FASN mediates crosstalk between autophagy and lipid metabolism via the AMPK-MTOR pathway in early age-related macular degeneration Authors: Jinyu Cai, Yanmei Jiang, Xingyu Liu, Tao Yang, Peizeng Yang, Ling Chen Published: 2026 ABSTRACT ABSTRACTAge-related macular degeneration (AMD) involves sub-retinal pigment epithelium (sub-RPE) lipid deposition in the early stage, with dysregulated lipid metabolism and impaired macroautophagy/autophagy implicated, yet the molecular mechanisms underlying their interaction remain unclear. In this study, transcriptomic analysis of human macular tissues identified FASN (fatty acid synthase), a regulator of lipid metabolism and lysosomal function, as a significantly upregulated key hub gene in early AMD. In apoe−/− mice fed a high-fat diet (HFD), retina-RPE-choroid complexes revealed elevated FASN alongside autophagy suppression, lysosomal dysfunction, and lipid accumulation. In vitro, FASN protein levels increased in RPE cells treated with the autophagy inhibitor 3-methyladenine (3-MA), but decreased with the autophagy activator rapamycin (RAPA), without transcriptional changes; lysosomal blockade with chloroquine (CQ) induced FASN accumulation, which was significantly delayed following autophagy inhibition. These findings indicate that FASN accumulation results from insufficient autophagic degradation. Conversely, FASN knockdown or pharmacological inhibition enhanced autophagic flux and promoted lysosomal lipid clearance in RPE cells. Mechanistically, FASN inhibition increased AMPK phosphorylation and decreased MTOR activity, thereby facilitating autophagy and lipophagy. Collectively, our findings reveal a self-amplifying pathological circuit in early AMD: autophagy impairment drives FASN accumulation, which in turn exacerbates lysosomal dysfunction and lipid accumulation. Targeting the FASN-AMPK-MTOR axis may offer a promising therapeutic strategy for early AMD. Introduction Age-related macular degeneration (AMD) is a degenerative macular disorder predominantly affecting the elderly worldwide and is characterized by progressive central vision loss [1]. Age is the foremost risk factor, with the majority of advanced cases and severe visual impairment occurring in individuals aged 60 or above [2,3]. The prevalence of AMD continues to rise in the context of global population aging, posing an increasing public health challenge worldwide. Clinically, AMD is classified into early, intermediate, and advanced stages. Early AMD, characterized by the presence of small drusen or mild retinal pigment epithelial (RPE) abnormalities, is typically asymptomatic. However, once the disease progresses to the advanced stage, it will manifest as geographic atrophy or macular neovascularization and result in severe vision loss or blindness [4]. Although current therapies can slow the progression of neovascular AMD, treatment options for other forms of AMD remain limited [5]. Therefore, elucidating the initial factors in the early stage is crucial for developing earlier interventions for AMD. Although genetic predisposition, complement dysregulation, chronic oxidative stress, and mitochondrial dysfunction have all been implicated in the progression of AMD, abnormal lipid metabolism in RPE has recently emerged as a key pathological factor [6–10]. Systematic gene sequencing studies have identified numerous genes involved in lipid transport and metabolism that are associated with AMD pathogenesis [11,12]. Furthermore, sub-RPE deposits, a hallmark pathological feature of AMD, accumulate progressively over the course of the disease [13]. These deposits, primarily composed of hydrophobic neutral lipids including esterified and unesterified cholesterol [14], evolve from diffuse lipid accumulation to focal soft drusen, ultimately impeding metabolic exchange between the outer retina and choroid [15]. This process constitutes a critical pathological basis for early AMD. Together, these findings underscore the significant role of dysregulated lipid metabolism in early AMD pathogenesis. However, the precise mechanisms underlying the formation of sub-RPE deposits remain unclear. Impaired macroautophagy/autophagy and dysfunctional lysosomal clearance in RPE may be the initiating factors in the formation of sub-RPE deposits [9,16–18]. As specialized phagocytic cells, RPE exhibits high autophagic activity, robust endo-lysosomal function, and the highest acidic lipase activity among ocular tissues [19]. These cells rely on efficient autophagy and lysosomal degradation to clear photoreceptor outer segments (POS), which are rich in