An overview of retinal light damage models for preclinical studies on age-related macular degeneration: identifying molecular hallmarks and therapeutic targets

1 Introduction

The visual system is involved in multiple disorders caused by the dysfunction of different components of the eye. The anterior part of the eye may develop cataracts (an opacification of the crystalline lens) or lesions of the cornea, with a consequent damage to the trigeminal nerve. Nevertheless, these damages could be repaired through surgery, corneal transplant or topical application of specific drugs (Kumar et al. 2022). Conversely, neurodegenerative disorders affect the retina, which is part of the sensory nervous system located at the back of the eye, with no regenerative capability. Retinitis pigmentosa (Newton and Megaw 2020), Stargardt’s disease (Tsang and Sharma 2018), glaucoma (He et al. 2018), diabetic retinopathy (Lin et al. 2020) and age-related macular degeneration (AMD) (Guymer and Campbell 2023) are the main degenerative diseases affecting the retina, with different aetiology and incidence, but with the common fate of causing vision loss. In this review, we will focus on AMD and the involvement of light as a risk factor triggering retinal neurodegeneration. We will discuss the use of light damage (LD) rodent models in preclinical studies to deeply explore AMD pathogenesis, therapeutics and molecular underpinnings, starting from the fundamental role of light for retinal function and then how it becomes harmful in certain conditions. We will review the main literature on the existing LD models and their ability to reproduce the features of human AMD both in early and late stages. We will also explore the advantages and limitations of the LD model for preclinical studies of AMD.

1.2 Age-related macular degeneration

Age-related macular degeneration is a multifactorial, neurodegenerative disorder characterized by a progressive impairment of the macula (Chichagova et al. 2018; Mitchell et al. 2018). This disorder is considered the major cause of blindness in Europe and developed countries, and its prevalence is approximately one person every 10, over 55 years (Zapata et al. 2021). It has been evaluated around 196 million cases for the 2020, with an estimated increase to 288 million by 2040 (Li et al. 2020; Wong et al. 2014). AMD consists of several alterations in neuronal retinal cells, RPE, BrM, choroid and retinal vessels. A lot of progress has been made in recent years, but to date, the exact pathogenesis of AMD is still under investigation, mainly to fully understand all the events occurring during the onset of the disease and its progression. According to clinical manifestations, AMD is classified in two main distinct forms (Figure 1): a dry form (atrophic or non-exudative) and a wet form (exudative or neovascular), and both can cause an irreversible visual impairment resulting in blindness in the most severe stages (Ambati and Fowler 2012). The most prevalent form is the dry AMD which accounts for 85–90 % of all cases (Evans and Syed 2013) and is characterized by a gradual vision loss. Although is less common (10 % of all cases), the exudative form of AMD is the most severe with a rapid progression, and it is characterized by abnormal choroidal neovascularization (CNV) in the macular region. The starting point of AMD is the dysfunction of RPE followed by photoreceptors death (Ferris et al. 2013; Wong et al. 2008). A Clinical Classification was proposed by Ferris et al. (2013) and it is based on the size and the degree of drusen deposits and pigmentary abnormalities, which are two main pathological hallmarks of AMD. According to this classification, AMD can be divided into early, intermediate (a critical clinical stage due to the increased risk for the progression of the pathology), and advanced stage. Starting from the intermediate stages, the pathology could be distinguished into dry or wet form (Ferris et al. 2013; Flaxel et al. 2020). The end stage of dry AMD is characterized by a typical lesion called geographic atrophy (GA) (Garcia-Garcia et al. 2022). On the other hand, in the wet or exudative form there is a dysregulated CNV. CNV could be classified into (i) type 1, which is typically localized under the RPE with its elevation, (ii) type 2 located in the subretinal space with the separation of the RPE from the neuroretina and (iii) type 3 characterized by oedema, haemorrhage and the aberrant connection between retinal and choroidal circulation (Spaide et al. 2020). The two forms of late AMD are not mutually exclusive; indeed, a secondary GA may appear following the neovascularization (Garcia-Garcia et al. 2022). A mixture of both types of AMD late-stages could be commonly found in the same eye (Stahl 2020).

Figure 1:

Schematic representation of the main features of DRY and WET AMD. Parts of the figure were drawn using pictures from Servier Medical Art. Servier Medical Art by Servier is licensed under a Creative Commons Attribution 3.0 Unported License (https://creativecommons.org/licenses/by/3.0/).

The incidence of AMD mainly correlates with the age (Blasiak et al. 2022; Jonasson et al. 2014), and lifestyle; indeed, smoking, light, higher body mass index with hyper lipidic diet, comorbidity with cardiovascular disease, previous cataract surgery, chronic kidney disease, hypertension, hyperthyroidism, diabetes, Parkinson’s and Alzheimer’s disease, genetic predisposition and racial background are classified as risk factors, as detailed elsewhere (Heesterbeek et al. 2020b). Recently thanks to technological advancements, including the single-cell sequencing (sc-RNAseq), genetic factors had also been identified to be involved in AMD development and progression (Collin et al. 2023; Voigt et al. 2021, 2022). Furthermore, several genetic studies have significantly contributed to the knowledge on AMD pathology. To date, the genomic heritability can be elucidated by both common and rare genetic variants in less than 40 genetic loci, which are primarily involved in the complement system, extracellular matrix remodelling, and lipid metabolism (reviewed by Dietzel et al. 2014). The potential benefit of identifying all risk factors could be the ability to predict the disease progression in individual patients, but mainly to improve the design of future clinical trials for the development of new effective therapeutic strategies. The following two paragraphs will describe the dual role played by light in its interaction with the visual system: the light as a necessary stimulus for vision, but also the light as a stressor to induce retinal degeneration.

