Autophagy in age-related macular degeneration

ABSTRACT

Age-related macular degeneration (AMD) is the leading cause of visual impairment in the aging population with limited understanding of its pathogenesis and a lack of effective treatment. The progression of AMD is initially characterized by atrophic alterations in the retinal pigment epithelium, as well as the formation of lysosomal lipofuscin and extracellular drusen deposits. Damage caused by chronic oxidative stress, protein aggregation and inflammatory processes may lead to geographic atrophy and/or choroidal neovascularization and fibrosis. The role of macroautophagy/autophagy in AMD pathology is steadily emerging. This review describes selective and secretory autophagy and their role in drusen biogenesis, senescence-associated secretory phenotype, inflammation and epithelial-mesenchymal transition in the pathogenesis of AMD.

Abbreviations: Aβ: amyloid-beta; AMBRA1: autophagy and beclin 1 regulator 1; AMD: age-related macular degeneration; ATF6: activating transcription factor 6; ATG: autophagy related; BACE1: beta-secretase 1; BHLHE40: basic helix-loop-helix family member e40; BNIP3: BCL2 interacting protein 3; BNIP3L/NIX: BCL2 interacting protein 3 like; C: complement; CALCOCO2/NDP52: calcium binding and coiled-coil domain 2; CARD: caspase recruitment domain; CDKN2A/p16: cyclin dependent kinase inhibitor 2A; CFB: complement factor B; DELEC1/Dec1; deleted in esophageal cancer 1; EIF2AK3/PERK: eukaryotic translation initiation factor 2 alpha kinase 3; EMT: epithelial-mesenchymal transition; ER: endoplasmic reticulum; ERN1/IRE1: endoplasmic reticulum to nucleus signaling 1; FUNDC1: FUN14 domain containing 1; GABARAP: GABA type A receptor-associated protein; HMGB1: high mobility group box 1; IL: interleukin; KEAP1: kelch like ECH associated protein 1; LAP: LC3-associated phagocytosis; LAMP2: lysosomal associated membrane protein 2; LIR: LC3-interacting region; MAP1LC3/LC3: microtubule associated protein 1 light chain 3; MTOR: mechanistic target of rapamycin kinase; NFE2L2: NFE2 like bZIP transcription factor 2; NLRP3; NLR family pyrin domain containing 3; NFKB/NFκB: nuclear factor kappa B; OPTN: optineurin; PARL: presenilin associated rhomboid like; PGAM5: PGAM family member 5, mitochondrial serine/threonine protein phosphatase; PINK1: PTEN induced kinase 1; POS: photoreceptor outer segment; PPARGC1A: PPARG coactivator 1 alpha; PRKN: parkin RBR E3 ubiquitin protein ligase; PYCARD/ASC: PYD and CARD domain containing; ROS: reactive oxygen species; RPE: retinal pigment epithelium; SA: secretory autophagy; SASP: senescence-associated secretory phenotype; SEC22B: SEC22 homolog B, vesicle trafficking protein; SNAP: synaptosome associated protein; SNARE: soluble N-ethylmaleimide-sensitive factor attachment protein receptor; SQSTM1/p62: sequestosome 1; STX: syntaxin; TGFB2: transforming growth factor beta 2; TRIM16: tripartite motif containing 16; TWIST: twist family bHLH transcription factor; Ub: ubiquitin; ULK: unc-51 like autophagy activating kinase; UPR: unfolded protein response; UPS: ubiquitin-proteasome system; V-ATPase: vacuolar-type H⁺-translocating ATPase; VIM: vimentin

Age-related macular degeneration (AMD)

Vision loss from age-related macular degeneration (AMD) is a major, expanding problem, especially in Western countries due to the aging population [1]. According to the World Health Organization, 1.3 billion people live with some form of vision impairment (https://www.who.int/blindness/en/), 188.5 million people have mild vision impairment, 217 million have moderate to severe vision impairment, and 36 million people are blind. In 2020, global projected cases of AMD were 196 million, rising to 288 million in 2040 [2]. In Western countries, one individual in seven over the age of 70 ‒ and one in three over the age of 80 ‒ is at risk of developing AMD; of these cases, over one-third will develop severe visual impairment or blindness in both eyes [3,4] The prevalence of AMD increased from 3.5% in people aged 55–59 years up to 17.6% in those 85 years and older in Caucasians [5].

