Inside out: Relations between the microbiome, nutrition, and eye health

Abstract

Age-related macular degeneration (AMD) is a complex disease with increasing numbers of individuals being afflicted and treatment modalities limited. There are strong interactions between diet, age, the metabolome, and gut microbiota, and all of these have roles in the pathogenesis of AMD. Communication axes exist between the gut microbiota and the eye, therefore, knowing how the microbiota influences the host metabolism during aging could guide a better understanding of AMD pathogenesis. While considerable experimental evidence exists for a diet-gut–eye axis from murine models of human ocular diseases, human diet-microbiome-metabolome studies are needed to elucidate changes in the gut microbiome at the taxonomic and functional levels that are functionally related to ocular pathology. Such studies will reveal new ways to diminish risk for progression of- or incidence of- AMD. Current data suggest that consuming diets rich in dark fish, fruits, vegetables, and low in glycemic index are most retina-healthful during aging.

1. Introduction

Age-related macular degeneration (AMD) is the leading cause of blindness in people over age 70 in the industrialized world, and effective therapies are limited. AMD is characterized by the progressive deterioration of the macula, an anatomical region of the retina that contains the highest density of cone photoreceptors and is responsible for high resolution visual acuity. AMD is a complex, progressive, retinal degenerative disease influenced by environmental, nutritional, and genetic factors, with age as the major risk factor. Current therapies for AMD target late stages that usually exhibit neovascular complications and fluid accumulation (Landowski and Bowes Rickman, 2021). Dysregulated lipid metabolism, diet and age-related disruption of glucose/-carbohydrate homeostasis and simple dietary insufficiency of several nutrients have been implicated in AMD pathogenesis suggesting points of intervention (Barbosa et al., 2014; Chiu and Taylor, 2011; Fragiotta et al., 2021; Kuan et al., 2021; Wang et al., 2010; Weikel et al., 2012).

Recently defined anatomic characteristics help refine our appreciation of early AMD. Cholesterol and cholesterol esters are lipophilic. They are central in the generation of drusen, their location in cone rich regions suggests pathologic dysregulated local and even systemic cholesterol homeostasis. While some epidemiological reports link intake of a high fat diet, drusen biogenesis and AMD, preclinical evidence points to vital role of local production of lipoproteins at the retinal pigment epithelial cells (RPE), thus, challenging the notion of a key role for dietary fat in prevention of AMD. Specifically, the RPE expresses the genes needed to synthesize and secrete ApoB100 and apolipoprotein E (Apo E) (Wang et al., 2006).

ApoB is a core backbone protein of hepatic VLDL and intestinal chylomicrons and can produce the large lipoprotein-like particles that accumulate with age in Bruch’s membrane (Fig. 1). Lipoprotein lipids are assembled not only from RPE, but also photoreceptor outer segments. There is also evidence for systemic lipid dysfunction in AMD and the microbiome affecting systemic lipids. Studies of human donor eyes show that large cholesteryl ester-rich lipoproteins accumulate in macular Bruch’s membrane of normal individuals starting as early as late adolescence (Curcio et al., 2009, 2010). Basal laminar deposit (BLamD) are lipid deposits between the RPE basal lamina (BL) and RPE. BLamD are clinically observed by OCT as a non-neovascular “split RPE-BL--Bruch’s membrane (BrM) complex” or “double-layer sign” (Sura et al., 2020). In comparison, Basal linear deposits or “BlinD” are lipid deposits in the inner collagenous layer. Drusen are the clinically seen version of BlinD (Curcio et al., 2009, 2010). High risk drusen are clustered centrally (central subfield and inner ring of the Early Treatment Diabetic Retinopathy Study (ETDRS) (Curcio, 2018b; Curcio et al., 2009; Pollreisz et al., 2021; Reiter et al., 2021; Wang et al., 2007), implicating the vulnerability of foveal cones and Müller glia, both of which are packed densely in the central fovea.

Fig. 1. Circulating lipids and RPE-derived lipids can contribute to pathological subretinal deposits.

Large cholesteryl ester-rich lipoproteins accumulate in macular Bruch’s membrane (BrM). These accumulate between the inner collagenous layer of Bruch’s membrane and the basal lamina of the RPE, where they fuse and pool to form drusen. Lipids from both systemic lipoproteins (dark green circles) and RPE derived lipoproteins (light green circles) contribute to drusen formation. Systemic lipoproteins are affected by the microbiome. The RPE is genetically capable of producing the large lipoprotein-like particles found in Bruch’s membrane and in drusen. It expresses apolipoprotein B and microsomal triglyceride transfer protein (MTTP) genes and proteins. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

Support for systemic lipid metabolism affecting BrM include observations that the entire retina and isolated outer segments have higher levels of the fatty acid docosahexaenoate than linoleate, whereas extracts of BrM or isolated lipoproteins from BrM have much higher levels of linoleate than docosahexaenoate. These findings were interpreted as evidence for dietary influence over BrM lipid accumulation (Curcio, 2018b; Curcio et al., 2009). In support of this, two studies have shown that well-differentiated and polarized RPE cells release apolipoprotein E, lipids, and minerals in Transwell chambers, even without the RPE cells being fed outer segments (Johnson et al., 2011; Pilgrim et al., 2017), suggesting diet has a role in formation of RPE-derived lipid deposits. Some investigators believe that the foveal centration of high-risk drusen invokes a dietary contribution (Curcio, 2018a); however this remains controversial.