1.3 Light as an environmental stimulus for visual perception

Light is fundamental for the correct functioning of the visual sensory system, which is sensitive to a spectrum of wavelength between 400 and 760 nm. The phototransduction is triggered when photons of light reach the retina and are focused on the macula (Figure 2). In the retina, photoreceptors convert light into graduated electrical signals, which become action potentials in the ganglion cells to be transferred to the higher visual cortex. The light triggers specific photochemical reactions which take place in the discs of photoreceptors outer segments. Rhodopsin and opsins are two kinds of photopigments rod and cone respectively. As reported in Figure 3, phototransduction involves six consequential steps: in the Step 1 the light converts the rhodopsin from opsin+11-cis retinal (the vitamin A aldehyde) to opsin+ All-trans retinal. In Step 2, the activated form of rhodopsin, called metarhodopsin, binds the transducin, which is a G-protein composed of three units (α, β, γ). In Step 3 transducin converts GTP (guanosine triphosphate) into GDP (guanosine diphosphate) while its α subunit detaches from β and γ ones. Then, in the Step 4, the α subunit of transducin activates the enzymatic reaction to convert cGMP into GMP through its binding to the phosphodiesterase E (PDE), counteracting the activity of guanylate cyclase (GC), able to convert GTP in cGMP (step 5). The cGMP is a key regulator of ion channels activity; in fact, the cGMP closes Na⁺/Ca²⁺ ion channels, blocking positive charges flux and causing the hyperpolarization of photoreceptors. In daylight conditions, this cascade of events determines a hyperpolarization of photoreceptors and the consequent decreased release of glutamate by the synaptic terminal. On the contrary, in dark conditions photoreceptors are depolarized and the release of glutamate increases. The subsequent depolarization or hyperpolarization of bipolar cells (classified in ON and OFF) is related to glutamate receptors activated in the postsynaptic membrane (ionotropic/excitatory, metabotropic/inhibitory). The ON bipolar cells, having metabotropic receptors, are depolarized when the light reaches the centre of the retinal receptive field, driving an excitatory stimulus to the ON ganglion cells. Conversely, the OFF bipolar cells, which have ionotropic receptors, are hyperpolarized when the light reaches the centre of the retinal receptive field, determining the inhibition of the OFF ganglion cells. This phenomenon is peculiar of retinal electrical properties and represents the way through which the retina elaborates the light/dark stimuli and discriminates the response from the centre/periphery visual field. This phenomenon is peculiar of retinal electrical properties and represents the way through which the retina elaborates the light/dark stimuli and discriminates the response from the centre/periphery visual field. During the last step of the visual cycle (step 6), the arrestin converts the metarhodopsin to rhodopsin to restart a new cycle. RPE also plays a pivotal role in the visual cycle participating in converting all-trans retinal to 11-cis and making it available for a new visual cycle and in the discs renewal. Therefore, the dysfunction of RPE causes an unavailability of the retinal-11-cis and an accumulation of waste material, called lipofuscin, in the subretinal space.

Figure 2:

Schematic representation of the interaction between light and eye. The light is able to cross the external components of the eye: the cornea, the pupil and the lens. In the back of the eye the light is focused on the fovea. Parts of the figure were drawn using pictures from Servier Medical Art. Servier Medical Art by Servier is licensed under a Creative Commons Attribution 3.0 Unported License (https://creativecommons.org/licenses/by/3.0/).

Figure 3:

Schematic representation of the phototransduction cascade. Abbreviations: hv: photon; R: rhodopsin; G: trasducin; PDE: phosphodiesterase E; GC: guanylyl cyclase. The asterisk indicates the activated forms of rhodopsin (R*) and trasducin (G*). Parts of the figure were drawn using pictures from Servier Medical Art. Servier Medical Art by Servier is licensed under a Creative Commons Attribution 3.0 Unported License (https://creativecommons.org/licenses/by/3.0/).

2 Animal models of AMD induced by light

The complex aetiology of AMD is reflected in the lack of therapies and the effort put in place to find good animal models which faithfully reproduce the pathology. During the years, several rodent models were developed to study the pathogenesis and possible therapeutic approaches for AMD. We briefly summarize the main in vivo models of AMD and the features studied in each of them in Supplementary Table S1. Unfortunately, in most of these models, the degeneration proceeds slowly and takes many weeks to months to reproduce some of the hallmarks of the disease. Thus, the need for more eloquent in vivo models able to recapitulate the main aspects of AMD encouraged many researchers in the field to make progress. In this context, several papers reported light damage (LD) as a suitable and reliable model to reach this goal. In our laboratory, we established a rat model of AMD by exposing albino Sprague Dawley (SD) rats to white light for 24 h (h) (1000 lux) (Tisi et al. 2019b, 2020a,b, 2022, 2023). Noell et al. first reported the light damage paradigm in the literature (Noell et al. 1966). They exposed the animals to different wavelengths of light at different time points, demonstrating that high intensity light damaged the rat retina. After light exposure there was the presence of oedematous areas in the light-exposed retinas, together with morphological alterations in the RPE and the outer nuclear layer (ONL). These findings were further corroborated by electroretinogram recordings in which the electrical response of the retina was decreased after light exposure, compared to the control animals (Noell et al. 1966). This pioneering research gave important insights into retinal degeneration processes. Some years later, Marc et al. demonstrated that light-induced retinal damage (LIRD) in albino rats recapitulates many features of human atrophic AMD, more than other murine AMD models (Marc et al. 2008). Since then, many research groups in the field started to develop LD models by using different animal strains or light damage protocols. The effects of light on retinal toxicity could be influenced by many factors: pigmentation, time and intensity of light, acute and chronic exposure, source of light and wavelengths (Youssef et al. 2011). Also, melanin is an endogenous protective factor for the retina, and it is in the iris (the light or dark color of the iris depends on its concentration) and in the RPE. Sanyal and Zeilmaker in 1988 performed an experiment on chimeric mice, with a different proportion of albino and pigmented cells, and demonstrated that following exposure to constant light for five weeks in mice with more pigmented cells there was increased photoreceptor survival (Sanyal and Zeilmaker 1988). Thus, RPE pigmentation is considered a protective strategy of the retina against oxidative stress, and this is reflected in the race-related incidence of AMD (Bressler et al. 2008). Based on this, the first LD models were made on albino strains, because the induction of retinal damage by light is easier than in pigmented animals, since some strains are particularly resistant (Zhong et al. 2016); however, LD paradigms on pigmented animals were also developed (Krigel et al. 2016; Wooff et al. 2023). Another influencing factor is represented by the light sources and wavelengths used to induce retinal degeneration. Fluorescent lamps were the most used at the beginning, but their use was then replaced by light emitting diodes (LED) lights, which facilitated the optimization of parameters, including wavelength and irradiance time. Although the blue component of commercially available LED lights is hazardous for retinal tissue, LEDs are largely used during our daily life (Contín et al. 2016; Ouyang et al. 2020; Van Norren and Gorgels 2011). Therefore, lights with specific wavelengths in the blue spectrum have been recently used to induce retinal degeneration in rodents. We are constantly exposed to LED illumination and, since AMD is a chronic disease, one more discriminant for the development of LD models was the acute or chronic exposure to light. In Table 1 we summarized the main established LD paradigms in the literature, indicating the rodent strains, timing, spectrum, light intensity and acute or chronic exposure. In the next paragraphs we will deeply explore AMD hallmarks reproduced by LD models and molecular markers used for their study (Table 2), underlining benefits of their use to explore AMD pathogenesis and therapeutics.

Table 2:

Molecular hallmarks of LD models related to human AMD phenotype.