neutral lipids [20]. An age-related decline in the autophagic capacity of RPE is believed to underlie the intracellular lipid accumulation, ultimately leading to RPE dysfunction and the development of sub-RPE deposits [21]. Beyond this classical degradative pathway, non-canonical forms of autophagy are implicated in AMD pathogenesis. Dysfunctional selective autophagy, particularly lipophagy (targeting degradation of lipids) and mitophagy (clearance of damaged mitochondria), drives AMD onset and progression [22,23]. Furthermore, pathogenic factors released through enhanced secretory autophagy or delivered via exosomes may also promote AMD pathology [18,24,25]. Nevertheless, the molecular mechanisms linking aberrant lipid metabolism to impaired autophagy in early AMD are poorly defined, and key regulatory nodes integrating these pathways remain elusive. In this study, we integrated transcriptomic analysis of macular tissues from AMD donors with functional experiments in apoe−/− mice fed a high-fat diet (HFD) and RPE cells, and identified FASN (fatty acid synthase), which was significantly upregulated in early AMD progression, as a critical regulator linking lipid metabolism and autophagy. Suppression of autophagy led to the accumulation of FASN, which in turn exacerbated lysosomal dysfunction and lipid accumulation in RPE cells. This cascade ultimately intensified sub-RPE lipid deposition and contributed to the pathogenesis of early AMD. Results Lipid metabolism and lysosomal function are linked to early AMD pathogenesis To elucidate the pathogenic mechanisms underlying early AMD, we focused on the transcriptomic data from the macular region of the RPE-choroid and retina, and conducted bioinformatic analyses to identify the differentially expressed genes (DEGs) in this disease. A total of 1018 DEGs were identified in RPE-choroid samples, including 515 upregulated and 503 downregulated genes (Figure 1A, Table S1). To further clarify the biological functions of these DEGs, we performed Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analyses. Figure 1B shows the top 10 significantly enriched GO terms, which were primarily associated with cholesterol biosynthetic and metabolic processes, with lysosomal function also playing a notable role. KEGG analysis revealed that steroid biosynthesis was the most significantly enriched pathway, along with marked enrichment in metabolic pathways (Figure 1C). Figure 1.Identification and functional enrichment of DEGs in early AMD. (A) Volcano plot showing DEGs in the RPE-choroid of early AMD patients compared to healthy controls. Significantly upregulated and downregulated genes are highlighted in red and blue, respectively. (B, C) GO and KEGG pathway enrichment analyses of DEGs from the RPE-choroid. The top ten significantly enriched terms are displayed, with the gene count indicating the number of DEGs in the specified pathway. (D, E) GO and KEGG enrichment analyses of DEGs from the retina of early AMD patients.The image features five plots analyzing differentially expressed genes (DEGs) in early age-related macular degeneration (AMD). Plot A is a volcano plot showing DEGs in the RPE-choroid of early AMD patients versus healthy controls, highlighting significant gene changes. Plot B presents Gene Ontology (GO) enrichment analysis of DEGs from the RPE-choroid, listing the top ten enriched terms with gene counts. Plot C shows KEGG pathway enrichment analysis of DEGs from the RPE-choroid, emphasizing pathways like steroid biosynthesis and metabolic pathways. Plot D provides GO enrichment analysis of DEGs from the retina, detailing biological processes, cellular components and molecular functions. Plot E illustrates KEGG pathway enrichment analysis of DEGs from the retina, indicating pathways such as ABC transporters and complement and coagulation cascades.Five plots on DEGs in early AMD: volcano plot, GO/KEGG analyses for RPE-choroid and retina. Identification and functional enrichment of DEGs in early AMD. (A) Volcano plot showing DEGs in the RPE-choroid of early AMD patients compared to healthy controls. Significantly upregulated and downregulated genes are highlighted in red and blue, respectively. (B, C) GO and KEGG pathway enrichment analyses of DEGs from the RPE-choroid. The top ten significantly enriched terms are displayed, with the gene count indicating the number of DEGs in the specified pathway. (D, E) GO and KEGG enrichment analyses of DEGs from the retina of early AMD patients. To verify whether these findings are generally applicable to retinal tissues, we analyzed DEGs from the