As indicated in its name, a feature of the macula lutea, or macula, is its yellow pigment (lutein and zeaxanthin), which is highly concentrated in the foveal center and plexiform layers extending into the parafovea. These xanthophyll carotenoids may be derived from diet and transported in plasma on circulating HDL and LDL particles (Loane et al., 2008). RPE have the capacity to uptake lutein and zeaxanthin from these particles via the scavenger receptor B1 (SRB1) and the LDL receptor and transfer them to the neurosensory retina (Harrison, 2019). The unused lipids are thought to be repackaged and released basolaterally as part of the large lipoproteins described above. Accumulation of this material may contribute to the nucleation of drusen by invading dendritic cells, as best shown in peripheral retina (Rudolf et al., 2008). Although further experimental confirmation and exploration are needed, hypotheses regarding dietary etiologies for AMD are intriguing and may provide insight into why the central macula is particularly vulnerable to high-risk drusen and why diet affects AMD progression. Additional dietary components that are associated with risk for onset or progress of AMD are discussed below.

2. The gut and its microbiome

The “microbiome” refers to the collective genomes of the micro-organisms in a particular environment whereas the “microbiota” is the actual community of micro-organisms representing approximately one hundred trillion micro-organisms, mostly bacteria but also viruses, fungi, and protozoa. While the human genome consists of about 23,000 genes, the microbiome encodes over three million genes producing thousands of key metabolites which influence the fitness, phenotype, and general health of the host. The largest microbiome resides within the digestive tract, where its genetic material is collectively referred to as the gut microbiome. The resident microbiota communities vary in number and diversity in luminal vs. mucosal locations (Albenberg et al., 2014; Friedman et al., 2018) along the longitudinal axis of the gut (Fig. 2A).

Fig. 2. The microbiota and gut physiology.

2A) The microbiota differs in luminal versus mucosal locations and by region of the gut. The microbiota community number and diversity vary along the longitudinal axis of the gut. The stomach has a unique microbiome. Descending from the stomach, the duodenum is a 20–25 cm long segment that begins the small intestine. It receives episodic food boluses from the acidic stomach and is subject to the rapidly changing conditions of metabolite load, digestive enzyme concentrations, and bile acids. Food digestion with absorption of nutrients primarily occurs in the jejunum and ilium. The entire small intestine receives input from the pancreas and liver. Beyond the small intestine is the large intestine or colon which has a unique microbiota compared to the small intestine. 2B) Diet and medication usage are dominant shapers of the gut microbiome. Although associated with chronological age, each individual and tissues within an individual age biologically at a different pace. Biological aging reflects the cumulative erosion damage of organs or systems that is dependent on genetics, environment, and lifestyle. Aging may also reflect medication use, access to food, nutrient quality, physical activity as well as lifestyle, including congregant or independent living.

The number of microbial species detected as specific operational taxonomic units (OTUs) and when measured within an individual’s intestine, termed “alpha diversity,” can dramatically differ among subjects, however it is typically used to compare experimental cohorts. While diet, environment, and medication usage are dominant shapers of the gut microbiome (Francisco et al., 2020; Rowan et al., 2017; Weikel et al., 2012), there are also microbe-microbe interactions through the gut (Albenberg and Wu, 2014; Muegge et al., 2011; Rothschild et al., 2018; Sholl et al., 2021; Zhernakova et al., 2016).

Food digestion with absorption of nutrients primarily occurs in the jejunum and ilium, distal sections of the small intestine. The small intestinal microbiota metagenome is enriched in genes related to carbohydrate metabolism as compared with the fecal metagenome (Zoetendal et al., 2012). The microbiome within the jejunum and ileum also modulates systemic lipid homeostasis (Lamichhane et al., 2021) and is a source for anti-inflammatory short chain fatty acid (SCFAs) metabolites. SCFA, consisting of chains of 2–7 carbons, are the major bacterial products of gut-based anaerobic fermentation of dietary carbohydrates and proteins resistant to degradation in the small intestine (Johnson et al., 2019). The production of SCFA can be augmented by increased consumption of fiber.