AMD features	Marker	Source	Modulation	References
Oxidative stress	Nrf-2	Retina		Yang et al. (2022)
	HO-1	Retina		Organisciak et al. (2013)
	SOD	Retina/blood		Qi et al. (2015) and Shin et al. (2022)
	NO	Blood		Wang et al. (2021)
	NOS	Blood		Wang et al. (2021)
	MDA	Retina/blood		Organisciak et al. (2013), Qi et al. (2015), Tanito et al. (2002), and Wang et al. (2021)
	CEP	Retina		Organisciak et al. (2013), Qi et al. (2015), and Tanito et al. (2002)
	GSH-Px	Blood		Wang et al. (2021)
	Trx	Retina		Tanito et al. (2002)
	8-OHG	Retina		Natoli et al. (2016)
	Acrolein	Retina		Tisi et al. (2019b)
Inflammation	IBA-1	Retina		Rutar et al. (2011b) and Tisi et al. (2019b)
	GFAP	Retina		Rutar et al. (2011b)
	IL-1β	Retina		Yang et al. (2022)
	IL-6	Retina		Wooff et al. (2023)
	IL-18	Retina		Yang et al. (2022)
	Ccl3	Retina		Rutar et al. (2015)
	Ccl4	Retina		Rutar et al. (2015)
	Ccl7	Retina		Rutar et al. (2015)
	Ccl2	Retina		Rutar et al. (2011b)
	CxCl1	Retina		Rutar et al. (2015)
	TNFα	Retina		Rutar et al. (2015)
	CxCl10	Retina		Rutar et al. (2015)
	NRLF3	Retina		Yang et al. (2022)
Lipids alterations	Omega-3	Retina/blood		Othman et al. (2015)
	Omega-6	Retina/blood		Othman et al. (2015)
	Endocannabinoid receptors	Retina		Maccarone et al. (2016)
	RvE1	Retina		Tisi et al. (2023)
		Blood
Blood retinal barrier alterations	(ZO-1)	Retina	Mislocation	Cachafeiro et al. (2013), Narimatsu et al. (2013), and Xie et al. (2021)
	Beta-catenin	Retina	Mislocation	Cachafeiro et al. (2013)
	N-cadherin	Retina	Mislocation	Cachafeiro et al. (2013)
	Serum albumin	Retina	Mislocation	Cachafeiro et al. (2013) and Krigel et al. (2016)
RPE alterations	RPE65	Retina		Song et al. (2020)
	ER stress markers	Retina		Jaadane et al. (2017) and Song et al. (2020)
	OPA-1	Retina		Wang et al. (2023)
	DRP1	Retina		Wang et al. (2023)
	LC3BII	Retina		Jaadane et al. (2017) and Tisi et al. (2020a)
	p62	Retina		Jaadane et al. (2017) and Tisi et al. (2020a)
	Apoptosis markers	Retina		Natoli et al. (2016), Tisi et al. (2019b), and Tisi et al. (2020a)
	RIPK3	Retina		Song et al. (2022)
Vascular alterations	VEGFA	Retina		Tisi et al. (2020b)
	VEGFR2	Retina		Tisi et al. (2020b)
	bFGF	Retina		Tisi et al. (2020b)
	FGFR1	Retina		Tisi et al. (2020b)
	PEDF	Retina		Cachafeiro et al. (2013)

3 Oxidative stress in human AMD

The retina is characterized by high oxygen consumption, because of its metabolic activity and constant exposure to light and UV radiation. Energy metabolism naturally generates oxidative stress through the production of ROS, which are chemicals containing oxygen (i.e. superoxide anion, hydroxyl radical, hydrogen peroxide and singlet oxygen) (Genestra 2007). ROS are first produced as a byproduct of mitochondrial metabolism, and during the phagocytosis of photoreceptors outer segments for their renewal (Brown et al. 2019; Datta et al. 2017). ROS serves as signalling molecules, regulating protein function and the activation of retinal antioxidant defences (Sies and Jones 2020). To counteract ROS accumulation, RPE synthesizes and releases antioxidant molecules, like glutathione, heme, antioxidant vitamins and redox proteins (i.e. thioredoxin) (part of Age-Related Eye Disease Study (AREDS) supplementation components). However, the retina eliminates ROS accumulation mainly through the nuclear factor erythroid-2-related factor 2 (NRF2)-related antioxidant enzymes, involving superoxide dismutase (SOD), catalase, and peroxidase, in addition to free radical scavengers (Lambros and Plafker 2016). Alongside antioxidants, autophagy also contributes to the redox defence of RPE (Ferguson and Green 2013). During aging, retinal antioxidant machinery is impaired or overwhelmed and neural retina and RPE are extensively exposed to oxidative damage (Tisi et al. 2021; Zhang et al. 2015). Furthermore, genetic polymorphism and lifestyle also exacerbate ROS accumulation (Othman et al. 2015; van Leeuwen et al. 2018).

4 Inflammation in human AMD

The term inflammation is usually related to a protective role from harmful stimuli, such as pathogens or injury, hence it is a cellular response against factors that perturbed the homeostasis of cells and tissues (Meizlish et al. 2021). The long-term inflammatory response is detrimental for the retina and is linked to AMD pathogenesis and progression (Ambati et al. 2013; Kauppinen et al. 2016; Nowak 2006). The breakdown of the blood–retinal barrier (BRB), as described in paragraph 7, causes the entry into the retinal space of inflammatory factors like inflammasomes, complement system, and immune cells. RPE and immune cells produce cytokines and chemokines leading to the inflammatory cascades (Wong et al. 2022). One of the first evidences that confirm the link between the inflammatory mechanism and the AMD were described by Green and Key (1977). Notably, local inflammatory events and abnormalities in inflammatory immune responses may contribute to drusen biogenesis (Hageman et al. 2001) and CNV (Kanda et al. 2008). In early AMD, the RPE activates the inflammasome (triggered by oxidative stress, see Section 3) and subsequently recruits macrophages to Bruch membrane, as reviewed elsewhere (Datta et al. 2017). To date it is well known that the acute inflammatory response is linked to dry AMD, rather than chronic inflammation (Tan et al. 2020) which is responsible for AMD progression to the late stages (Donoso et al. 2006; Kauppinen et al. 2016). Several studies investigated the role of chronic systemic inflammation in intermediate and neovascular AMD, discovering elevated plasma levels of cytokines, such as IL-4, IL-13 and IL-33 (Gotfredsen et al. 2023; Rajeswaren et al. 2023). Some of them (e.g. IL-4), were up-regulated in the plasma of GA patients, suggesting its potential role as a marker of dry AMD; on the other hand, the tumour necrosis factor-alpha (TNF-α) was recently correlated to CNV but not with dry AMD lesions (Khan et al. 2022). Thus, these markers might be considered for their use for a differential diagnosis of the two AMD forms. All the above-mentioned cytokines (IL-4, IL-13 and IL-33) are related to both types of macrophages, recruited to damaged tissues that polarize to functional phenotypes like M1 classified as pro-inflammatory and the M2 classified as anti-inflammatory (Zhou et al. 2017). Interestingly, both types of macrophages are present in AMD (dry and wet forms) human eyes as demonstrated for the first time by Cao et al. (2011). Furthermore, it should be mentioned the role of microglia cells (MG), the immunocompetent resident macrophages normally localized nearly to retinal blood vessels in the inner layers of the neural retina (Provis et al. 1995). Resting (or inactive) microglia, which display a ramified shape, lays in the inner layers of normal human retinas (Gupta et al. 2003); while the activated MG cells (balloon shaped, ameboid) start to invade the outer retina and were found in the photoreceptor layer and phagocytosed rod debris (Gupta et al. 2003). As discussed by the authors, the MG cells are activated in response to photoreceptors apoptosis, releasing cytotoxic factors (Gupta et al. 2003) and different inflammatory mediators such as: chemokines, ROS and NO, that all together provide a chronic neuroinflammatory environment (Cuenca et al. 2014). In addition, microglia infiltration contributes to the choroidal neovascularization in wet AMD. The role of the complement system is defending the human body from dangerous stimuli and participating in the tissue homeostasis. Under abnormal circumstances, it can lead to uncontrolled inflammatory response, resulting in disease phenotypes, as happens in AMD. The activation of the complement system is related to the proteolytic cascade triggered by different pathways (classic, lectin and alternative) resulting in the formation of several complement proteins, as reviewed elsewhere (Walport 2001). A multicenter study conducted by Heesterbeek et al. (2020a) on AMD patients at different stages and control subjects showed a correlation between the complement activation levels and the severity of AMD. Furthermore, mRNA levels of complement proteins were studied in patients with late AMD (Demirs et al. 2021). The use of RNA-scope followed by immunohistochemical analysis, showed that the microglia\macrophages cells produce C3 mRNA with the highest overlapping around the GA lesion.

