Vitamin D and Age-Related Macular Degeneration

In recent years, the relationship between vitamin D and health has received growing attention from the scientific and medical communities. Vitamin D deficiencies have been repeatedly associated with various acute and chronic diseases, including age-related macular degeneration (AMD). Its active metabolite, 1α,25-dihydoxy vitamin D, acts as a modulator of cell proliferation, differentiation and apoptosis, and cumulative data from experimental and observational studies suggest that relatively a lower vitamin D status could be a potential risk factor for the development of early and/or late AMD. Herein, we made a narrative review of the mechanisms linking a potential role of vitamin D with the current concepts of AMD pathophysiology.


Introduction
Age-related macular degeneration (AMD) is a chronic, progressive degenerative disease affecting the macula and reducing central visual acuity in advanced stages. This is the leading cause of irreversible visual impairment in the elderly population in developed countries, accounting for 7% of keratinocytes exposed to ultraviolet radiation (UV-B), leading to the photolysis of 7-dehydrocholesterol (7-DHC) to pre-vitamin D3. Excess pre-vitamin D3 is converted into various inactive metabolites [22]. Once produced by the skin, vitamin D binds to a specific binding protein (DBP) and is released into the bloodstream. Part of the vitamin D produced is stored in fat cells [23]. Vitamin D, produced by the skin or supplied by food, is biologically inert and requires two subsequent hydroxylation processes in the liver and kidneys to produce active metabolites, as described in Figure 1. The liver is generally considered the primary, and likely the sole source, of 25(OH)D production. However, other enzymes have 25-hydroxylase activity and may potentially affect levels of 25(OH)D in the blood and possibly other tissues [24]. Contrary to the liver, 25-hydroxylases, both renal enzymes-CYP27B1 and CYP24A1-are tightly controlled [24]. Negative feed-back is produced by 1,25(OH) 2 D, through down-regulation of CYP27B1 and up-regulation of CYP24A1. The half-life of 1,25(OH) 2 D in plasma is relatively short (only several hours) [25], compared to the half-life of 25(OH)D (about 3 weeks). Therefore, serum 25(OH)D is considered the best biomarker of vitamin D status [25]. Proximal tubular epithelial (renal) cells are not the only source of 1,25(OH) 2 D production. CYP27B1 is also expressed in a number of extrarenal sites, including the gastrointestinal tract, skin, vasculature, placenta and immune cells [26]. Although the physiological impact of extrarenal CYP27A1 is still controversial [26], local synthesis of 1,25(OH) 2 D provides the basis for a paracrine or autocrine function.

Mode of Action
Vitamin D is a secosteroid that is structurally analogous to the steroid hormones (e.g., estradiol, cortisol, and aldosterone), but with an open B-ring. Similar to other steroid hormones, vitamin D functions according to two modes of action: a mechanism mediating gene transcription (genomic action) and a rapid non-transcriptional action, mediated by the activation of secondary messengers and phosphokinase activation (non-genomic action) [27,28]. The genomic pathway is mediated by the binding of 1,25(OH) 2 D with a high affinity vitamin D receptor (VDR). When activated, the VDR acts as a transcriptional factor [24,27] and may directly or indirectly control 200 to 2000 genes in various tissues and cells [29]. This includes genes involved in mineral and bone homeostasis, but also genes controlling cell proliferation, differentiation, and apoptosis [27,29,30]. The VDR is ubiquitously expressed throughout the human body [31], including in immune cells, endothelial cells and vascular smooth muscle cells [32], but also in eye tissues, including the retina [33]. It was recently demonstrated that vitamin D3 supplementation (400 to 2000 IU/day for 8 weeks) is associated with related alterations of 291 genes, including 17 genes known to play important roles in transcriptional regulation, immune function, apoptosis, and responses to stress [34]. The non-genomic pathway involves the interaction of 1,25(OH) 2 D with a specific receptor localized to the plasma membrane of target cells [28]. Based on cell type, signal transduction may involve different secondary messengers and cytosolic kinase systems, leading to biological effects that include the regulation of cell proliferation, differentiation, or apoptosis [30]. Vitamin D (ergocalciferol and/or cholecalciferol) is produced and excreted by basal skin keratinocytes exposed to ultraviolet radiation (UV-B), or directly provided by food. While skin vitamin D is transported into the liver bound to binding proteins (DBP), dietary vitamin D is absorbed by the gastro-intestinal tract and transported to the liver via the venous circulation and chylomicron remnants. Part of the vitamin D produced is stored in fat cells and may serve as an endogenous source of vitamin D. In the liver, vitamin D2 and vitamin D3 are hydroxylated in position 25 by several enzymes found in microsomal or mitochondrial fractions. Once produced in the liver, 25(OH)D is released into the bloodstream whilst bound to DBP. Alternatively, vitamin D can be metabolized in 25(OH)D in other tissues. In the kidney, 25(OH)D is converted to the active metabolite, 1,25(OH)2D, through the action of the enzyme 1-alpha-hydroxylase (CYP27B1), located in the proximal tubules. In excess, 1,25(OH)2D and 25(OH)D activate 24-hydroxylase (CYP24A1) and are degraded into 24-hydroxylated products, i.e., 24,25(OH)2D and 1,24,25(OH)3D, which have no biological activity. Once produced in the kidney, 1,25(OH)2D is released and transported into the bloodstream and is mainly bound to DBP until it reaches target tissues expressing the vitamin D receptor. Once produced in the kidney, 1,25(OH) 2 D is released and transported into the bloodstream and is mainly bound to DBP until it reaches target tissues expressing the vitamin D receptor.