SCFAs are associated with many functions. They are an important fuel for intestinal epithelial cells and affect water and electrolyte absorption (Blaak et al., 2020; Vinolo et al., 2011). The production of SCFAs modulate gut hormone release, impacting lipid and lipoprotein metabolism (Gribble and Reimann, 2019). SCFAs suppress inflammation, promote antibody production, resist pathogen invasion, and modulate energy expenditure (Rothschild et al., 2018). The most copious of the SCFAs detected in fecal samples are acetate, propionate, and butyrate (Barton et al., 2019). Butyrate is the primary source of energy for the colonic epithelial cell. Propionate can be converted to glucose by gluconeogenesis in the intestine (Francisco et al., 2020). Acetate serves as a co-factor/microbial metabolite enhancing the growth of other bacteria, but it is also distributed to peripheral tissues (Rothschild et al., 2018) where it is used for cholesterol metabolism and lipogenesis. Acetate and propionate also contribute to appetite regulation (Zhernakova et al., 2016). Dietary fiber intake and probiotic intake favorably impact the production of SCFA and antioxidants, modulate lipid metabolism, and enhance detoxification pathways (such as indoxyl sulfate) (Hung et al., 2017).

The epithelial barrier of the intestine includes a glycocalyx of mucins and other glycoproteins that coat the intestinal epithelial cells and trap bacteria, limiting their further penetration; thus, protecting the host from microbe invasion and potential systemic bacteremia. Included in the epithelial layer, at the bottom of the intestinal crypts, are Paneth cells. These immunomodulatory cells secrete anti-microbial peptides including defensins. In response to bacterial entry, intestinal epithelial cells act as microbial sensors and secrete interleukins-8 (IL-8), monocyte chemoattractant protein-1 (MCP-1), RANTES (an IL-8 family cytokine), tumor necrosis factor-alpha (TNF), and IL-6 that direct the recruitment of immune cells including IgA-secreting plasma cells, CD4⁺ and CD8⁺ T cells and regulatory T cells (Tregs) (Keppler et al., 2021) to the gut. Local, non-inflammatory macrophages ingest and kill the rare commensals which enter the intestinal epithelium. On the other hand, pathogenic bacteria express pathogen associated molecular patterns (PAMPs) which are small molecular motifs conserved within a class of microbes. PAMPs extravasate across the leaky gut epithelium and are recognized by toll-like receptors (TLRs) and other pattern recognition receptors (PRRs), serving to intensify inflammation. When PAMPs enter the circulation, they travel to the bone marrow. Bone marrow stem/progenitor cells express TLRs and are responsive to systemic levels of the microbial products, particularly LPS and peptidoglycan (PGN) (Joshi et al., 1979; Khosravi et al., 2014; Nicaise et al., 1998; Tada et al., 1996). Further, steady state myelopoiesis is dynamically regulated by microbiota products (Balmer et al., 2014; Bosco and Noti, 2021; Odamaki et al., 2016). Thus, linking the gut to the systemic immune response.

3. Relations between the gut microbiome and AMD

In aged populations a decrease in appetite, loss of teeth, and decreased efficiency of the digestive system all serve to modulate the gut microbiome (Bosco and Noti, 2021; Odamaki et al., 2016). Aging affects the microbiome and is an “uneven” process with different individuals aging at different rates and tissues within an individual also aging to different degrees. The rate of aging is dependent on genetics (Martin et al., 2007), medication use, access to food, nutrient quality, physical activity as well as lifestyle, including congregant or independent living (Fig. 2B). Chronological aging and biological aging progress in tandem, but clearly there is individual variability, with some people showing signs of biological or functional decline at earlier chronological ages than others. The intestinal epithelium, although composed of only a single cell layer, forms a robust barrier that typically protects against the penetration of all microbes. However, defects limiting its effectiveness occur in physiological aging which is associated with chronic inflammation (Cox et al., 2021; Langfeld et al., 2021). While physiological aging impacts the nature of commensal bacteria, these bacteria are typically not harmful and are “noninvasive” (Mosconi et al., 2013). Thus, diminished intestinal barrier function contributes to increased gut permeability driving not only local inflammation, but microbial extravasation to systemic sites results in “inflammaging.” Inflammaging is defined as low-grade chronic systemic inflammation established during physiological aging which may contribute to AMD pathology. Specific relations between diet, the microbiome and risk for AMD are discussed below.