5 Lipid alterations in human AMD

Lipids play multiple functions in the human body and are involved in the composition of cell membranes, serve as energy storage, and take part in the network of cellular signalling (these lipids are known as “bioactive lipids”). Accumulated evidence from human studies indicates that lipid metabolism is involved in the pathogenesis of AMD (van Leeuwen et al. 2018). The ratio of plasma omega-6/omega-3 was found to be increased in patients with AMD (Leung et al. 2019). It should be considered that the levels of omega-3 and -6 may also depend on their intake through food (for instance fish is rich in omega-3) and therefore data on their quantification should be carefully interpreted. Moreover, dietary intake of lipids may influence the development of AMD or prevent its progression. Indeed, the intake of omega-3 polyunsaturated fatty acids (PUFAs) (like alpha linolenic acid, ALA; eicosapentaenoic acid, EPA; docosahexaenoic acid, DHA) has been shown to prevent AMD development (Age-Related Eye Disease Study 2 Research Group, 2013. Lutein +zeaxanthin and omega-3 fatty acids for age-related macular degeneration: the Age-Related Eye Disease Study 2) (AREDS2) randomized clinical trial (Chew et al. 2013). Consistently, epidemiological studies indicate that patients suffering from AMD used to have lower fish intake (Augood et al. 2008; Christen et al. 2011). Moreover, DHA was found to be decreased in the serum of AMD patients (Orban et al. 2015). Omega-3 fatty acids are also the source of other bioactive lipids, known as “specialized pro-resolving lipid mediators” (SPMs), which are emerging as key regulators of the inflammatory processes by actively promoting the resolution of acute inflammation. SPMs in AMD have been poorly investigated, and there is a lack of data from humans. However, based on the protective effects exerted by omega-3, it is expected that their derivatives may help in the retinal protection of those patients. Differently, the intake of omega-6 fatty acids (like linoleic acid) is generally associated with an increased risk of AMD due their pro-inflammatory properties, although few studies demonstrated a significant correlation between those two (Christen et al. 2011; Seddon et al. 2001; Tan et al. 2009). Yet, the omega-6 arachidonic acid (AA) was found to be increased in the serum of AMD patients (Orban et al. 2015). Omega-6 are also the source of eicosanoids (EICs), a class of bioactive lipids (i.e. prostaglandins, PGs; prostacyclins, PGIs; thromboxanes, TXs; leukotriens, LTs), which exerts mainly a pro-inflammatory function (Leuti et al. 2020). Consistently, dysregulation of EICs has been shown in AMD patients, although the studies are still limited to few pieces of evidence. For instance, increased serum levels of PGE₂ have been identified in AMD patients (Orban et al. 2015). Alongside lipids, also lipoproteins, involved in lipid transportation through the circulation, have been associated with AMD (van Leeuwen et al. 2018). Specifically, epidemiological data showed that high levels of high-density lipoprotein cholesterol (HDL-C) in the blood of patients suffering from AMD correlate with an increased risk of the disease, while circulating oxidized low-density lipoproteins (ox-LDL) have not been associated with AMD development/worsening (Klein et al. 2019). Yet, increased levels of triglycerides seem to be negatively correlated with the risk of AMD (Colijn et al. 2019; Wang et al. 2016). Importantly, a large genome-wide association study has identified multiple genetic variants of genes involved in the lipid metabolism in AMD patients, including the ATP binding cassette subfamily A member 1 (ABCA1), APOE, cholesteryl ester transfer protein (CETP) and hepatic lipase C (LIPC) genes (Fritsche et al. 2016), further supporting the involvement of lipids in the pathogenesis of AMD.

6 Drusen and lipofuscin in human dry AMD

One of the main features of dry AMD is the presence of drusen, a yellowish compound rich lipids and proteins, detectable in correspondence of the macular region of the retina, between RPE and BrM membrane. In a first classification, drusen were divided in two types: soft indistinct drusen or reticular pseudodrusen, and soft distinct drusen associated with RPE abnormalities; this classification was proposed by several epidemiological studies including the Blue Mountain Eye Study, Beaver Dam Eye Study and Rotterdam Study (Klein et al. 1992; Mitchell et al. 1995; Vingerling et al. 1995). Subsequently, after the 2000s, a new grading system was developed by the Age-Related Eye Disease Study (AREDS) and Beckman Initiative for Macular Research Classification Committee. Drusen deposits were classified into three types depending on their size: small drusen (<63 μm), intermediate/medium drusen (63 ≦ μm, ≦125 μm) and large drusen (>125 μm) (Davis et al. 2005; Ferris et al. 2013). Drusen deposition increased the risk for progression to end-stages of GA and neovascularization (Curcio 2018). Schmitz et al. reported the prevalence of the reticular drusen analysing 458 subjects (418 patients had bilateral GA) thanks to the high-resolution imaging technology using a confocal scanning laser ophthalmoscope (cSLO) (Schmitz-Valckenberg et al. 2011). The prevalence of drusen detection was higher on cSLO images (62 %, at least in one eye) when compared with conventional fundus photography technique (18 %) and these results were similar to those described by Cohen and co-workers (Cohen et al. 2007), hence the reticular drusen can be considered a common phenotypic hallmark in eyes with GA due to AMD. The dimension of drusen has been used to provide a clinical estimation of the pathology progression; indeed, the presence of larger drusen indicates a higher risk for AMD progression to the advanced form. As the disease progresses, these drusen expand (Mitchell et al. 2018) and result in GA, characterized by irreversible damage to retinal cells (paragraph 9). The drusen composition has been well documented, identifying several drusen components (Wang et al. 2010); at least 40 % of drusen content is esterified cholesterol and phosphatidylcholine. Other components are vitronectin, TIMP metallopeptidase inhibitor 3, clusterin, Apolipoprotein E, complement factor H, complement factor 5, 6, 8 and 9. In this study was not confirmed the presence of others drusen components as Apolipoprotein B (Malek et al. 2003), amyloid oligomers (Luibl et al. 2006) complement component 3 (Johnson et al. 2001). Lipofuscin comprises abundant and long-lasting intracellular inclusion bodies, related to lysosomes, which contains mainly bisretinoids (Eidred 1993; Weirer et al. 1986), such as vitamin A derivates and N-retinylidene-N-retinylethanolamine (A2E) firstly described by Sakai and coworkers in 1996 (Sakai et al. 1996). Furthermore, the presence of the Lipofuscin within the RPE cells is responsible for the fundus autofluorescence (AF) in imaging (Delori et al. 1995). The Wisconsin Age-Related Maculopathy grading scheme (Klein et al. 1991), thanks to colour fundus photography, described in 1991 reticular soft drusen, then named reticular pseudodrusen (RPD), and known as subretinal drusenoid deposits (Sakurada et al. 2023). The RPD are composed of lipofuscin and extracellular deposits between the photoreceptor and the RPE and are associated with increasing risk to developed late stage of AMD, in particular GA (Finger et al. 2014; Keenan 2023). RPD are related to several retinal changes such as photoreceptors loss because of their localization in the subretinal space, which extended into the outer nuclear layer, resulting in a thinner retinal thickness in the affected area compared to the RPD-free region (Greferath et al. 2016). Similar results were obtained from a longitudinal perspective study (Chiang et al. 2020) in which was found a detectable retinal thinning in AMD patient, with a decrement about 5 % in the central macular region over four years. Furthermore, several choroidal changes are associated with RPD as reported such as: choroidal thickness (CT), choroidal vascular thickness (CVT) and choroidal vascularity index (CVI), highlighted through optical coherence tomography. In AMD choroidal vascular thickness is lost and CVI is significantly reduced (Laíns et al. 2017). Unlike drusen, to date RPD is incompletely characterized and needs further investigation to highlight a clear association with late AMD.