Association between Vitamin D and Health Outcomes
Vitamin D is well known for its major role in bone mineral homeostasis; it promotes the transport of calcium and phosphate, ensuring adequate bone mineralization [35]. Vitamin D is thought to have other biological functions (Figure 2), and observational studies have suggested an inverse relationship between the plasma level of 25(OH)D and the risk of developing various chronic diseases, including cancers, infections, cardiovascular diseases, auto-immune diseases, and diabetes [22,24]. Although a link is supported by experiments in vitro or animal studies, as reviewed by others [32,[36][37][38], not all literature supports a protective association between vitamin D and chronic disease outcomes. There has still no obvious causality link shown in interventional studies, and consistent data from randomized clinical trials are still scarce, as reviewed by Bikle [24]. Thus, to date, the Scientific Advisory Committee on Nutrition (SACN) has recognized the critical role of vitamin D in bone health, but not in other chronic diseases [39].

Association between Vitamin D and Health Outcomes
Vitamin D is well known for its major role in bone mineral homeostasis; it promotes the transport of calcium and phosphate, ensuring adequate bone mineralization [35]. Vitamin D is thought to have other biological functions (Figure 2), and observational studies have suggested an inverse relationship between the plasma level of 25(OH)D and the risk of developing various chronic diseases, including cancers, infections, cardiovascular diseases, auto-immune diseases, and diabetes [22,24]. Although a link is supported by experiments in vitro or animal studies, as reviewed by others [32,[36][37][38], not all literature supports a protective association between vitamin D and chronic disease outcomes. There has still no obvious causality link shown in interventional studies, and consistent data from randomized clinical trials are still scarce, as reviewed by Bikle [24]. Thus, to date, the Scientific Advisory Committee on Nutrition (SACN) has recognized the critical role of vitamin D in bone health, but not in other chronic diseases [39].

Vitamin D and AMD Pathophysiology
A possible physiological role of vitamin D in the retina is supported by evidence that the vitamin D receptor (VDR) and the enzymes involved in the metabolism of vitamin D (CYP27B1 and CYP24A1) are expressed in the retina, RPE and choroid. The presence of VDR was demonstrated through immunohistochemical staining in the inner and outer segments of retinal photoreceptors and the RPE [33]. The presence of the VDR gene in resting retinal and choroidal endothelial cells was then confirmed by molecular biology [40,41]. Interestingly, Morrison et al. [40] reported that VDR expression in RPE/choroid tissues was generally observed at or below the level of detection, but was at significantly higher levels in several donors, concomitant with an increased expression of genes involved in local inflammatory responses. More recently, it was shown that the human RPE cell line, ARPE-19, was able to convert 25(OH)D to 1,25(OH)2D [42]. The presence of the VDR and enzyme 1alpha-hydroxylase (CYP27B1) in retinal and RPE/choroidal cells strongly suggests that vitamin D may function in a paracrine/autocrine manner. Thus, a major limiting step in vitamin D function may be the bioavailability of 25(OH)D, as lower systemic levels may directly diminish the amount of active vitamin D in the macula [42]. Protective effects of vitamin D, which prevent the development of early