macular retina of early AMD patients (Table S2). GO enrichment analysis indicated that, in addition to lipid-related processes, phagocytosis, receptor-mediated endocytosis, complement activation, and inflammatory response were strongly associated with early AMD (Figure 1D). Consistent with observations in the RPE-choroid complex, organelles and membrane structures, including lysosomes, endoplasmic reticulum, and exosomes, were functionally implicated in early AMD pathogenesis. KEGG analysis of the macular retina (Figure 1E) further highlighted significant enrichment in the complement and coagulation cascades pathway, followed by the phagosome pathway and cholesterol metabolism. These results corroborate and extend the findings from the RPE-choroid complex, revealing a clear connection between lipid synthesis and metabolism, phagosome function, complement activation, inflammatory response, and early AMD. Furthermore, these results highlighted the important pathogenic role of lysosomes in AMD development. Lipid-related processes and tissue morphogenesis are key pathways in early AMD We subsequently constructed a protein-protein interaction (PPI) network based on the identified DEGs (Figure 2A), in which node size and color intensity reflect the strength of association with early AMD. Using the MCODE plugin in Cytoscape, we identified a key functional module closely related to lipid synthesis and metabolism (score: 14.714) (Figure 2B). Functional analysis of DEGs was further performed using the ClueGO and CluePedia plugins. Figure 2C shows the top 10 most relevant GO biological processes (BP) terms related to early AMD pathogenesis. Among these, the biosynthetic process of organic hydroxy compounds (such as cholesterol and steroids) (red network) and tissue morphogenesis (green network) exhibited the most extensive interactions, suggesting their critical roles in early AMD pathogenesis and progression. Figure 2D illustrates the GO-BP terms most relevant to upregulated (red) and downregulated (blue) DEGs. The results indicated that biological regulation of lipids and other organic nitrogen compounds was primarily driven by upregulated DEGs, while downregulated DEGs were more closely linked to tissue morphology. Furthermore, we identified major DEGs involved in these pathways (Figure 2E), revealing 24 DEGs related to lipid biosynthesis and metabolism, most of which were upregulated. Figure 2.PPI network analysis of DEGs in the RPE-choroid of early AMD patients. (A) Overview of the PPI network. Nodes represent upregulated (yellow) or downregulated (blue) genes; larger and darker nodes indicate higher disease relevance. (B) A key module from the PPI network, strongly associated with lipid biosynthesis and metabolism. (C) Interaction network of significantly enriched GO-BP terms. Nodes of the same color represent functionally related pathways. (D) The GO-BP network highlights terms predominantly enriched by upregulated (red) or downregulated (blue) genes. (E) Key pathways and their associated DEGs. Large nodes represent pathways (red: upregulated gene-dominated; blue: downregulated gene-dominated; gray: no clear dominance). Small dots represent individual genes (red: upregulated; blue: downregulated).Image A presents a protein-protein interaction network analysis of genes in the retinal pigment epithelium-choroid of early age-related macular degeneration patients. Nodes indicate gene expression changes, with larger, darker nodes showing higher disease relevance. Image B highlights a module linked to lipid biosynthesis and metabolism. Image C displays an interaction network of enriched Gene Ontology biological process terms, with nodes of the same color indicating related pathways. Image D emphasizes Gene Ontology terms enriched by gene expression changes. Image E illustrates key pathways and associated genes, with large nodes for pathways and small dots for genes. These diagrams underscore the importance of lipid biosynthesis and metabolism in the pathogenesis of early age-related macular degeneration.Diagrams of PPI network analysis in early AMD patients, focusing on lipid biosynthesis and metabolism. PPI network analysis of DEGs in the RPE-choroid of early AMD patients. (A) Overview of the PPI network. Nodes represent upregulated (yellow) or downregulated (blue) genes; larger and darker nodes indicate higher disease relevance. (B) A key module from the PPI network, strongly associated with lipid biosynthesis and metabolism. (C) Interaction network of significantly enriched GO-BP terms. Nodes of the same color represent functionally