7 Blood-retinal barrier and its alterations in human AMD

Retinal microenvironment is finely controlled by the presence of the blood–retinal barrier (BRB), which regulates the transport of ions, water and proteins from the bloodstream to the retina and the translocation of waste material from the retina to the subretinal space and then to the systemic circulation. The BRB could be divided into inner BRB (iBRB) and oBRB and both contribute to the fine regulation of the retinal environment (O’Leary and Campbell 2021). The oBRB is the results of the interaction between the CC, the BrM and RPE. The structure of the choriocapillaris is characterized by extensive fenestrations and for this reason does not carry out a selective function but contributes to the BRB allowing the exchange of nutrients between the systemic circulations and the retina (Guymer et al. 2004). The RPE actively contributes to the formation of the BRB (and for this reason detailed in paragraph 8); indeed, the presence of Tight Junctions (TJ) at its apical surface ensures the maintenance of BRB integrity, allowing the controlled transport of substances to the neuroretina (Naylor et al. 2019). TJs are the regulators of the barrier functions and are composed by transmembrane proteins (such as the claudins, the MARVEL family and the JAMs family) and cytoplasmic proteins (including the members of the Zonula Occludens family) (Naylor et al. 2019). RPE also secretes numerous growth factors from its basolateral surface, such as VEGF, contributing to the homeostasis of CC (Caceres and Rodriguez-Boulan 2020). RPE cells also contribute to the synthesis of BrM components. In particular, the BrM is composed of five layers divided in collagenous layers and elastic layers. The main roles of the BrM in the functioning of BRB is allowing the passive diffusion of molecules through its layers and blocking the diffusion of big molecules (Booij et al. 2010). The other part of the BRB is the iBRB, consisting of the retinal vasculature, which originates from the retinal artery and supplies the retinal layers. The blood vessels that take part in the iBRB display a different structure compared to that of the CC, with no fenestration and a high presence of TJs which ensure molecules exchange mainly through a trans-cellular transport. Furthermore, retinal vessels are surrounded by pericytes and glial cells, which contribute to the formation of the so-called neovascular unit (NVU) (O’Leary and Campbell 2021). AMD is characterized by extensive structural and functional changes in both inner and outer BRB, with the consequent loss of its function. BRB alterations occurring during AMD could be different in presence of the dry or wet form. For instance, during the onset and progression of dry AMD, the choroid and choriocapillaris undergo a progressive thinning, which causes reduced nutrient’s supplementation (Tisi et al. 2021). On the other hand, during exudative AMD, the over expression of VEGF causes a dysregulated proliferation of the CC and the resultant invasion of the neuroretina, known as CNV (Cabral et al. 2017). More information about vascular changes occurring during wet AMD could be found in the dedicated paragraph (Section 9). The BrM, as a structure with a dynamic nature, is subjected to numerous changes during AMD pathogenesis. Notably, there is the accumulation of metabolism by products which cannot be eliminated and therefore contribute to drusen and basal deposits formation. Alterations in the extracellular matrix, including the imbalance between the secretion of metallopeptidases (MMP) and their inhibitors (TIMPs), also contribute to the formation of drusen and basal deposits in the BrM (Garcia-Garcia et al. 2022). Moreover, other structural changes involved the increased thickness of BrM and the reduction of the elastic properties. In terms of function, these changes imply a loss of the BrM barrier function with altered transport through the BRB (Booij et al. 2010). As mentioned in Section 8, RPE functions start to deteriorate with the accumulation of waste materials together with an auto-fluorescent pigment called lipofuscin (see paragraph 6). Late RPE contribution to the breakdown of BRB during AMD could be ascribed to the morphological and functional alterations, in a process known as epithelial-mesenchymal transition (EMT) (Shu et al. 2020). During EMT, retinal pigment epithelial cells lose their polarity and start to de-differentiate with the subsequent disruption of the junctional complex characteristic of this structure, and the uncontrolled exchange through the BRB (O’Leary and Campbell 2021). During AMD, multiple alterations also occur in the iBRB with differences between the dry and wet forms. In the past years, it was commonly accepted that in non-exudative AMD, the BRB integrity was not affected because there were no signs from the clinical analysis in patients with an intermediate form. Nevertheless, analysis of eyes obtained from AMD donors showed the presence of elevated iron levels and proteins of the systemic circulations in the eyecup, indicating a subclinical sign of BRB alterations in these patients. Conversely, during exudative AMD the signs of BRB breakdown are clearly and easily detectable with clinical imaging methods as fluorescein angiography (FA) (Schultz et al. 2019). Alterations of the retinal vasculature are deeply discussed in the dedicated paragraph (Section 9).