Vitamin D and AMD Pathophysiology
A possible physiological role of vitamin D in the retina is supported by evidence that the vitamin D receptor (VDR) and the enzymes involved in the metabolism of vitamin D (CYP27B1 and CYP24A1) are expressed in the retina, RPE and choroid. The presence of VDR was demonstrated through immunohistochemical staining in the inner and outer segments of retinal photoreceptors and the RPE [33]. The presence of the VDR gene in resting retinal and choroidal endothelial cells was then confirmed by molecular biology [40,41]. Interestingly, Morrison et al. [40] reported that VDR expression in RPE/choroid tissues was generally observed at or below the level of detection, but was at significantly higher levels in several donors, concomitant with an increased expression of genes involved in local inflammatory responses. More recently, it was shown that the human RPE cell line, ARPE-19, was able to convert 25(OH)D to 1,25(OH) 2 D [42]. The presence of the VDR and enzyme 1-alpha-hydroxylase (CYP27B1) in retinal and RPE/choroidal cells strongly suggests that vitamin D may function in a paracrine/autocrine manner. Thus, a major limiting step in vitamin D function may be the bioavailability of 25(OH)D, as lower systemic levels may directly diminish the amount of active vitamin D in the macula [42]. Protective effects of vitamin D, which prevent the development of early and/or late AMD, have been proposed, based on experimental in vitro and in vivo studies, as reviewed by Wang & Hartnett [4] and Parmeggiani & Romano [43]. These include pathogenic steps centered on oxidation, inflammation and angiogenesis, as illustrated in Figure 3. and/or late AMD, have been proposed, based on experimental in vitro and in vivo studies, as reviewed by Wang & Hartnett [4] and Parmeggiani & Romano [43]. These include pathogenic steps centered on oxidation, inflammation and angiogenesis, as illustrated in Figure 3.

Inhibition of Chronic Oxidative Stress
One major physiological role of RPE cells is phagocytosis and intracellular degradation of shed photoreceptor outer segments [44]. Intracellular degradation is mediated by the generation of free reactive oxygen species (ROS) and lysosomal enzymes. Cellular debris is normally digested by the RPE cells. However, in senescent RPE cells, oxidative stress is increased, leading to cell damage, RPE dysfunction and release of abnormal extracellular matrix deposits. Vitamin D has been shown to be

Inhibition of Chronic Oxidative Stress
One major physiological role of RPE cells is phagocytosis and intracellular degradation of shed photoreceptor outer segments [44]. Intracellular degradation is mediated by the generation of free reactive oxygen species (ROS) and lysosomal enzymes. Cellular debris is normally digested by the RPE cells. However, in senescent RPE cells, oxidative stress is increased, leading to cell damage, RPE dysfunction and release of abnormal extracellular matrix deposits. Vitamin D has been shown to be protective against oxidative stress in various cell lines and animal models [45,46]. Recently, in a mouse cone cell line (661W), 1,25(OH) 2 D was shown to decrease the generation of ROS in H 2 O 2 -stimulated cells, by modulating the expression of antioxidant enzymes (catalase, superoxide dismutase, and glutathione peroxidase) [18].

Inhibition of Amyloid Beta Protein Deposits
A pathogenic role for amyloid beta protein (Aβ) in AMD was proposed in studies showing that Aβ increases VEGF expression in human RPE cells [47]. A large amount of Aβ is released by senescent RPE cells [48] and Aβ is a major component of drusen [49]. In addition, Aβ is considered a primary activator of the complement cascade and inflammation, as reviewed by Akiyama et al. [50]. The deleterious role of Aβ in the pathophysiology of AMD has been suggested by animal studies, which have shown that intraperitoneal administration of anti-Aβ antibodies in aged mice leads to a reduction of Aβ deposition in the retina and improves electroretinogram (ERG) deficits [51]. On the other hand, vitamin D has been recently associated with the clearance of Aβ deposits and the improvement of retinal function in aged mice treated with subcutaneous injections of vitamin D [52]. The proposed mechanism involves the activation of macrophage phagocytosis of Aβ deposits by 1,25(OH) 2 D, and their removal from the Bruch's membrane. These results may indicate a role of vitamin D supplementation in prevention or treatment of early AMD.