The macula is responsible for our most important vision (sharpness, contrast and color vision) (Figure 1). AMD is usually classified into dry and wet forms. Early clinical hallmarks in both forms include the accumulation of lysosomal lipofuscin and extracellular drusen deposits and the retinal pigment epithelium (RPE) degeneration. Dry AMD (85–90% of cases) is currently untreatable. In wet AMD (10–15%), new blood vessels (neovascularization) sprout from the choriocapillaris into the retina and evoke detrimental edema [5,6]. Intravitreal injections of anti-VEGF (vascular endothelial growth factor) drugs (bevacizumab, ranibizumab, aflibercept, brolucizumab) have proven to be effective in the suppression of vascular growth and activity and decrease edema in the wet AMD [7]. Worldwide administration of anti-VEGF injections has been proven to slow the progression of wet AMD. The age-standardized prevalence of blindness due to AMD has decreased by almost 30% from 1990 to 2020 [8]. However, due to the increase in life expectancy and the concomitant demographic change in global human societies, AMD causes a tremendous impact on the physical and mental health of the geriatric population and their families. On top of that, the need of repetitive injections over a lifetime is a burden to both the patient and the health care system. The total worldwide cost of AMD treatment is estimated at 350 billion USD per year (www.brightfocus.org/sources-macular-degeneration-facts-figures).

Figure 1.

Color fundus photographs (upper panel) and optical coherent tomographs (lower panel) of (A) normal eye, (B) dry AMD with drusen (white arrow), and (C) wet AMD with hemorrhage, intraretinal, subretinal and sub-RPE fluids (red arrow). Abbreviations: AMD: age-related macular degeneration; PR: photoreceptors; RPE/BM: retinal pigment epithelium/Bruch’s membrane; Ch: choroid.

AMD etiology is multifactorial, with a combination of a strong genetic component and environmental risk factors, including aging, smoking, arteriosclerosis, obesity, hypertension, hypercholesterolemia and a fat-rich diet [9]. Disrupted lysosomal clearance and increased accumulated waste material are the major cellular factors contributing to AMD [10–12]. Two main mechanisms are responsible for managing the waste clearance in cells – autophagy and the ubiquitin-proteasome system (UPS). For a long time these systems were viewed independently from each other; however, the most recent progress in the field suggests that the clearance defects that lead to the progression of AMD should be considered in the context of autophagy-UPS interplay rather than in the context of any of these systems alone [13].

In addition to the well-documented involvement of autophagy in the dysregulation of cellular waste clearance leading to AMD [13], a few other considerations, such as oxidative stress, inflammation and RPE death, link defective autophagy with AMD [11,14–16]. Reactive oxygen species (ROS) have been described to be involved in nutrient-starvation-induced autophagy as signaling molecules [17]. In turn, the cellular antioxidant response is regulated by autophagy through the SQSTM1/p62 (sequestosome 1)-KEAP1-NFE2L2/Nrf2 pathway [18–20]. Importantly, SQSTM1 plays a special role bridging the autophagy-UPS systems [21]. Autophagy has also been reported to associate with antioxidant defense through the DNA damage response mechanisms [22].

Retinal pigment epithelium in AMD

RPE degeneration is one of the hallmarks of AMD [23]. In addition to apoptosis, necroptosis seems to be an important RPE cell death type in AMD [24–26]. Autophagy is involved in the regulation of all cell death routes that have been described in AMD pathogenesis [25].

The quiescent RPE cells form a single cell layer located between the photoreceptors (rod and cones) at the front and Bruch’s membrane and choriocapillaris (Figure 1). One of the main functions of the RPE cells is to maintain the homeostasis of the overlying photoreceptors [27]; the RPE transports nutrients and oxygen from the choriocapillaris and regulates the visual cycle according to circadian rhythm. The RPE produces and secretes several growth factors including VEGF, therefore playing a key role in the development of wet AMD. RPE cells have evolutionarily developed a high concentration of many cytoprotective pigments, such as melanin, flavins and retinoids likely due to constant light exposure [28,29]. These compounds are especially abundant in the macula and are seen as a yellowish pigmentation and thus called macula lutea. Although, RPE has been accepted to be a driver of early AMD, pathobiology compositional and cell-type-specific gene expression changes in different macular cells can be observed during AMD development [9,30].