8 Retinal pigment epithelium in human AMD

RPE is a monolayer of post-mitotic, pigmented, polygonal and polarized cells placed behind the retina. RPE is one of the main components of the outer blood retinal barrier (oBRB), together with the choriocapillaris and the Bruch’s membrane, it is closely interconnected with the outer segments of photoreceptors, playing a key role in the proper functioning of the visual cycle. In addition, RPE ensures retinal homeostasis through its multiple functions; indeed, RPE cells corroborates photoreceptors functions through the absorption of scattered light, the shedding of photoreceptors outer segments, the renewal of photosensitive pigments and the secretion of growth factors and antioxidant molecules (Strauss 2005). RPE dysfunction and degeneration is considered one of the first events occurring in AMD (Kim et al. 2021) and, together with other outer retinal layers, is more subjected to light damage (Behar-Cohen et al. 2011). In particular, in RPE cells, light exposure induces an excessive production of ROS because of the high metabolic rate and oxygen consumption of this tissue; due to these features, under physiological conditions RPE is able to counteract ROS accumulation through its antioxidant machinery, which also involved the nuclear factor erythroid-derived factor 2/antioxidant response element (NRF-2/ARE) and peroxisome proliferator-activated receptor co-activator protein 1 (PGC-1) (Garcia-Garcia et al. 2022). Nevertheless, with aging the antioxidant system of RPE cells is less responsive, leading to ROS accumulation and oxidative stress burden. Moreover, oxidative stress and the accumulation of misfolded proteins, together with the impairment of RPE antioxidant machinery, lead to a “metabolic crisis” also associated with mitochondrial damage (Tong et al. 2022). Fusion and fission processes of mitochondria allow the maintenance of ROS levels into a physiological range, and an impaired balance between these two processes is the key to the establishment of the vicious cycle which leads to RPE/photoreceptors degeneration. This hypothesis was supported by numerous studies in which the analysis of mitochondrial function on AMD donor tissues showed fewer mitochondrial, reduced surface area and membrane potential, disrupted cristae together with altered mitochondrial proteome including the dysregulated expression of fusion/fission associated protein (Fisher et al. 2022). Furthermore, Ferrington et al. showed that RPE from AMD donors presents less ATP production and respiration, as evaluated by the Mito Stress assay, signs of a declined mitochondrial function (Ferrington et al. 2017). Oxidative stress could also compromise the correct function of the endoplasmic reticulum (ER), which is responsible for the protein synthesis and folding. During AMD, the dysregulation of these processes leads to the accumulation of misfolded proteins and waste materials (McLaughlin et al. 2022). RPE cells are also specialized in the renewal of photoreceptors’ outer segments by its digestion and recycling through the autophagy process. This process is also one of the key mechanisms involved in mitochondrial control, and it is responsible for the digestion of damaged mitochondria; in addition, this process also plays a fundamental role in the removal of misfolded proteins and waste materials accumulated following oxidative damage and ER stress. Recently, was assessed the role of impaired autophagy in the degeneration of RPE cells during AMD pathogenesis (Intartaglia et al. 2021). In the RPE of AMD donors, the content of Ubiquitin and sequestosome-1 (p62/SQSTM1) was up-regulated compared to healthy donors (Ferrington et al. 2016), suggesting an impairment of autophagy. Furthermore, there is evidence of an ER-mitochondria-autophagy axis in RPE cells, involved in AMD progression (Kaarniranta et al. 2023). The late stage of dry AMD is the GA (Zarbin et al. 2014) which is related to the occurrence of incomplete/complete RPE and outer retinal atrophy (iRORA/cRORA) (Garcia-Garcia et al. 2022; Savastano et al. 2020). Light induced animal models reproducing RPE dysfunction/alteration during AMD will be elucidated in the next paragraph.

9 Vascular alterations in human AMD

Retinal tissue needs a well-organized blood supply system; in particular, the outer retina is irrorated by the choroid, composed of different layers: the suprachoroidal, large and medium vessels layers and the choriocapillaris. The choroid is the structure responsible for nutrient exchange between the neural retina and the systemic circulation (Guymer et al. 2004). The retinal artery, one of the iBRB components (see paragraph 7), branches out into a deep, inner and superficial vascular plexuses placed in the outer plexiform layer, inner plexiform layer and in the ganglion cell layers respectively (Coorey et al. 2012). Alterations in the retinal supply system are mainly involved in the pathogenesis of wet AMD. As described in the paragraph 1.2, choroid could undergo different changes during AMD pathogenesis leading to the occurrence of CNV. It has widely been shown that isoform A of vascular endothelial growth factor (VEGFA) plays a key role in this pathology; it is a prominent angiogenic factor that has been the subject of intense research since its discovery. Other angiogenic factors include insulin like growth factor 1 (IGF-1), basic fibroblast growth factor (bFGF), TNF-α, angiopoietin-2 and MMPs (Paraoan et al. 2020). The stimulus for retinal neovascularization is often local hypoxia induced by oxidative stress, with blood vessel generation as an attempt to re-oxygenate the ocular tissue (Coorey et al. 2012). Retinal and choroidal hypoxia may induce an upregulation of VEGF production by RPE cells (Paraoan et al. 2020). Unfortunately, the newly formed blood vessels are leaky and prone to haemorrhage, allowing the infiltration of potentially detrimental factors. As mentioned in paragraph 7 (iBRB), retinal vessels are surrounded by pericytes and glial cells, which contribute to the correct functioning of the iBRB. During CNV, glial cells also participate in these pathogenic events. Indeed, CNV occurring in wet AMD is a response to RPE damage and the activation of an immune activation (see paragraph 4) (Ricci et al. 2020). Activated microglial cells are capable to express many pro-angiogenic factors, including VEGF, promoting angiogenesis and affecting retinal vasculature structure, thus promoting BRB breakdown (Coorey et al. 2012).

10 Diagnostic techniques: morphological and functional analysis for AMD

The diagnosis of AMD is mainly based on ophthalmoscopy (also known as fundoscopy). Colour fundus photography, optical coherence tomography (OCT), and fluorescein angiography are some of the clinical approaches used for the identification of morphological and structural changes in patients (Jaadane et al. 2023; Zarbin et al. 2014). These analyses are often related to a functional assessment of the retina through the electroretinogram apparatus (Forshaw et al. 2021; Robson et al. 2018) (see paragraph 11). OCT is a non-invasive, rapid technique used in order to obtain cross-sectional and 3-D images of the retina. Conventional OCT provides limited insights into retinal and choroidal changes during AMD pathology. Nevertheless, to obtain high resolution images of RPE, retinal and choroidal structures, new OCT-based approaches were developed (as deepened elsewhere (Zarbin et al. 2014)). The alignment of the cross-sectional OCT images with en face images and the high resolution (5-μm axial) of spectral-domain optical coherence tomography (SD-OCT) allows to determine the correlation between the histological features and clinical findings of human pathology, including AMD. Recently it has been proposed by Chen et al. (2023), the using of an ultrahigh resolution spectral domain-OCT (UHR SD-OCT) instrument to visualize and measure outer retina alterations at micrometer scale, as a marker for normal aging versus early AMD. FFA is the gold standard technique to investigate AMD-related vascular alterations, but it is an invasive procedure based on intravenous injections of sodium fluorescein dye. Thus, a less invasive OCT-based technique was developed: optical coherence tomography angiography (OCTA), providing a rapid and non-invasive tool for the diagnosis of retinal vascular pathologies (Jaadane et al. 2023). These clinical imaging tools enable the in-vivo evaluation of key AMD alterations, which could be otherwise detectable only ex-vivo. High-resolution spectral domain OCT is a tool applicable also to the light damaged rodent model to characterize the in-vivo morphological changes that will be compared to the functional analysis and the subsequent ex-vivo studies. Following this experimental approach Benthal et al. (2022), characterized the aetiology of decreased cone-driven vision in the light damage SD exposed to a chronic LD. They used the OCT for the in-vivo measurements of ONL thicknesses correlated with the assessments of visual function though the ERG recordings. Furthermore, FFA is currently used to assess vascular alterations in LD animal models (Gong et al. 2022). Morphological assessment, such as the measurement of the ONL thickness (or counting the photoreceptors layer) can be performed both in-vivo with the OCT, or ex-vivo through the analysis of the cryosections, as several research groups already performed (Rubin et al. 2022; Rutar et al. 2010; Tisi et al. 2019b). Furthermore, another important evidence obtained exclusively through the cryosections is the localization of selected markers through the visualizing of the entire retinal structure gained by immunohistochemistry technique, as Benthal et al. did in his study (Benthal et al. 2022).