Inhibition of Chronic Inflammation
AMD can be considered a low-grade chronic inflammatory disease, where immunocompetent cells, such as macrophages and lymphocytes, are present in the chorioretinal tissues affected by AMD (for review see Parmeggiani et al. [43]). In animal experiments, inflammatory cells are located in the RPE, Bruch's membrane, choroid, and outer and inner segments, and were more abundant in areas with increasing damage [53]. In the aging retina, two inflammatory pathways are affected as a result of age-related tissue stress: the complement cascade (for review see Khandadia et al. [54]) and the tissue resident macrophage (retinal microglia) activation pathway [55]. Inflammatory events ultimately lead to protein damage, aggregation and degeneration of RPE and are therefore key elements in the pathogenesis of AMD [3]. On the other hand, there are several lines of evidence consistent with an anti-inflammatory role of vitamin D in the pathogenesis of AMD. Vitamin D was shown to inhibit immune reactions in chronic inflammatory diseases, as reviewed by Hewison et al. [56]. Cells of the myeloid lineage are known to produce CYP27B1 hydroxylase, leading to local production of 1,25(OH) 2 D. Activated macrophages express the VDR and are thus vitamin D responsive cells. One major effect of vitamin D on activated macrophages is potent suppression of pro-inflammatory events mediated by interferon-gamma (INF-γ) [57]. This is consistent with the study in mice, which showed that vitamin D administration reduces the number of activated macrophages and promotes a shift from pro-inflammatory cytokine-secreting macrophages to an amoeboid cell type, still capable of phagocytosis [52]. In the RPE/choroid, the inhibition of the inflammatory reaction mediated by cytokines, which are secreted by macrophages/microglia (Interleukin(IL)-1, IL-6 and tumor necrosis factor -1 (TNFα)), may prevent RPE degeneration and apoptosis [58,59].
Adaptive immunity in AMD is a recent investigational field, and data suggests that dysregulation of immune responses could contribute to different processes associated with AMD, such as RPE atrophy, neovascularization, and retinal degeneration [60]. It was reported that the C5a complement component induced IL-22 and IL-17 expression by human peripheral CD4+ T cells, and that elevated levels of these cytokines were present in AMD patients [61]. This provides a potential link between innate and adaptive immunities in the pathogenesis of AMD. Vitamin D, as a potential inhibitor of immune cell recruitment, proliferation and activation, may protect against the development of AMD. Indeed, vitamin D metabolizing enzymes and vitamin D receptors are present in many immune cells, including antigen-presenting cells, T cells, B cells and monocytes [38,56]. Vitamin D (1,25(OH) 2 D) has been shown to suppress proinflammatory cytokines, in part by altering T-cell function toward a Th2 (anti-inflammatory) rather than a Th1 (pro-inflammatory) response [56]. As reviewed by Prietl et al. [38], treatment of T cells with 1,25(OH) 2 D or its analogs, inhibits the secretion of proinflammatory Th1 (IL-2, INFγ, TNFα), Th9 (IL-9) and Th22 (IL-22) cytokines, but promotes the production of more anti-inflammatory Th2 cytokines (IL-3, IL-4, IL-5, IL-10). The production of IL-17 produced by Th17 cells is also inhibited by vitamin D [62].