One of the central homeostatic functions of RPE cells is phagocytosing lipid-rich end-products of the visual cycle coming from photoreceptor outer segments. This process can be interpreted as MAP1LC3/LC3 (microtubule associated protein 1 light chain 3)-associated phagocytosis (LAP; Figure 2) [31,32]. The decreased LAP and an excess of mitochondria-derived ROS have been suggested to be key factors in triggering RPE degeneration in AMD [9,33]. Chronic oxidative stress makes proteins vulnerable to damage and leads to the detrimental accumulation of aberrant proteins. This is especially important for normal homeostasis in cells that are not dividing as part of normal homeostasis such as RPE cells. A clear clinical sign of the disturbed proteostasis is the accumulation of lysosomal lipofuscin in the RPE and extracellular drusen deposits between the RPE and choriocapillaris (Figures 1 and 2). Proteasomes and lysosomes play a main role in the prevention of protein aggregation and cellular clearance [34]. Currently, emerging evidence suggests that autophagy plays an important role in the various stages of AMD development and progression [6,10,11,33–35]. Because lysosomal enzyme activity is crucial to autophagy flux, lipofuscin components may ultimately decrease the capacity for autophagy clearance [36]. This is supported by reports showing impaired autophagy that coincides with lipofuscin granules and drusen deposits in RPE cells of the nfe2l2/nrf2 and ppargc1a/pgc-1α knockout mouse and human cadaver AMD samples [11,37]. In addition, a decline in autophagic capacity accompanied by an increase in ROS production can elicit activation of inflammation in RPE cells [38].

Figure 2.

Proteostasis in a degenerative RPE (retinal pigment epithelial) cell. Macula POS (photoreceptor outer segments) are exposed to a high load of light-induced reactive oxidative species. To prevent accumulation of agents the daily phagocytosis of POSs called heterophagy or LC3 (microtubule associated protein 1 light chain 3 alpha)-associated phagocytosis. In aged RPE cells, degradation of the discs decreases, and lipofuscin start to accumulate in lysosomes because of the coincident decline of lysosomal enzyme activity. The impaired lysosomal enzyme activity inhibits autophagic flux and increases oxidative stress that lead to protein misfolding, cellular organelle damages and protein aggregation. Individual polypeptides can be degraded by the ubiquitin (Ub)-targeted proteasome, while aggregates are degraded by selective autophagy. Autophagosome membranes are thought to originate in part from the endoplasmic reticulum (ER). SQSTM1/p62 (sequestosome 1) sorts proteins between proteasomal and autophagic clearance pathways; this receptor binds to Ub cargos and to LC3.

Selective autophagy in AMD

The critical role of autophagic clearance mechanism in AMD pathogenesis is emerging [9]. Selective autophagy signaling has been thoroughly reviewed elsewhere [39]. Briefly, MTOR (mechanistic target of rapamycin kinase), ATG (autophagy related) genes, ULK1 (unc-51 like autophagy activating kinase 1), LC3 and SQSTM1 (as well as other receptor proteins) are key molecules for regulating selective autophagy. Autophagy is initiated by the formation of a transient sequestering compartment termed the phagophore, that expands and closes to become a double-membrane vesicle, an autophagosome, enclosing material to be degraded (cargo) that is delivered to the lysosome, where degradation and recycling occur. MTOR is the primary negative regulator that controls the initiation signals of autophagy. Active ULK1 participates in autophagosome initiation to create the phagophore. ATG7 activates Atg8-family proteins (including LC3 and GABARAP subfamilies; we will refer to LC3 hereafter for simplicity) and ATG12; the latter forms a complex with ATG5 and ATG16L1, which functions in part as an E3 ligase that conjugates the LC3 to phosphatidylethanolamine on the phagophore, allowing membrane expansion and autophagosome maturation. The lipid conjugated form of LC3 is referred to as LC3-II. The conjugated LC3 also play an important role in cargo recognition; LC3 has ubiquitin and SQSTM1 binding sites that connect cargo to the phagophore, as well as connecting the process of autophagy to the proteasomal clearance system. Once the specific cargo is ubiquitinated and recognized by SQSTM1, which then binds LC3, it undergoes autolysosomal degradation (Figure 2).