4. Complement cascade, inflammation, and the microbiome

In addition to inflammaging, there are associations between risk for AMD and malfunction of the complement cascade. It is plausible that gut dysbiosis results in chronic over-activation of the complement cascade. For example, Lin and colleagues have recently shown the correlation between ARMS2 and CFH risk alleles. Lin also found significant gut microbial alterations in those with late AMD (Lin, 2019). The ARMS2 risk allele was associated with significantly higher rates of IgA-bound gut bacteria, and this was interpreted as directly impacting the host immune system since IgA binds bacteria and protects against pathogens (Lin et al., 2021). Corroborating these findings, individuals with wet AMD show altered microbiome signatures relative to controls (Zinkernagel et al., 2017; Zysset-Burri et al., 2020). Shotgun metagenomics analysis of fifty-seven neovascular AMD and fifty-eight healthy controls indicated that Negativicutes was more abundant in AMD subjects whereas the genus Oscillibacter and species Bacteroides were more prevalent in individuals without AMD. They also identified that SNPs within the complement factor B gene were more abundant in healthy individuals. In contrast, SNPs within the elevated temperature requirement A serine peptidase 1 and complement factor H (CFH) genes were seen in subjects with neovascular AMD (Zinkernagel et al., 2017; Zysset-Burri et al., 2020). Adding a metabolomic observation, the intestinal bacteria associated with AMD subjects appeared to be related to fatty acid metabolism and carotenoid biosynthesis when compared with controls. These findings provide possible mechanism for how gut microbiota may contribute to the pathogenesis of AMD (Chew, 2017).

5. Using the microbiome to predict eye disease

While some human gut bacteria seem to be associated with “good health” (e.g., Faecalibacterium prausnitzii) and others universally with “poor health” (e.g. Fusobacterium nucleatum), for most organisms, relative advantage is context-dependent (Conway and N, 2021). The aging microbiome is characterized by a lack of resilience and diversity (O’Toole and Jeffery, 2015). In people who have healthy aging, Akkermansia is increased while Faecalbacterium, Bacterodaceae and Lachnospirraceae are decreased (Badal et al., 2020). In contrast, age-related dysbiosis is characterized by a loss in Clostridialis and Bifidobacterium and elevation of proteobacteria such as enterobacteriaceae (Jeffery et al., 2016; O’Toole and Jeffery, 2015). In the “oldest” (>90 years old) adults that remain robust, butyricimonas and other taxa associated with greater anti-inflammatory properties tend to be increased. The microbiome of individuals with “healthy” aging is also associated with more SCFA, enhanced pathways related to cellular respiration, and increased vitamin synthesis (Badal et al., 2020).

A flurry of publications in the last 5-years implicates and connects the gut microbiome to several retinal diseases. These studies began as logical extensions of research connecting diet to AMD. Initial findings connecting gut microbiome to AMD came from a murine study linking a high-fat diet to an exacerbation of laser-induced choroidal neovascularization (CNV) in the mouse (Andriessen et al., 2016).

It is now appreciated that there is also an ocular microbiome independent of the ocular surface microbiome. Current dogma in ophthalmology and vision research presumes the intraocular environment is sterile. However, intestinal bacterial products translocation into the bloodstream and into many internal organs including healthy eyes and eyes from diseased animal models, suggests that a microbial community may also inhabit the intraocular cavity (Wen et al., 2018). Deng et al. evaluated intraocular samples from over one thousand human eyes. They used quantitative PCR, negative staining transmission electron microscopy, direct culture, and high-throughput sequencing technologies to demonstrate the presence of intraocular bacteria. Importantly, they excluded the possibility that the microbiome from these low-biomass communities could be contaminated from other tissues and reagents (Deng et al., 2021). They detected a disease-specific microbial signature for the intraocular environment of individuals with AMD and glaucoma, suggesting that either spontaneous or pathogenic bacterial translocation may be associated with these conditions (Deng et al., 2021). Prasad et al. showed the importance of the plasma microbiome and the intraocular microbiome in a rodent model of diabetic retinopathy (Prasad et al., 2022). The presence of an intraocular microbiome was established after birth in normal eyes from non-human primates and a variety of species including rat, rabbit, and pig (Deng et al., 2021).

6. “Let food be thy medicine and medicine be thy food” Hippocrates

So, what should we eat to maintain healthy eyes and a healthful microbiome upon aging? Various patterns of diet have been evaluated for their impact on disease progression (Chiu et al., 2014; Francisco et al., 2020). More recently, relationships between specific nutrient intake, the gut microbiome, the metabolome and retinal vitality have been explored and it was observed that many metabolites produced by gut microbiota reach the retina and potentially impact retinal function (Rowan et al., 2017). The host benefits when gut microbiota express enzymes and generate metabolites that are useful but otherwise absent (Rowan et al., 2017; Sedghi et al., 2019). Here, we first recall basic digestion and discuss data relating intake of individual nutrients to AMD. Then, we summarize information about diets, the microbiome, the metabolome and risk for onset or progress of AMD.