Vitamin D and Angiogenesis in AMD
It is well established that angiogenesis plays a major role in the development and progression of AMD. VEGF, the most potent inducer of endothelial activation, is a modulator of vascular permeability in the macula. The RPE is capable of secreting a variety of growth factors, including VEGF, in order to maintain the physiological angiogenesis necessary for physiological development of the choroidal vasculature [63]. RPE dysregulation due to oxidative stress and inflammatory reactions may lead to abnormal angiogenesis [64]. On the other hand, hypoxia, caused by drusen accumulation in the Bruch's membrane, is considered a strong activator of VEGF secretion by RPE cells [65]. Inflammatory cytokines, such as TNFα released by activated microglia and IL-17 produced by Th17 cells, may also induce angiogenesis in choroidal endothelial cells [4]. As demonstrated in tumor cells, vitamin D is a potent inhibitor of angiogenesis due to its effects on endothelial cells and interruption of the angiogenesis signaling pathway [66]. In various human cancer cells, it has been shown that the anti-angiogenic activity of 1,25(OH) 2 D is mediated by inhibition of the transcription of hypoxia-inducible factor (HIF-1), which is a well-known inducer of VEGF under conditions of hypoxia [67]. A similar regulatory role for HIF-1 was shown in the pathogenesis of AMD [68]. In vitro and in vivo experiments, performed by Mantel et al., showed that 1,25(OH) 2 D inhibits specific stages of the angiogenic process in a dose-dependent manner [69]. Vitamin D (1,25(OH) 2 D) inhibits VEGF-induced endothelial cell sprouting and elongation and also has a small, but significant, inhibitory effect on VEGF-induced endothelial cell proliferation. In experiments carried out by Albert et al. using a mouse model of oxygen-induced ischemic retinopathy, animals treated with 1,25(OH) 2 D had significant decreases in retinal neovascularization relative to control animals, although the levels of ocular VEGF were similar in treated and control animals [70]. A possible mechanism may include induction of endothelial cell apoptosis [69,70]. On the other hand, inhibition of the production of the metalloproteinase, MMP-9, released by inflammatory cells may be another pathway controlled by vitamin D. In TNFα-stimulated keratinocytes, vitamin D was shown to attenuate the production of the metalloproteinase, MMP-9 [71], which is suspected to play a role in choroidal neovascularization, by altering the Bruch's membrane [72,73].

Association between Vitamin D Plasma Levels and AMD
The association between vitamin D and AMD is a relatively new investigational field. Observational studies conducted over the past decade, to investigate the relationship between vitamin D status and AMD risk, have led to controversial results. A recent meta-analysis, of 11 observational studies [74] indicated that low circulating levels of 25(OH)D (<50 nmol/L (20 ng/mL)) were significantly associated with late AMD, with an odds ratio (OR) of 2.18 (95% CI: 1.34-3.56). However, the association was not statistically significant in early AMD. In another recent review, with slightly different study criteria, Wu et al. concluded that there was no evidence to indicate an inverse association between blood vitamin D levels and risk of any stages or subphenotypes of AMD [75].
Although such meta-analyses overcome some statistical issues, the selected studies showed high heterogeneity in design, including population representativeness, ethnicity, time and methods to determine vitamin D status as well as AMD classification. In their meta-analysis, Wu et al. reviewed the quality of individual observational studies, based on 10 methodologic criteria that are not always fulfilled [75]. There are several well conducted observational studies which have shown an inverse relationship between vitamin D plasma levels and early AMD [20,76], or late atrophic and/or neovascular AMD [19,77]. However, other studies, with lower quality scores, have not shown significant associations between vitamin D status and prevalent AMD [78][79][80].
Finally, all but one of the published epidemiological studies (in particular those included in the two meta-analyses) were cross-sectional or case-control studies, with 25(OH)D being measured after the diagnosis of AMD. Although cross-sectional studies are useful for exploring a possible association between AMD and vitamin D, they are not designed to determine causality. As recognized by others, the reduced vitamin D levels observed in cross-sectional analyses of patients with AMD may be related to visual impairment, more sedentary indoor activities, less exposure to sunlight and therefore, less vitamin D skin production [19]. For this reason, prospective studies are preferable and avoid issues related to assessment of vitamin D levels after the development of advanced disease.