Autophagy impairment, caused by the depletion of the core autophagy genes ATG5 and ATG7, is associated with an AMD-like phenotype in mouse RPE cells [40]. This phenotype is manifested by RPE thickening, hypertrophy or hypotrophy, pigmentary abnormalities and the accumulation of oxidized proteins. Oxidative stress is a canonical stimulus to induce autophagy in RPE cells [11,35,37,41,42]. Autophagy flux in late AMD occurs at a lower rate than in early stages of the disease [11,38]. Chronic oxidative stress seems to decrease lysosomal enzyme activity and autophagic flux in the RPE cells [11,36,37,43].

Because the RPE cells have high rate of phagocytosis, they require efficient autophagy process [6]. It is likely that the normal functioning of lysosomes, organelles which must fuse with autophagosomes to deliver the hydrolases that degrade the cargo, is indispensable in the degradation and recycling processes. The lysosomal degradation pathway declines with age in the human brain, contributing to the pathogenesis of neurodegenerative diseases [44,45] and is now considered as a significant risk factor for AMD [9]. The acidification of lysosomes is established by the vacuolar-type H⁺-translocating ATPase (V-ATPase) [46,47] which are multi-subunit complexes, composed of a peripheral V₁ domain that hydrolyzes ATP and an integral V₀ domain, that translocate protons from the cytoplasm to the lumen [47]. Interestingly, CRYBA1/βA3/A1-crystallin binds to the V₀ domain of V-ATPase and regulates lysosome-mediated degradation in the RPE [12]. Furthermore, CRYBA1/βA3/A1-crystallin is also a master regulator for amino acid sensing in lysosomes and for translocation of amino acids [48] into the lumen, a requirement for MTORC1 activation [49]. These findings suggest that CRYBA1/βA3/A1-crystallin is essential for MTORC1 signaling in the lysosomes of RPE. The MTORC1 signaling pathway gene defects are linked to wet AMD [50].

Most of the autophagy studies for AMD pathology have been done in cell culture and animal models or cadaver tissue samples. Care should be taken in interpreting autophagy protein profiles in animal models since there is significant diurnal rhythmicity in ATG7, ATG9, LC3 and BECN1 (beclin 1) expression in wild-type mouse and rat retina [51]. Furthermore, the distinctive diurnal rhythmicity of these autophagy proteins was significantly impaired, and phase shifted in diabetic animals, but it is not clear if a similar diurnal dysregulation is observed in AMD.

Mitophagy in AMD

The precise cause of RPE degeneration and the onset and progression of AMD are not fully understood. However, mitochondria dysfunction, increased ROS production, and mitochondrial DNA/mtDNA damage are observed together with increased protein aggregation and inflammation in AMD [9]. To avoid the vicious cycle of extensive ROS production with further mitochondrial damage and ROS release, dysfunctional mitochondria must be removed from cells via mitophagy, a specialized form of selective macroautophagy, where the main cargo are mitochondria (Figure 3) [43,52].

Figure 3.

Age-related macular degeneration is associated with damaged and dysfunctional mitochondria in retinal pigment epithelial cells that are a source of reactive oxygen species that may further damage the organelle. To avoid a vicious cycle, damaged mitochondria are removed by mitophagy, a pathway of selective autophagy with mitochondria as the main cargo. Damage to mitochondria is associated with a decrease in mitochondrial membrane potential (Δψ_m). In PINK1 (PTEN induced kinase 1)-PRKN (parkin RBR E3 ubiquitin protein ligase)-dependent mitophagy, PINK1 detects Δψ_m and damaged mitochondria and recruits, phosphorylates (P) and activates PRKN, which then amplifies the signal and ubiquitinates (Ub) various mitochondrial surface proteins represented here by a single OMM (outer mitochondrial membrane) protein. These ubiquitinated proteins are subsequently recognized by autophagic receptors, including AMBRA1 (autophagy and beclin 1 regulator 1), OPTN (optineurin), CALCOCO2/NDP52 (calcium binding and coiled-coil domain 2) and SQSTM1 (sequestosome 1) that target mitochondria to phagophores for degradation via direct interaction with MAP1LC3/LC3 (microtubule associated protein 1 light chain 3) protein. TBK1 (TANK binding kinase 1) phosphorylates autophagic receptors to increase mitophagy. PINK1-PRKN-independent mitophagy is mediated by various mitophagy receptor proteins, including BNIP3L/NIX (BCL2 interacting protein 3 like), FUNDC1 (FUN14 domain containing 1), and BNIP3 (BCL2 interacting protein 3) that on phosphorylation interact with LC3 independent of Ub.