Some portion of the macronutrients (carbohydrates, proteins, and fat) can reach the colon if the intake surpasses the rate of digestion in the upper gastrointestinal tract or if they are not digested by host enzymes (Sedghi et al., 2019). Whole foods (i.e., whole grains) as opposed to highly processed foods (i.e., white breads, pastries, high fructose sweetened beverages) provide macronutrients to the colonic microbiota because of the proportion of their composition that is indigestible in the upper intestine (Francisco et al., 2020; Rowan et al., 2017). As compared to the human host, gut bacteria possess more enzymes able to degrade dietary polysaccharides; thus, the host relies on the microbiome to garner energy from such complex carbohydrates (Flint et al., 2008). Conversely, protein digestion by proteases within human intestines is very efficient (Mills et al., 2019; Sedghi et al., 2019).

While macronutrients are energy-providing components of the diet, micronutrients work in concert to promote metabolic activity, cellular growth, and differentiation (Muegge et al., 2011). The bioavailability of micronutrients to the host is also determined in part by gut microbial metabolic processes (Sedghi et al., 2019). In addition to vitamins, some key micronutrients are plant polyphenols including flavonoids such as quercetin and anthocyanins, and non-flavonoids such as resveratrol (Bosco and Noti, 2021). Polyphenols exhibit little bioavailability in the small intestine and thus pass into the colon where they are catabolized by microbiota to smaller bioactive molecules and absorbed into the circulation (Bosco and Noti, 2021; Cardona et al., 2013; Filosa et al., 2018; Odamaki et al., 2016; Zhernakova et al., 2016). Polyphenols can also exhibit effects without being absorbed by chelating iron, for example. They may also serve as prebiotic metabolites, inhibiting unwanted bacteria and favoring beneficial bacteria (Xiao et al., 2020). Water-soluble vitamins provided in the diet are absorbed in the small intestine while others are synthesized in the colon (DeJong et al., 2020). For example, colonic microbiota synthesize vitamin K and some B vitamins (Buford, 2017; DeJong et al., 2020). Microbially produced vitamins are utilized by the host and by other non-vitamin producing bacteria (Buford, 2017; DeJong et al., 2020).

One water-soluble compound and amino acid analog, taurine, is present in the retina at extremely high concentrations, and taurine is essential to eye health. Taurine is also present at exceedingly elevated levels in the eye lens, a tissue with far slower overall metabolism than the retina. Decades ago, it was demonstrated that taurine deficiency led to retinal based blindness in cats, and recent experiments indicate that there are interactions between taurine and drugs that are also related to vision (Froger et al., 2014; Pardue and Allen, 2018). More recently, taurine and its deoxycholate conjugate were shown to have neuroprotective effects.

Vitamin A and vitamin D deficiencies result in less diverse microbial populations, having adverse effects on the gut. Conversely, supplementation with vitamin D was shown to increase fecal levels of SCFA. Relative sufficiency versus deficiency in calcium, magnesium, and zinc can also re-shape the microbiota (Wilmanski et al., 2019). Agron et al. recently found that elevated intake of vitamins A, B6, C, folate, beta-carotene, lutein, zeaxanthin, magnesium, and copper were also associated with decreased risk for AMD (Agron et al., 2021).

7. Mediterranean diet, AMD, and microbiome

The role of diet and AMD has been discussed in the literature for decades and multiple reviews have summarized epidemiologic and clinical studies (Francisco et al., 2020; Weikel et al., 2012). Dietary studies that followed large numbers of subjects are noted below. The classical “Mediterranean” diet is rich in green vegetables and fish. Dark fish are a rich source of omega-3 polyunsaturated fatty acids fatty acids (PUFAs). These include eicosapentanoic acid and docosahexaenoic acid and are anti-inflammatory. In comparison, omega-6 PUFAs such as docosapentanoic acid, arachidonic and linoleic acids are often associated with proinflammatory effects, albeit they are also found in many “healthy” foods. Omega-6 PUFAs have been shown to increase NF-kB activity influencing production of interleukins, TNF-α and MCP-1 (Patterson et al., 2012; Wall et al., 2010). Over the last several decades, the content of omega-6 fatty acids in typical American diets has increased. In older adults, along with protection against AMD, a diet rich in micronutrients, fruits, vegetables, and fish, and low in saturated fats–such as the Mediterranean diet–is more commonly found in countries with long life expectancies.

The impact of the Mediterranean diet on AMD has been evaluated in many studies (Barber et al., 2021; Beam et al., 2021; Clark et al., 2021; Keenan et al., 2020; Merle et al., 2015, 2019). Most studies point to a beneficial effect associated with a decreased risk of progressing to late AMD by 25–50% in consumers of such micronutrient rich diets. Consuming the Mediterranean diet is particularly associated with reduced progression to geographic atrophy, a condition for which no effective therapy exists (Keenan et al., 2020). Higher genetic risk for AMD was associated with lower intestinal bacterial diversity, an indicator of compromised health (Lin et al., 2021). Intake of fatty fish such as tuna, herring, etc., at least twice per week, has been routinely found to be associated with decreased risk of developing AMD (Barber et al., 2021; Beam et al., 2021; Chiu et al., 2009; Chong et al., 2008; Christen et al., 2011; Clark et al., 2021; Keenan et al., 2020; Zhu et al., 2016). Encouragingly, individuals with high genetic risk can reduce their risk of late AMD by eating a favorable (Mediterranean) diet (Colijn et al., 2021; Ho et al., 2011). This indicates that diet is a strong means to counterbalance the otherwise extremely high genetic risk. Since diet and microbiome are related a role microbiota is at the intersection of genetics and AMD.

