote a cascade of advances in the brain and vice versa. 

Acknowledgments

Thanks to Izhar Livne-Bar for helpful comments and suggestions. 

Footnotes

 Supported by the Canadian National Institute for the Blind (CNIB), the Glaucoma Research Society of Canada (GRSC), and the Canadian Institutes of Health Research (CIHR). Dr Sivak is the Toronto General and Western Hospital Foundation Glaucoma Research Chair.

Footnotes

 Disclosure: J.M. Sivak, None