PINK1 (PTEN induced kinase 1), PRKN (parkin RBR E3 ubiquitin protein ligase) and OPTN (optineurin) are some of the main proteins of the mitophagy pathway [53]. Once mitochondrial damage occurs, the inner mitochondrial membrane is depolarized, PINK1 import into the mitochondria and its degradation by the protease PARL (presenilin associated rhomboid like) are suppressed, leading to its increased accumulation in the outer mitochondrial membrane (Figure 3) [54,55]. Subsequently, PRKN is recruited from the cytosol. PINK1 phosphorylates ubiquitin and the ubiquitin-like domain of PRKN at serine 65 [56,57]. Once recruited and activated, PRKN ubiquitinates numerous targets and stimulates the local formation of phagophores via receptor interactions with LC3 via their LC3-interacting region (LIR) motifs (Figure 3) [58]. Receptor-mediated mitophagy regulated with AMBRA1 (autophagy and beclin 1 regulator 1), BNIP3 (BCL2 interacting protein 3), BNIP3L/NIX (BCL2 interacting protein 3 like) and FUNDC1 (FUN14 domain containing 1) coexist with PINK1-PRKN-mediated mitophagy and/or other types of mitophagy [59]. Oxidative stress significantly increased key molecules involved in mitophagy (PINK1 and PRKN) in RPE cells to segregate damaged mitochondria and remove them via autophagy. Interestingly, knockdown of BACE1 (beta-secretase 1), a key enzyme in amyloid formation in Alzheimer disease that is also expressed in the retina, results in significantly increased levels of PINK1 and PRKN in oxidatively stressed RPE cells and is associated with a large increase of mitochondria within autophagosomes [60]. Alterations in the mitophagic pathway by oxidative stress are similarly observed in other degenerative diseases involving BACE1, such as Alzheimer disease and Parkinson disease [61].

A significant amount of macroautophagy, including mitophagy, takes place in the outer retina [59]. Interestingly, mitochondrial redox states seem to affect retinal mitophagy stronger than do normal circadian rhythm [62,63]. This is supported with the dry AMD resembling nfe2l2 ppargc1a double-knockout mice that show an upregulation of PINK1 and PRKN together with damaged mitochondria [43]. Because neither evidence of increased autolysosome formation in transmission electron micrographs nor colocalization of lysosomal marker LAMP2 (lysosomal associated membrane protein 2) and the mitochondrial marker ATP synthase in confocal micrographs are observed, the authors suggested that there is a decrease of mitophagy in response to disturbed antioxidant defense and dysfunction of energy metabolism. Although, upregulation of the receptor-mediated mitophagy markers were not detected in the nfe2l2 ppargc1a double-knockout mice, their role in AMD pathology should be considered [43,64].

Endoplasmic reticulum-selective autophagy in AMD

Chronic oxidative stress, impaired autophagy, increased protein aggregation and RPE cell degeneration coincide with endoplasmic reticulum (ER) stress in AMD pathology [37,65,66]. To maintain proteostasis and cell function, the ER activates an adaptive quality control mechanism known as the unfolded protein response/UPR. The unfolded protein response is initiated by three independent transmembrane stress transducers; 1) ERN1/IRE1 (endoplasmic reticulum to nucleus signaling 1), 2) EIF2AK3/PERK (eukaryotic translation initiation factor 2 alpha kinase 3), and 3) ATF6 (activating transcription factor 6) [67]. Numerous proteins moving to the ER, Golgi apparatus, cell membrane, or lysosomes are synthesized on the ER membrane. Those proteins undergo various post-translational modifications and chaperone stabilization to obtain the final functional structure. Proteins having ER-leader sequence in their amino-terminal tail are secreted canonically outside of the cell via the Golgi apparatus [68]. Unconventional secretion occurs for leaderless sequence proteins directly from the cytoplasm. Both conventional and unconventional secretory pathways are interfaced with autophagy. Several ATG proteins have been implicated in the process of unconventional secretion [68].