The Mediterranean diet also plays a role in shifting gut microbial composition. It increases the number of SCFA-producing bacteria that are considered to have anti-inflammatory properties. Communities of subjects who have more diversified microbiota or higher beta-diversity (of their microbiotas) tend to have less dysbiosis. As described below, Mediterranean diets also tend to be lower in simple carbohydrates than typical American diets. This may be related to the similar beneficial effects on AMD risk for low glycemic index diets (Albenberg and Wu, 2014; Chiu et al., 2009, 2014).

Additional insights and proposed interventions were gleaned from multiple studies. In 1994, the Eye Disease Case-Control Study (EDCC) indicated that patients who consumed the highest quintile of green leafy vegetables and foods containing carotenoids (lutein, zeaxanthin) were 43% less likely to have advanced AMD than controls (N = 356 cases, N = 520 controls). Data from the large and prospective Nurses’ Health Study and the Health Professionals Follow-up Study showed 54% higher risk for vision-impacting AMD in the highest vs. the lowest quintiles of total fat intake and intake of linolenic acid (omega-6). Conversely, ≥4 servings of fish per week were associated with a 35% lower risk (Seddon et al., 1994). Importantly, the benefit gained from eating fish with high omega-3 fatty acids is in addition to that provided by consuming a lower glycemic index diet (see below). Therefore, approaches to manipulating dietary risk for AMD are more complex than simply increasing the intake of omega-3 PUFAs (Chiu et al., 2009).

Measures of overall dietary quality have also been related to risk for AMD. The Healthy Eating Index (HEI) is a measure of diet quality used to assess how well a set of foods aligns with key recommendations of the Dietary Guidelines for Americans (DGA) (Krebs-Smith et al., 2018). The 2015–2020 Dietary Guidelines for Americans include explicit recommendations to limit intakes of both added sugars and saturated fats to <10% of energy and was developed as a general measure of diet quality (Snetselaar et al., 2021). Using HEI and the Alternate Healthy Eating Index (AHEI; adjusted to capture dietary factors specifically related to cardiovascular disease risk), Montgomery et al. showed that AHEI participants in the highest quartile of diet quality had a significantly reduced odds of AMD (0.54, 95% confidence interval 0.30–0.99) and that HEI participants had non-significantly reduced odds of AMD (0.75, 0.41–1.38). Odds of AMD were also 51% lower in the highest quartile of fish intake compared to the lowest quartile (odds ratio = 0.49, 0.26–0.90). They concluded that HEI and AHEI score may be a useful instrument for assessing AMD risk due to diet, and both scores could potentially be improved by incorporating more specific information regarding micronutrient intake (Montgomery et al., 2010).

In summary, data from multiple studies of diet and AMD demonstrated protective (left side of unity) or harmful (right side of unity) relationships between specific dietary intake patterns and risk of AMD (Fig. 3) (Francisco et al., 2020). Diets rich in dark-yellow, cruciferous, and green leafy vegetables, legumes, fruits, other vegetables, whole grains, tomatoes, fish, poultry, soup, low-fat dairy were collectively protective for AMD, as were Mediterranean diets (rich in olive oil, fresh produce, nuts, legumes, fish, and fruits).

Fig. 3. Odds ratios or relative risk ratios for age-related macular degeneration derived from multiple studies.

Summary of studies of dietary patterns and AMD. Diets were grouped into four overall classes and difference among them were associated with AMD. Prudent diets (high in dark-yellow, cruciferous, and green leafy vegetables, legumes, fruits, other vegetables, whole grains, tomatoes, fish, poultry, soup, low-fat dairy) were collectively protective against AMD, as were Mediterranean diets (high in olive oil, fresh produce, nuts, legumes, fish). Healthy Eating Index (HEI) is a measure of diet quality used to assess how well a set of foods aligns with key recommendations of the Dietary Guidelines for Americans (DGA) (Krebs-Smith et al., 2018). The 2015–2020 DGA include explicit recommendations to limit intakes of both added sugars and saturated Fats to <10% of energy and was developed as a general measure of diet quality (Snetselaar et al., 2021). Western diets (high in red meat, saturated fats, highly processed foods, sweets and desserts, sugar-sweetened beverages) were all highly associated with AMD risk. Risk for smoking is added for comparison. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