related pathways. (D) The GO-BP network highlights terms predominantly enriched by upregulated (red) or downregulated (blue) genes. (E) Key pathways and their associated DEGs. Large nodes represent pathways (red: upregulated gene-dominated; blue: downregulated gene-dominated; gray: no clear dominance). Small dots represent individual genes (red: upregulated; blue: downregulated). FASN is a central hub gene linking lipid metabolism and autophagy Our analyses revealed a functional distinction between upregulated and downregulated DEGs in early AMD progression, with lipid synthesis and metabolism being predominantly associated with upregulated DEGs. We therefore conducted a targeted bioinformatic analysis focusing on the upregulated DEGs. GO analysis confirmed their strong involvement in lipid-related processes, including the biosynthesis and metabolism of cholesterol, ergosterol, and phospholipids (Figure 3A). These upregulated DEGs were also implicated in the regulation of organelles and membrane structures, such as lysosomes, the endoplasmic reticulum, and exosomes. KEGG analysis further indicated significant enrichment in metabolic pathways, particularly those related to lipid biosynthesis and metabolism (Figure 3B). Subsequent functional analysis of PPI network revealed a relatively focused interaction network, also predominantly centered on lipid-related pathways (Figure 3C). The key upregulated DEGs within this network are displayed in Figure 3D. Importantly, a key module identified by upregulated DEGs was closely associated with lipid functions and membrane structure regulation (Figure 3E). This module involved 17 DEGs, among which FASN exhibited the strongest association with early AMD. Figure 3.Bioinformatic analysis of upregulated genes and identification of hub genes in early AMD. (A, B) Enrichment analysis of GO terms and KEGG pathways for upregulated DEGs. (C) Interaction network of enriched GO-BP terms for upregulated DEGs. (D) Core genes within the highlighted pathways. (E) A key module from the PPI network of upregulated genes. Node size and color intensity correspond to the strength of association with early AMD. (F) Hub genes identified by 12 algorithms using the cytoHubba plugin. (G) The PPI network constructed between FASN and the 43 autophagy-related DEGs identified in macular RPE-choroid and retinal tissues from early AMD patients. Red nodes represent genes directly connected to FASN, while blue nodes denote those indirectly connected. Node size and color intensity correspond to the number of direct interactions within the network.The image features seven panels analyzing upregulated genes in early AMD. Panel A displays a bar graph of GO term enrichment for biological processes, cellular components and molecular functions. Panel B shows a dot plot of KEGG pathway enrichment. Panel C illustrates a network of enriched GO-BP terms, focusing on lipid processes. Panel D highlights core genes in pathways related to cholesterol and sterol biosynthesis. Panel E presents a PPI network module with FASN as a central hub. Panel F identifies hub genes using 12 algorithms, emphasizing FASN. Panel G shows a PPI network between FASN and 43 autophagy-related DEGs, with red nodes for direct and blue for indirect connections.Bioinformatic analysis of early AMD: upregulated genes, enrichment, networks, hub genes. Bioinformatic analysis of upregulated genes and identification of hub genes in early AMD. (A, B) Enrichment analysis of GO terms and KEGG pathways for upregulated DEGs. (C) Interaction network of enriched GO-BP terms for upregulated DEGs. (D) Core genes within the highlighted pathways. (E) A key module from the PPI network of upregulated genes. Node size and color intensity correspond to the strength of association with early AMD. (F) Hub genes identified by 12 algorithms using the cytoHubba plugin. (G) The PPI network constructed between FASN and the 43 autophagy-related DEGs identified in macular RPE-choroid and retinal tissues from early AMD patients. Red nodes represent genes directly connected to FASN, while blue nodes denote those indirectly connected. Node size and color intensity correspond to the number of direct interactions within the network. To enhance the sensitivity and specificity of DEGs identification, we applied 12 algorithms provided by the cytoHubba plugin to pinpoint the most significant upregulated genes (Table S3). Most hub genes are associated with cholesterol synthesis and steroid metabolism, as well as cell metabolism and energy (Table S4). ACAT2, GFAP, and DHCR7 were recognized by 7 