11 Retinal function in human AMD

Electroretinogram (ERG) is a non-invasive technique performed to evaluate light-evoked electrical activity of retinal cells that can be performed in both preclinical and clinical studies. Thanks to the light-dependent activity of cones and rods and their connections with the inner retinal circuitry that the ERG response is generated. The cones mediate a photopic vision operating in a range of light intensities between 10⁻¹ and 10⁵ lux. The cones are highly concentrated in the center of the retina, especially in the fovea, they are characterized by high spatial and temporal frequency and permit to perceive the colours. Contrary, low resolution and high sensitivity are specific for rods, which guarantee a scotopic vision up to 10⁻² lux. The electroretinogram is mainly composed by three different waves such as: A-, B- and C-waves that represent the electrical response of photoreceptors (negative a-wave), second order neurons located in the inner retinal (positive b-wave) and the epithelial cells (late-onset positive c-wave) respectively (Robson et al. 2018). The International Society for Clinical Electrophysiology of Vision (ISCEV) determined a guideline on how to obtain, for each of the standard tests including the full-field flash electroretinogram (ffERG), the pattern electroretinogram (pattern ERG or PERG), the multifocal electroretinogram (mfERG), the electrooculogram (EOG) and the cortical-derived visual evoked potential (VEP), a clinical response with a standard protocol, thus enabling the comparison of ERG data set from different laboratories (Forshaw et al. 2021; Robson et al. 2018). The ISCEV standard full-field ERG represents a global retinal response to brief flashes of light, in scotopic or photopic conditions (Robson et al. 2018). Furthermore, in the ffERG the oscillatory potentials (OPs) can be also identified and they are characterized by being a high-frequency, low-amplitude response, which is superimposed on the rising phase of the b-wave (Liao et al. 2023); the OPs activity is related to the inner retinal circuitry that involve bipolar cells, amacrine cells, and/or ganglion cells (Liao et al. 2023; Wachtmeister 1998). Among many clinical applications the ERG is commonly used to diagnose and follow-up the AMD. The focal ERG (fERG) is a more specific test to specifically evaluate the electro-functional response of the macular region and it provides information exclusively from the fovea, by using a small flickering spot to elicit a local response (Hogg and Chakravarthy 2006). The fERG were recently used in combination with the visual acuity and OCT analysis from Savastano et al. (2022) for early diagnosis and prognosis of non-exudative AMD. Notably, the authors described with the longitudinal analysis that both fERG amplitudes and outer retinal thickness decrease after a one-year follow-up. Recently, Messenio et al. evaluated the macular function through the fERG in 47 patients with intermediate AMD compared to 65 healthy subjects, with preserved visual acuity (Messenio et al. 2022). Unsurprisingly was found a statistically significant difference in fERG mean amplitude between the two groups confirming the reduction in the macular function even if the visual acuity is preserved. Interestingly, by the fERG modulation sensitivity can be also evaluated for the cone function abnormalities through the temporal cone flicker sensitivity, as a function of flicker modulation depth as described by Falsini et al. (2000). The authors found that in AMD patients the fERG response relating to stimulus modulation depth, compared to the healthy controls, was altered displaying a different pattern of abnormalities correlated to the macular lesions. The early degenerative changes of cone photoreceptors, that their number should be still almost normal, were associated with gain losses and a modulation depth dependent phase delay, with normal thresholds. The authors speculated that the response gain losses and modulation depth dependent phase delays, with normal thresholds, were associated with early lesions, such as early degenerative changes of cone photoreceptors that the number should be still almost normal. Contrariwise, the more advanced stages were correlated to the increased thresholds, in addition to gain and phase abnormalities. Another important approach is the multifocal ERG (mfERG), described firstly by Sutter and Tran (1992), that allows topographically synchronized stimulation of multiple retinal regions, whose electrical response can be correlated to the function of specific retinal area. For this reason, to date mfERG is considered a useful tool for evaluating and monitoring the retinal function in AMD patients. Recently, Parisi et al. (2020), analysed 27 patients with intermediate AMD and 20 age-matched control eyes in order to investigate the macular function through the mfERG. The macular function in AMD patients was clearly impaired only in the fovea and parafovea area, not significant changes were found in more peripheral retinal areas. Similar results were obtained previously by Gin et al. (2011) that analysed the mfERG in 15 patients with intermediate AMD compared with 14 aged similar controls. At the same time another interesting study was published by Gonzàlez-Garcìa et al. (2016), where the authors combined the mfERG with an imaging technique, the optical coherence tomography (OCT) in 30 patients with dry AMD followed-up for two years. Unsurprisingly both techniques show clear changes over the time, such as increased drusen formation together with the functional impairment. The ffERG can therefore investigate the impact of peripheral retinal lesions leading to a better understanding of overall retinal function in AMD patients (Berrow et al. 2010) where only a few rods seem to remain within the parafoveal area in the late stages due to a preferential vulnerability and loss of rod over cone photoreceptors in both aging and AMD progression, as demonstrated by Curcio et al. (1993, 1996. In addition to the photoreceptors damaging, reactive oxygen species and oxidative damage contribute to modify the ERG response due to the inhibition of the sodium ion channel, thus resulting in the impairment of the rods-mediated phototransduction and prolonged dark adaptation. Regarding the wet AMD functional evaluation, Nishihara et al. (2008) assessed macular function in 157 patients by using focal macular ERG (fmERG). The amplitudes of fmERGs in the AMD patients displayed a reduction compared to the healthy controls. As well, the analysis of the a- and b-waves showed a clear reduction in the amplitudes of the corresponding waves in the healthy controls in addition to the implicit times becoming 3.8 ms (a-wave) and 10.2 ms (b-wave) longer compared to age-similar healthy subjects. This evidence suggests that the retinal function of these patients is severely impaired. The ERG technique has been performed also in animal models of AMD for preclinical studies as described in the next paragraph.