8. AREDS, AMD and the microbiome

The largest and most systematic evaluation of effects of micronutrients on risk for AMD were done in the Age Related Eye Diseases Studies (AREDS). The first AREDS study evaluated an oral supplement containing vitamin C, vitamin E, beta-carotene, and zinc, in a large sample (N = 3640 persons). The supplement reduced the risk of advanced AMD, primarily neovascular AMD, by about 25% over a 5-year period in participants with intermediate AMD. There was no effect in participants with early or no AMD (Age-Related Eye Disease Study Research, 2001). AREDS2 was a second trial (4203 participants) that evaluated an oral supplement containing lutein and zeaxanthin, in addition to vitamins C and E, and evaluated whether beta-carotene could be eliminated, or zinc reduced. A subgroup receiving lutein and zeaxanthin without beta-carotene were 25% less likely to progress to neovascular AMD, although there was no statistically significant reduction in the risk of progressing to geographic atrophy (Age-Related Eye Disease Study 2 Research, G, 2013; Chew et al., 2014). Consequently, the doses of 10 mg of lutein, 2 mg of zeaxanthin, 80 mg of zinc oxide, 2 mg of cupric oxide, 500 mg of vitamin C, and 400 IU (268 mg) of vitamin E (these levels are 5-10-fold higher than usual dietary intake), as used in the AREDS2 study have become the standard-of-care for individuals at significant risk for visual loss from AMD (Gibbons et al., 2017). While the AREDS were not “diet” studies per se, they support observations that individuals who consumed higher amounts of fruits, vegetables, and fish are most protected from developing early or advanced AMD.

Recalling the inflammation-microbiome-AMD connections noted above, it is of interest that vitamins E and C can promote an anti-inflammatory gut microbiome in humans, reducing risk for systemic oxidative stress, which could conceivably lessen risk for AMD (Chew, 2017; Lin, 2019; Poyet et al., 2019). Zinc is a cofactor for numerous antioxidant enzymes in both the host and gut microbiome. Studies in laboratory animals demonstrate large alterations of the composition and diversity of the gut microbiome in response to both zinc deficiency and high-dose supplementation (Bielik and Kolisek, 2021; Pajarillo et al., 2021). Considering the extremely high dose of zinc in AREDS2 formulations, the impact of zinc supplementation on the gut microbiome in AMD patients deserves further investigation. Lutein and zeaxanthin are of particular interest because they are widely utilized by gut micro-organisms and even constitutively synthesized by some of them, combating oxidative stress. Thus, there is likely to be a complex interplay between the host and gut microbiome in response to supplementation with 5–10 times the average American’s dietary intake of lutein and zeaxanthin in AREDS2 vitamins. This is further complicated by the metabolic regulation of carotenoid cleavage and uptake in the intestine of these lipid soluble micronutrients (Duvallet et al., 2017). One recent study in pregnant women found that dietary and plasma carotenoids are positively associated with alpha diversity in their fecal microbiota (Pakpour et al., 2017) and similar studies should be considered in the future for the AMD population. A caution is worth mentioning. Humans actively accumulate xanthophyll carotenoids in the eye and other tissues, relative to most other mammals, thus, human studies are preferred to laboratory animal studies.

9. Western diet, behavior, AMD risk and microbiome

As noted earlier, Americans, generally, eat a more carbohydrate-rich and fat-rich diet than was consumed by Mediterranean cultures. Western diets (rich in red meat, saturated fats, highly processed foods, sweets and desserts, sugar-sweetened beverages) were associated with increased AMD risk (Chiu et al., 2014; Froger et al., 2014) and the risk for AMD associated with consuming Western diets is as high as the increased risk associated with smoking (Fig. 4). Moreover, the elevated ratio of omega-6 to omega-3 PUFA in the Western diet (Patterson et al., 2012) has potentially “pro-inflammatory” consequences. Additionally, staples of the Western diet including refined sugars and a variety of soft carbohydrate-rich foods that result in less diverse microbiomes, which also contain higher levels of some microbiota that raise glycated stress and have been described as unhealthy (Cordain et al., 2005; Edlund et al., 2015; Hujoel, 2009; Najeeb et al., 2016). Common ingestion of drugs such as proton pump inhibitors and antibiotics, can lead to altered gut pH and permeability with increased gut-derived toxins entering the circulation (Alhajri et al., 2021). Disease-driven gut pathology can increase the circulation of gut microbial antigens leading to increased systemic inflammation, insulin resistance, cognitive decline, and greater risk of CVD (Fig. 4).

Fig. 4. The gut can impact AMD progression.

The consumption of a lower glycemic index diet or ketogenic diet serves to increase beta-hydroxybutyrate and beneficially impacts gut physiology. These beneficial effects may offset the detrimental effects of AMD risk alleles, smoking, obesity, and a history of previously ingesting a western diet.