algorithms; HMGCS1, EEF2, and FDPS were detected by 6 algorithms. FASN, however, was the only gene identified by 8 algorithms and consistently ranked among the top 3 across all analyses, confirming its status as a central hub gene in early AMD pathogenesis (Figure 3F). As illustrated in Figure 3E, FASN also contributed to the regulation of lipid metabolism and membrane function (such as lysosomes), suggesting its potential role in mediating the crosstalk between lipid-related processes and autophagy. To further validate this, we identified 43 autophagy-related DEGs from the macular RPE-choroid and retinal tissues of early AMD patients, based on a set of 604 previously reported autophagy-related genes [26]. The PPI analysis revealed that 15 of these DEGs directly interacted with FASN (red nodes), while the remaining genes exhibited indirect interactions (blue nodes) (Figure 3G; Table S5). Significantly, the lysosomal marker LAMP1 (lysosome associated membrane protein 1), a direct effector of FASN, occupied a central hub within the autophagy-lysosome regulatory network (engaging with the highest number of genes in this network), suggesting that FASN might directly modulate autophagy, particularly by influencing lysosomal function. FASN is the rate-limiting enzyme for endogenous fatty acid synthesis [27], and its dysfunction, leading to disrupted lipid homeostasis, has been well-documented in cancer [28–31]. However, evidence supporting a role for FASN in mediating the interplay between lipid metabolism and autophagy in AMD pathogenesis remains scarce. Based on these findings, we selected FASN as a key hub gene for further investigation into the mechanisms of early AMD. Early AMD mouse model exhibits FASN accumulation and autophagy impairment in the ocular fundus Transcriptomic sequencing revealed upregulation of FASN in ocular tissues from early AMD patients. To validate these findings, we generated an early AMD model using B6/JGpt-Apoeem1Cd82/Gpt (apoe−/−) mice fed a HFD. Compared to wild type (WT) controls, apoe−/−-HFD mice exhibited increased body weight and larger body size (Figure 4A-B). Color fundus photography (CFP) showed numerous diffusely distributed drusen-like deposits in the fundus of apoe−/−-HFD mice (Figure 4C), which presented as a rough RPE layer on optical coherence tomography (OCT) (Figure 4D). RPE flat mounts from apoe−/−-HFD mice displayed disrupted cuboidal RPE morphology, multinuclear enlarged cells (yellow arrows) (Figure 4E-F), and patchy cytoplasmic distribution (pink arrows), suggesting impaired RPE integrity. Transmission electron microscopy (TEM) images further demonstrated an increase in vacuoles (green arrows) and lipid droplet (LD)-like particles (yellow arrows) within RPE cells of apoe−/−-HFD mice, along with partially undigested POS (blue arrow) and swollen endoplasmic reticulum (dark blue arrow) (Figure 4G). In contrast, RPE cells of WT mice showed minimal abnormalities mentioned above but more autophagy-related structures such as autolysosomes (red arrows), which were rarely observed in apoe−/−-HFD mice. Additionally, Bruch’s membrane (BrM) was generally thickened in apoe−/−-HFD mice (Figure 4H), with localized drusen-like deposits. Overall, these results confirm that apoe−/−-HFD mice recapitulate key early AMD phenotypes, including RPE degeneration and sub-RPE lipid deposition, and represent a suitable model for studying early AMD. Figure 4.The apoe−/−-HFD mouse model recapitulated pathological features of early AMD. (A) Body weight comparison between WT and apoe−/−-HFD mice (n = 10). (B) Representative photographs of mouse physiques. (C) Representative CFP images showing drusen-like deposits in apoe−/−-HFD mice. (D) Representative OCT images revealing rough RPE layer (red frame) in apoe−/−-HFD mice. (E) Representative retinal flat mounts stained for TJP1/ZO-1. Yellow arrows indicate multinuclear enlarged cells, while pink arrows highlight abnormal cytoplasmic staining observed in apoe−/−-HFD mice. Scale bar: 100 μm. (F) Quantification of relative RPE cell area between WT and apoe−/−-HFD mice (n = 3 flat mounts, 3 fields per mount). (G) TEM images of the RPE-choroid complex. Apoe−/−-HFD mice exhibited RPE abnormalities, including LDs accumulation (yellow arrow), vacuolization (green arrow), swollen mitochondria (dark blue arrow), thickened BrM, and undigested POS (light blue arrow). In contrast, WT mice displayed fewer abnormalities but more autophagy-related structures, such as autolysosomes (red arrows). Scale bars: 1 μm and 500 nm. (H) Quantification of BrM thickness (3 random measurements per