There is evidence of autophagy-ER-mitochondrial crosstalk in RPE cells and during AMD development (Figures 2 and 3) [37,43,66,69]. ER-mitochondria contact sites are involved in autophagosome formation, and many proteins in the mitochondria-associated ER membrane/MAM compartments are necessary for autophagic vesicle formation (Figure 2) [70]. Autophagy pathways also regulate cytokine production and secretion, particularly IL1B/IL-1β (interleukin 1 beta) [71]. Vice versa, cytokines, including IL1B, induce autophagy and activate autophagosome formation [71]. Therefore, it seems that autophagy is closely involved in the non-canonical secretion of leaderless proteins ranging from those involved in the control of mitochondrial health to inflammation signaling. One can also anticipate that they have a role in drusen formation, a key clinical sign in AMD.

Decreased autophagy and senescence in AMD

Decreased autophagy is associated with a cell phenotype shifting to senescent cells that contribute to a loss of tissue homeostasis [37,72]. Accumulation of lipofuscin in the RPE is a sign of senescence in AMD. Senescence-associated secretory phenotype (SASP) is associated with the release of ROS, selective growth factors, and inflammatory cytokines, chemokines and proteases (Figure 4) [72]. HMGB1 (high mobility group box 1) is a senescence marker that is upregulated in iPSC-RPE in response to lipofuscin component treatment [73]. Mitochondrial damage or ER stress also upregulate HMGB1 [70]. A recent observation involving mitochondrial phosphatase PGAM5 confirms the role of mitochondria in the modulation of cellular senescence by regulating mitochondrial dynamics [74]. AMD-associated changes in the cells, such as telomere dysfunction, DNA damage, and metabolic disturbance may lead to the persistent accumulation of senescent cells [68,75]. Because senescent cells display disturbed metabolic function and autophagy activity that coincide with proinflammatory secretome, they may regulate drusen biogenesis and advanced AMD development for certain patients [75]. Cellular senescence has context-dependent beneficial or detrimental role, but under chronic oxidative stress and during RPE degeneration, senescence may progress AMD [76]. Therefore, removing of senescent RPE cells or suppression of their activity is a potential therapeutic option in AMD [77,78]. For example, an CRYBA1/αB crystallin chaperone peptide functions as a senolytic agent (inducing death in senescent cells) in experimental AMD [78]. Interestingly, cocultures of embryonic stem cells reversed senescence of RPE cells [79]. Our recent observations reveal that senescence pathways may also intersect with epithelial mesenchymal transition phenotype in the retinas of nfe2l2 ppargc1a double-knockout mice [68].

Figure 4.

Interaction of cytoplasmic waste clearing and secretion systems in a degenerative RPE cells. Constant oxidative stress and disturbed proteostasis may lead to the activation of secretory autophagy and SASP (senescence-associated secretory phenotype) that may regulate drusen biogenesis and progression of age-related macular degeneration via increased inflammation and EMT (epithelial-mesenchymal transition). Secretory autophagy cargo IL1B/IL-1β (interleukin 1 beta) is recognized by TRIM16 (tripartite motif containing 16) that interacts with the soluble N-ethylmaleimide-sensitive factor attachment protein receptor (SNARE) that promotes LC3 (microtubule associated protein 1 light chain 3 alpha)-linked cargo degradation. SEC22B (SEC22 homolog B, vesicle trafficking protein), STX3 (syntaxin 3), STX4, SNAP23 (synaptosome associated protein 23), and SNAP29 complete cargo secretion. IL18, Aβ and HMGB1 (high mobility group box 1) are key proteins of secretory autophagy. In SASP, CDKN2A/p16 (cyclin dependent kinase inhibitor 2A), DELEC1/Dec1 (deleted in esophageal cancer 1)-BHLHE40 (basic helix-loop-helix family member e40) and HMGB1 are secreted in the age-related macular degeneration degenerative process.