10. Low glycemic index diets and AMD

In addition to providing energy, sugars and their metabolites modify many biomolecules such as proteins, diminishing their functional capacity. Low glycemic index with more whole grains, vegetables, and fruit, have been evaluated regarding risk for AMD. The glycemic index measures how fast sugars from the diet enter the blood stream, relative to 50g of ingested pure sugar. Every human epidemiologic study published to date has found that consuming lower glycemic index diets preserves retina function upon aging. Consumers of lower glycemic index diets enjoy protection for development of all grades of AMD as well as for progression of early or intermediate AMD to more advanced stages (Chiu et al., 2007) (Chiu and Taylor, 2011) (reviewed in (Francisco et al., 2020)). To elucidate the etiology of these relationships, this was explored in multiple animal studies. Each showed that AMD-like retinal lesions were less prevalent in animals that consumed a lower glycemic index diet and, the lower glycemic index diet impacted the microbiome dramatically (Rowan et al., 2017, 2020). This allowed description of a diet-gut/microbiome-retina axis. Many of the microbial cometabolites affect retinal physiology. Importantly, changing from a higher to a lower glycemic index diet even at mature ages, could prevent progression of retinal lesions (Rowan et al., 2017), suggesting that changing from higher to lower glycemic index diets even at relatively late stages of life might bring benefit to the retina. The salutary effect of the lower glycemia diets was mechanistically attributed to lower levels of advanced glycation end products (AGEs) and other oxidized moieties, as well as metabolism that resulted in less accumulation of lipid products. Low glycemic index diets are also associated with reduced HbA1c, decreased fasting glucose, total cholesterol, LDL, high sensitivity C-reactive protein, and IL-6 in the blood, and decreased insulin resistance (Dave, 2019). Lower levels of lipid peroxidation products (4-hydroxynonenal, HNE) and lipofuscin were found in retinae in low glycemic diet fed mice (Rowan et al., 2017).

11. Summary

The etiology of AMD is multifactorial including nutritional factors with associated intestinal microbiome and metabolome, genetic variants in complement pathway and HTRA/ARMS pathways, and environmental risk factors, particularly, smoking. Both micronutrient and macronutrient consumption impact the progression and severity of AMD. “Food is medicine.” The aging population is more vulnerable to nutrient deficiencies. Risks for AMD due to poor nutrition are as great as the risk imposed by smoking (Figs. 3 and 4). Diets alter the gut microbiome, and it may soon be possible to associate specific microbial communities, gene function and metabolic products with AMD and dietary patterns given that it is now possible to determine individual responses to diet. To exploit this opportunity to optimize health and guide drug discovery, it will be advantageous to better define interactions between nutrition, metabolomes, microbiomes, physiological responses, and clinical ophthalmologic data. Such an “interactome” network will also help identify novel biomarkers of AMD progression. This may allow identifying the optimal gut microbiome for a particular individual and hasten the move toward personalized medicine. The positive effects of consuming lower glycemic index diets, increasing beta-hydroxybutyrate, and intermittent fasting are mediated through changes in the gut microbiome. Clinical studies using plasma metabolites (lipids, advanced glycation end products, lutein, markers of barrier dysfunction) and inflammatory molecules will help elucidate interactions between the gut microbiome and AMD. The future holds promise as the pursuit of longitudinal human studies will help define the dynamics of the diet, gut microbiome, and the metabolome in risk for AMD and its severity.

Glossary

Microbiome

The collective genomes of the micro-organisms in a particular environment

Microbiota

The community of micro-organisms themselves

Alpha diversity

Interchangeably used with the term ‘species diversity,’ summarizes the distribution of species abundances in each sample into a single number that depends on species richness and evenness

Beta diversity

Quantifies (dis-) similarities between multiple subjects (samples)

Operational taxonomic unit (OTU)

A definition used to classify groups of closely related organisms. DNA sequences can be clustered according to their similarity to one another, and operational taxonomic units are defined based on the similarity threshold (usually 97% similarity) set by the researcher

Germ-free animals (GFA)

Animals that have no micro-organisms living in or on them

Short chain fatty acids (SCFA)

Fatty acids with two to six carbon atoms that are produced by bacterial fermentation of dietary fibers and include formic acid, acetic acid, propionic acid, butyric acid, isobutyric acid, valeric acid, isovaleric acid and 2-methylbutyric acid

Glycemic index

The glycemic index (GI) is a rating system for foods containing carbohydrates. It shows how quickly each food affects your blood sugar (glucose) level when that food is eaten on its own. Blood sugar released over 2 h is compared to the blood sugar observed when the same subject consumed 50g of glucose

Data availability

No data was used for the research described in the article.