mouse; n = 3 independent experiments). Data are presented as mean ± SEM. Statistical significance was determined by unpaired t-test; ***p < 0.001, ****p < 0.0001.Image A compares body weight between wild type (WT) and apoe-/- HFD mice, showing WT mice at 25-35g and apoe-/- HFD mice at 35-50g. Image B illustrates the physical differences between WT and apoe-/- HFD mice. Image C presents fundus photos of retinas, highlighting drusen-like deposits in apoe-/- HFD mice. Image D shows optical coherence tomography of the retinal pigment epithelium, indicating roughness in apoe-/- HFD mice. Image E displays retinal flat mounts stained for TJP1/ZO-1, with yellow arrows pointing to enlarged cells and pink arrows to abnormal staining in apoe-/- HFD mice. Image F quantifies retinal pigment epithelium cell area, comparing WT and apoe-/- HFD mice. Image G uses transmission electron microscopy to show abnormalities like lipid droplet accumulation and swollen mitochondria in apoe-/- HFD mice. Image H quantifies Bruch’s membrane thickness, comparing WT and apoe-/- HFD mice.Images compare wild type vs apoe-/- HFD mice: body weight, retina, cell structures. The apoe−/−-HFD mouse model recapitulated pathological features of early AMD. (A) Body weight comparison between WT and apoe−/−-HFD mice (n = 10). (B) Representative photographs of mouse physiques. (C) Representative CFP images showing drusen-like deposits in apoe−/−-HFD mice. (D) Representative OCT images revealing rough RPE layer (red frame) in apoe−/−-HFD mice. (E) Representative retinal flat mounts stained for TJP1/ZO-1. Yellow arrows indicate multinuclear enlarged cells, while pink arrows highlight abnormal cytoplasmic staining observed in apoe−/−-HFD mice. Scale bar: 100 μm. (F) Quantification of relative RPE cell area between WT and apoe−/−-HFD mice (n = 3 flat mounts, 3 fields per mount). (G) TEM images of the RPE-choroid complex. Apoe−/−-HFD mice exhibited RPE abnormalities, including LDs accumulation (yellow arrow), vacuolization (green arrow), swollen mitochondria (dark blue arrow), thickened BrM, and undigested POS (light blue arrow). In contrast, WT mice displayed fewer abnormalities but more autophagy-related structures, such as autolysosomes (red arrows). Scale bars: 1 μm and 500 nm. (H) Quantification of BrM thickness (3 random measurements per mouse; n = 3 independent experiments). Data are presented as mean ± SEM. Statistical significance was determined by unpaired t-test; ***p < 0.001, ****p < 0.0001. We then examined the expression and localization of FASN and key autophagy markers, involving MAP1LC3B/LC3B (microtubule associated protein 1 light chain 3 beta), SQSTM1/p62 (sequestosome 1), and LAMP1 in ocular tissues (Figure 5A, E, and I). FASN, MAP1LC3B/LC3B, and SQSTM1/p62 were widely expressed in the neural retina and RPE layer. Although LAMP1 was also observed in the neural retina, it was predominantly localized to the RPE layer, which was consistent with the phagocytic feature of RPE cells [21]. Quantitative fluorescence analysis revealed significant accumulation of FASN, MAP1LC3B/LC3B, and SQSTM1/p62 in retina-RPE-choroid complex of apoe−/−-HFD mice, while LAMP1 expression was decreased (Figure 5C, G, and K). Remarkable colocalization was observed between FASN and all three autophagy markers, suggesting that FASN might be subject to autophagic degradation. The increased colocalization of FASN with MAP1LC3B/LC3B and SQSTM1/p62, together with decreased colocalization with LAMP1 in apoe−/−-HFD mice, indicated autophagic flux blockade caused by lysosomal dysfunction as a potential mechanism underlying FASN accumulation. Given the irreplaceable role of RPE cells in AMD pathology, we also examined expression levels within the RPE layer. Fluorescent quantification results in the RPE layer were similar to those in the retina-RPE-choroid complex (Figure 5B, D, F, H, J, and L). However, compared to WT mice, apoe−/−-HFD mice exhibited more pronounced accumulation of MAP1LC3B/LC3B and SQSTM1/p62 in the RPE layer relative to the entire retina. Figure 5.The apoe−/−-HFD mice exhibited FASN accumulation and autophagy impairment in the fundus. (A, E, and I) Representative immunofluorescence images of mouse retinal-RPE-choroid complexes co-stained for FASN (red) and MAP1LC3B/LC3B, SQSTM1/p62, or LAMP1 (green). Colocalization appears yellow in the merged panels. Scale bar: 100 μm. (C, G, and K) Quantification analysis of relative levels of FASN, MAP1LC3B/LC3B, SQSTM1/p62, LAMP1, and PCC between WT and apoe−/−-HFD mice (co-staining results shown in A, E, and I) (n = 3 independent experiments; 