Relationship between Dietary Vitamin D and AMD
To date, interventional studies assessing the effect of vitamin D supplementation in preventing the onset or progression of AMD are lacking. Thus, there are no specific dietary recommendations regarding vitamin D for primary or secondary prevention of AMD, despite the high prevalence of vitamin D deficiency or insufficiency in the general population [81,82]. However, the relationship between dietary vitamin D and AMD has been investigated in several studies. Parekh et al. found that milk intake (fortified in vitamin D in the USA) was inversely associated with early AMD (OR: 0.75; 95% CI, 0.6-0.9) and fish intake was inversely associated with advanced AMD (OR: 0.41; 95% CI, 0.2-0.9) [20]. In Caucasian male monozygotic twin pairs with discordant AMD phenotypes, Seddon et al. reported that higher dietary intakes of vitamin D (assessed using a food frequency questionnaire) were present in twins with less severe AMD (p = 0.01) and smaller drusen sizes (p = 0.05) relative to co-twins-adjusted for smoking and age [83]-providing evidence that vitamin D could be involved in the etiology of AMD. In a recent Japanese case-control study, patients with neovascular AMD and control subjects, randomly selected from the population, aged ≥65 years, were assessed, using a brief-type self-administered questionnaire (BDHQ) on diet history. Logistic regression analyses, adjusted for smoking, age, sex, chronic diseases, supplement use, and alcohol consumption, demonstrated that a low intake of vitamin D, together with other nutrients, including n-3 fatty acid, alpha-tocopherol, zinc, vitamin C, and beta-carotene, was significantly (p < 0.001) associated with neovascular AMD [84].

Genetic Link between Vitamin D and AMD
Recent insight into pathophysiologic triggers has indicated that AMD involves the interaction of multiple genetic and environmental factors [5,12,85]. The interrelationship between gene polymorphisms and vitamin D has been suggested in several studies, providing additional links between vitamin D and AMD pathophysiology. A study performed in extremely discordant siblings (i.e., one sibling had the neovascular form of AMD while the discordant sibling had no sign of AMD and was older than 65 years) suggested that a single point variant in CYP24A1 (the gene encoding the catabolizing enzyme of the vitamin D pathway) influenced AMD risk, although this association has not been replicated [40]. Several gene polymorphisms in CYP2Rl (the 25-hydroxylase which converts vitamin D into 25(OH)D) and in VDR were also suggested to increase risk [86]. Moreover, an interaction between vitamin D status and single nucleotide polymorphisms (SNPs) in the complement factor H (CFH Y402H) and complement factor I (CFI) genes has been suggested [86]. Overall, compared to subjects with adequate blood 25(OH)D levels (i.e., ≥30 ng/mL), there was a 2.6-fold (95% CI: 1.3-5.2) increased likelihood of AMD in 25(OH)D deficient subjects (i.e., <12 ng/mL), a 3.4-fold (95% CI: 1.1-10.9) increased likelihood in subjects carrying one risk allele for CFH Y402H and a 6.7-fold (95% CI: 1.6-28.2) increased likelihood in subjects carrying two risk alleles. Similar trends were found with a CFI SNP. This is noteworthy, since these two proteins-CFH and CFI-work in interconnection to inhibit the complement cascade, which plays a central role in drusen formation in AMD [3,43]. Therefore, in the case of vitamin D deficiency, the ability to suppress the local inflammatory response may be impaired, leading to an increased risk of AMD. Another gene product, HTRA1, a serine protease which was shown to be expressed in the drusen of AMD patients [87], may be upregulated by vitamin D. A gene polymorphism in a vitamin D-responsive element in the promoter region of the HTRA1 gene may suppress this effect [88]. These possible associations support a link between vitamin D and genetic susceptibility to AMD.

Conclusions and Future Perspectives
Vitamin D can be considered a steroid hormone which binds to high affinity receptors. Experimental studies have suggested that vitamin D can control the expression of genes involved in oxidative stress, inflammation, and angiogenesis. In the macula, vitamin D may preserve the function of the retinal pigmentary epithelium and choroidal cells, through a paracrine/autocrine pathway. It is thus possible that the bioavailability of 25(OH)D circulating in blood is a limiting step in the protective effect of vitamin D. On the other hand, observational studies, including population-based studies, suggest an association between vitamin D deficiency and a higher risk of early and/or late AMD. This is consistent for a role of vitamin D in the pathophysiology of AMD. The potential causal association between vitamin D and AMD encourages future clinical research. In the interim, due to insufficient data, there is still no recommendation to screen for vitamin D deficiencies in patients at risk of AMD. However, all individuals may benefit by increasing their levels of vitamin D, through all possible means, including sun exposure, dietary recommendations, vitamin D-enriched foods, and vitamin D supplementation.