Autophagy in the prevention of epithelial-mesenchymal transition

Epithelial-mesenchymal transition (EMT) is a process in which polarized epithelial cells change their structural and biochemical properties into more motile and adaptable mesenchymal cells [80]. Autophagy prevents EMT by clearing SNAI/snail and TWIST, EMT-inducing transcription factors. For example, in differentiated hepatocytes, autophagy degrades SNAI, whereas lysosomal inhibition impairs the autophagic clearance of SNAI and TWIST to induce EMT. It is likely that RPE cells escape from the stressful microenvironment produced by oxidative insult via initiation of type 2 EMT, a mechanism that the cells adopt to avoid cell death [68]. However, in AMD patients, the cells die due to the chronic nature of the insult as the disease progresses [25]. The RPE may switch on EMT due to malfunctioning of various pathways/processes that are required to maintain homeostasis of the RPE cells [68]. One such process that is immensely important for RPE cell survival is lysosomal-mediated clearance by phagocytosis and autophagy [6].

Type 2 EMT occurs in organ fibrosis and is mediated by inflammatory cells and fibroblasts that release different inflammatory mediators and components of the extracellular matrix (Figure 4) [68,80]. Once RPE cells are degenerated, they lose cell-to-cell contact and cell polarity, causing them to migrate into the retinal and subretinal RPE space [81]. Profibrotic cells migrate to and proliferate at the Bruch’s membrane-RPE complex, where they form subretinal fibrosis with the involvement of other cells, mainly matrix-producing mesenchymal cells [82]. Progression of AMD to its advanced stage, resulting from oxidative stress, RPE degeneration and low-level chronic inflammation, displays some common features with aberrant wound healing [83]. EMT-related changes have been observed in AMD-simulating cell lines and tissue samples [84], but also in serum isolated from wet AMD patients [85]. In nfe2l2ppargc1a double-knockout mice, a model of atrophic AMD, increased mesenchymal markers (SNAI1, SNAI2/Slug, VIM [vimentin] and CDH11/OB-cadherin) are associated with senescence markers (Figure 4; CDKN2A/p16, DELEC1/Dec1-BHLHE40 and HMGB1), which together might drive EMT in aged RPE cells [68]. EMT may be initiated in RPE cells as they try to increase their plasticity to better cope with the detrimental changes in their environment. As EMT strongly disrupts the normal RPE cell function, it eventually leads to the death of photoreceptor cells and loss of vision. Autophagy is a putative therapy target to prevent TGFB2/TGF-β2 (transforming growth factor beta 2)-induced EMT in human retinal pigment epithelium cells [86].

Crosstalk between secretory autophagy and inflammation in AMD

Because autophagy plays an important role in the homeostasis of RPE cells, its impairment can lead to accumulation of damaged organelles and various nonfunctional or toxic proteins, including lysosomal lipofuscin, and promote the formation of extracellular drusen [33]. Currently, the role of different autophagic mechanisms in retinal physiology and pathology leading to AMD is not completely clear. It is well established that lysosomal degradative functions can be inhibited in RPE cells by a lipofuscin component [87]. In addition to lysosomal degradation, autophagy may play a role in non-canonical secretion of leaderless cytosolic proteins, a process called secretory autophagy (SA). In response to lysosomal damage, the prototypical cytosolic SA cargo IL1B is recognized by the specialized SA cargo receptor TRIM16 (tripartite motif containing 16) that interacts with the soluble N-ethylmaleimide-sensitive factor attachment protein receptor (SNARE) system to deliver the cargo to LC3-II (Figure 4) [88]. Cargo secretion is unaffected by downregulation of STX17 (syntaxin 17), a SNARE promoting autophagosome-lysosome fusion and cargo degradation protein. SEC22B, in combination with plasma membrane STX3 and STX4, as well as SNAP23 (synaptosome associated protein 23) and SNAP29, completes the cargo secretion. Thus, SA utilizes a specialized cytosolic cargo receptor and a dedicated SNARE system. Other cargo, such as ferritin, is secreted via the same pathway [88]. Moreover, a subset of leaderless cytosolic proteins, including IL18, IL33 and amyloid-beta (Aβ) can be secreted through SA.

The AMD genetic risk variants in cluster genes of the complement system and complement activation products are reported in AMD patients [89]. IL1B promotes complement alternative pathway activation in RPE cells [90]. Aβ upregulates CFB (complement factor B) in RPE cells through cytokines released from recruited macrophages/microglia (Figures 4 and 5) [34]. Interestingly, iron chelator protects against light-induced retinal degeneration that is associated with C3 (complement C3) downregulation [91].