11.1 Electroretinogram in LD models

The functional consequences of the light exposure in rodents are generally assessed using the full-field (or flash) ERG. Many authors (Maccarone et al. 2008, 2016; Tisi et al. 2019a, 2020b) demonstrated the impairment of electrical activity of retinal cells caused by high light exposure. It is well described in our previous papers (Tisi et al. 2020b) that seven days after high light exposure (1000 lux for 24 h) of SD rats is evident a significant decrease of all parameters recorded by ERG, a-, b-waves and oscillatory potentials. The result is the visual function impairment with a similar residual retinal response from seven days up to 120 days, when compared to the unexposed animals as shown in Figure 4. Recently, Riccitelli et al. (2021) analysed the retinal function responses in both scotopic and photopic conditions with the ERG recording at different days of recovery such as 7, 15, 30, 45 and 90 up to seven months after LD (1000 lux for 24 h) in albino SD rats. Interestingly, the authors revealed a partial retinal recovery, maximal after 45 days (34 % for the a-wave, 45 % for the b-wave and 50 % for the OPs) compared to the control group, and after this time it dropped down up to seven months. This evidence was even corroborated with the flicker ERG analysis at frequency domain as the results showed a lower response, mainly after day 90 up to seven months, compared to the control group. In addition, the authors made a comparison between the ERG results performed seven days after different times of light exposure such as 12, 15, 18 or 24 h at 1000 lux. As discussed by the authors the short time of LD is not able to lead to a clearly retinal function deterioration as instead obtained with the 24 h of exposure time. Similar results were obtained by Cachafeiro et al. (2013) that exposed the BALB\c mice at 5000 lux for only 1 h. The authors performed the ERG in scotopic condition 10 days after the LD and both a- and b-waves were compromised due to the photooxidative damage. Despite the known resistance of C57BL/6J mice to light damage, Natoli et al. (2016), demonstrated a clear change of retinal function in response to seven days light damaged (100.000 lux up to seven days), the mice shown a significant reduction in both a- and b-wave (around less 25 % at highest flash intensity) comparable with those elicited in albino rat models. All the mentioned studies highlighted how the exposure to the light leading to a retinal function impairment because of all events triggered by the photooxidative stimulus, as fully explained throughout our review.

Figure 4:

Electroretinogram recordings analysis in a LD model of AMD. (A) Representative fERG recordings of CTRL (control), LD+7rec (LD 24 h+7 recovery days) and LD+60rec (LD 24 h+7 recovery days) experimental groups at 1 cd × s/m²; fERG recordings (B) a-wave amplitude and (C) b-wave amplitude; adapted with permission from (Fiorani et al. 2015; Tisi et al. 2020a,b).

12 Discussion

In vivo studies based on animal models are fundamental to investigate biomolecular, genetic and physiological changes responsible for human pathologies. However, a synergic work is necessary, between preclinical and clinical studies, to (i) identify new therapeutic targets, (ii) discover early pathology markers and (iii) test new pharmacological approaches. Having an animal model of pathology could be easier for genetic diseases or disorders with a well-known aetiology, but this is not true for multifactorial/neurodegenerative diseases, including AMD. Indeed, AMD does not have a defined pathogenesis, thus every affected patient might have a unique interplay between risk factors and genetic predisposition, which is not reproducible in animals. Furthermore, anatomical difference between humans and rodents must be also considered: rodents do not have the macula, the region responsible for visual acuity and where the degeneration occurs. Reproduce every single AMD feature in rodents is not possible, however, an ideal animal model should recapitulate the main histological and functional changes of AMD, being inexpensive and with a rapid evolution, as well.

In this review we mainly considered LD paradigms, but we also reported other existing AMD models (Table S1). Many of these are genetic based, even if their development is not an easy task, due to the complex aetiology of AMD. In some cases, AMD features were reproduced in these models through gene defects not involved in the human pathology; on the other hand, their use was fundamental to explore the pathogenic mechanisms related to one or more genes. A critic point is that, in genetic models, desired pathological hallmarks are often slowly developed, making difficult evaluate the disease progression.

LD models has allowed our and other research groups to deep understand novel mechanisms involved in AMD-induced retinal degeneration, using a well-known environmental risk factor for AMD. Light pollution has always been considered harmful for the retina; moreover, there was an increasing concern since LED lights were marketed for everyday use. In this review we summarized the main LD models used, which have been developed starting from different strains, using different light sources and timetable. We discussed the use of mice and rats as models for light-induced retinal degeneration. Although rodents display anatomical difference compared to human retina, rats develop retinal damage in a limited region (known as hotspot), partially resembling macular degeneration. Conversely, LD causes the degeneration in the entire retinal length in mice. Pigmentation also plays a pivotal role during AMD, acting as a ROS scavenger. This role is reflected in a biased incidence of AMD, which mainly affects people with a reduced melanin content. Pigmentation is protective also in rodents; indeed, reproduce AMD features in pigmented rodents is more difficult, requiring more intense light as well as chronic exposure of animals. In these cases, other factors should be considered such as circadian rhythm or excessive heat generated by light sources, which can negatively impact the experimental design. Furthermore, some pigmented strains as C57BL/6J mice are naturally resistant to light exposure due to genetic polymorphisms, not present in humans. Noteworthy, although pigmentation and genetic background are important in light damage paradigms, the age of animals and beginning of light damage are the two main critical factors (Polosa et al. 2016). In our previous studies we used an LD model based on the exposure of albino SD rats to white light (1000 lux, 24 h). We demonstrated that this model reproduces both the atrophic and the neovascular forms of AMD (Rutar et al. 2010; Tisi et al. 2020b, 2020c). Furthermore, the degeneration process is time-dependent allowing us to characterize different stages of the process. We demonstrated that immediately after light exposure there was a peak of apoptotic cells in the ONL, together with the appearance of subretinal debris in the subretinal space, which are early signs of AMD. Seven days after light damage, the “hotspot” area is clearly visible in a limited region of the rat’s retina, with an overall reduction of the ONL layer. At the same timepoint, electroretinogram recordings showed a significant decrease in both a- and b-wave amplitudes, consistent with visual function loss in AMD patients. From seven days after LD, neovascular lesions start to appear, with newly formed vessels invading photoreceptors layer and an up-regulation of pro-angiogenic factors (VEGF and VEGFR2). Notably, 120 days after LD the degeneration involved all superior retina, from dorsal edge to the area close to the optic nerve and affecting also electroretinogram response (Tisi et al. 2020b). Noteworthy, the inferior retina is preserved in our model as it happens in humans.

We can conclude that LD is an eloquent model to explore AMD, which helped our and other research groups to underline several novel mechanisms involved in AMD degeneration (i.e. autophagy, oxidative stress, metabolic lipids pathway and inflammation) (Table 2). We know the criticisms deriving from the use of LD models, but on the other hand these disadvantages lose their relevance, if compared with the advantage of being able to study all the characteristics of such a complex disease within a limited period. Furthermore, LD models actively contribute to the development of an effective strategy to slow down/treat both the early and the advanced forms of AMD. This will be a widespread degenerative disease and, if not properly treated, it will become highly disabling for millions of people worldwide. Expanding the knowledge on AMD is still crucial to discover new early biomarkers useful for a rapid diagnosis and for their use as therapeutic targets as well.