Figure 5.

Crosstalk of the complement system and inflammation in degenerative retinal pigment epithelium cells in oxidative stress and dysfunctional autophagy symbolized by a multicolor thunder within an autophagosome. Oxidative stress may impair autophagy, and impaired autophagy may contribute to oxidative stress. NLRP3 (NLR family pyrin domain containing 3) interacts with the adaptor protein PYCARD/ASC (PYD and CARD domain containing), which then recruits pro-CASP1 (caspase 1). Activation and assembly of the inflammasome promotes proteolytic cleavage with CASP1, maturation and secretion of pro-inflammatory cytokines IL1B (interleukin 1 beta) and IL18 (interleukin 18). NFKB/NFκB (nuclear factor kappa B) is a key protein complex to prime the inflammasome response. Together with many cytokines, complement proteins C5/C5a, C3 and CFB (complement factor B) are involved in inflammasome regulation. Chronic inflammation may lead to the epithelial-mesenchymal transition (EMT) that may play a role in age-related macular degeneration.

In addition to complement system-regulated inflammation with oxidative stress and impaired autophagy, the activation of NLRP3 (NLR family pyrin domain containing 3) inflammasome occurs in AMD progression (Figure 5) [92]. The activation and assembly of the inflammasome promotes proteolytic cleavage, maturation and secretion of pro-inflammatory cytokines IL1B and IL18 [93]. The autophagy receptor SQSTM1 is a significant factor in inflammatory responses, especially following inflammasome activation [93]. Pro-IL18 is constitutively expressed by ARPE-19 cells, whereas the expression of pro-IL1B is induced by IL1A/IL-1α priming. The release of mature IL1B and IL18 is observed together with increased ROS and DNA damage. NLRP3 inflammasome signaling has a key role in the processing of IL1B [94]. IL33 is induced by Aβ stimulation that coincides with inflammatory cytokines production in RPE cells [95]. Interestingly, drusen have a wide spectrum of amyloid structures that may lead to local toxicity in the RPE and induce local inflammatory events in the pathogenesis of AMD [96]. It is suggested that functional autophagy can limit activation of the inflammasome and other inflammatory responses [97,98]. The complement system presents inflammasomes with C5/complement component C5a, which primes RPE cells for inflammasome activation by lipofuscin-mediated photooxidative damage [99,100]. Supporting a role for secretory autophagy in AMD, Aβ has been reported to regulate inflammasome activation [101,102]. Secretory autophagy machinery and vesicular trafficking are also involved in HMGB1 secretion and a senescence phenotype in AMD and an AMD animal model [68,103].

Future directions

Recent studies support the involvement of increased oxidative stress, protein aggregation, cellular senescence, chronic inflammation and, for some patients, the development of choroidal neovascularization and EMT phenotype in AMD pathogenesis. Because autophagy has a central role in the regulation of all these cellular processes, we should better understand the different autophagic mechanisms in retinal physiology and pathology underlying AMD. Secretory autophagy that overlaps with inflammation may open new ways to understand AMD pathology and project therapy targets. Current clinical focus is shifting to dry AMD therapy development, once wet AMD anti-VEGF treatments have been improved [7,104]. Although, autophagy has a key role in AMD pathology MTOR inhibition in clinical trials show no evidence of efficacy in advanced AMD [105,106]. One can speculate that AMD patient were selected too late to the trials. Recent observations provide promising results to prevent AMD-related cellular signs by activating autophagy without direct MTOR inhibition [104]. In addition to pharmacological developments miRNAs that regulate autophagy has been documented as putative future treatment options for AMD [107,108].

Genetic and environmental risk factor studies do not sufficiently explain progression differences between dry and wet AMD phenotypes or treatment responses. Thus, there is a need to apply precision medicine and understand the genetic, epigenetic, metabolomic and proteomic etiology of autophagy regulators to prevent or treat AMD.

Funding Statement

This work was supported by the Academy of Finland [333302];Narodowym Centrum Nauki [2017/27/B/NZ3/00872];National Institute for Health Care Management Foundation [1R01EY031594-01A1];

Disclosure statement

No potential conflict of interest was reported by the authors.