3 fields per condition). (B, F, and J) Enlarged views of the RPE layer from A, E, and I. Scale bar: 50 μm. (D, H, and L) Quantification analysis of relative levels of FASN, MAP1LC3B/LC3B, SQSTM1/p62, LAMP1, and PCC between WT and apoe−/−-HFD mice (co-staining results shown in B, F, and J) (n = 3 independent experiments; 3 fields per condition). (M) WB analysis showing FASN protein accumulation in the retina-RPE-choroid complexes of apoe−/−-HFD mice (n = 4). (N) WB analysis of the LC3-II:LC3-I ratio and SQSTM1/p62 protein level revealing autophagic flux blockage in the retina-RPE-choroid complexes of apoe−/−-HFD mice (n = 3). (O) WB analysis of BECN1 (an early autophagy-related marker) and LAMP1 (a late autophagy-related marker), demonstrating autophagy impairment in the retina-RPE-choroid complexes of apoe−/−-HFD mice (n = 4). Data are presented as mean ± SEM. Statistical significance was determined by unpaired t-test; *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.Image A shows immunofluorescence of WT and -HFD mice retinal-RPE-choroid complexes with FASN and MAP1LC3B/LC3B staining. Scale: 100μm. Image B focuses on the RPE layer, scale: 50μm. Image C compares relative FASN, LC3B, PCC levels in WT and -HFD mice complexes. Image D provides similar data for the RPE layer. Image E features SQSTM1/p62 staining. Image F enlarges the RPE layer from E. Image G compares FASN, SQSTM1, PCC levels in the complex. Image H shows similar data for the RPE layer. Image I displays LAMP1 staining. Image J enlarges the RPE layer from I. Image K compares FASN, LAMP1, PCC levels in the complex. Image L shows similar data for the RPE layer. Image M presents WB analysis of FASN protein levels. Image N shows WB analysis of LC3-II:LC3-I ratio and SQSTM1/p62 protein levels. Image O provides WB analysis of BECN1 and LAMP1 protein levels. Statistical significance is indicated by asterisks.Images/graphs of FASN, autophagy markers in WT/-HFD mice retinal-RPE-choroid complexes. The apoe−/−-HFD mice exhibited FASN accumulation and autophagy impairment in the fundus. (A, E, and I) Representative immunofluorescence images of mouse retinal-RPE-choroid complexes co-stained for FASN (red) and MAP1LC3B/LC3B, SQSTM1/p62, or LAMP1 (green). Colocalization appears yellow in the merged panels. Scale bar: 100 μm. (C, G, and K) Quantification analysis of relative levels of FASN, MAP1LC3B/LC3B, SQSTM1/p62, LAMP1, and PCC between WT and apoe−/−-HFD mice (co-staining results shown in A, E, and I) (n = 3 independent experiments; 3 fields per condition). (B, F, and J) Enlarged views of the RPE layer from A, E, and I. Scale bar: 50 μm. (D, H, and L) Quantification analysis of relative levels of FASN, MAP1LC3B/LC3B, SQSTM1/p62, LAMP1, and PCC between WT and apoe−/−-HFD mice (co-staining results shown in B, F, and J) (n = 3 independent experiments; 3 fields per condition). (M) WB analysis showing FASN protein accumulation in the retina-RPE-choroid complexes of apoe−/−-HFD mice (n = 4). (N) WB analysis of the LC3-II:LC3-I ratio and SQSTM1/p62 protein level revealing autophagic flux blockage in the retina-RPE-choroid complexes of apoe−/−-HFD mice (n = 3). (O) WB analysis of BECN1 (an early autophagy-related marker) and LAMP1 (a late autophagy-related marker), demonstrating autophagy impairment in the retina-RPE-choroid complexes of apoe−/−-HFD mice (n = 4). Data are presented as mean ± SEM. Statistical significance was determined by unpaired t-test; *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. Western blotting (WB) analysis of the retina-RPE-choroid complex further confirmed elevated FASN expression, an increased LC3-II:LC3-I ratio, and accumulation of SQSTM1/p62 in apoe−/−-HFD mice (Figure 5M–N), supporting the notion that FASN aggregation was accompanied by blocked autophagic flux and lysosomal dysfunction. Additionally, downregulation of BECN1 (beclin 1; an early autophagy marker) and LAMP1 (lysosomal marker) in apoe−/−-HFD mice (Figure 5O) suggested the defects in autophagy initiation and lysosomal function. Defective autophagic-lysosomal degradation drives FASN accumulation Given the fundus-wide negative correlation between FASN levels and autophagy/lysosomal activity, we investigated whether autophagy directly regulates FASN levels in RPE cells. To modulate autophagic activity, adult retinal pigment epithelial cell line-19 (ARPE-19) cells were treated with either the autophagy activator rapamycin (RAPA) or the autophagy inhibitor 3-methyladenine (3-MA). Subsequent WB analysis revealed decreased FASN levels in RAPA-treated cells and increased FASN levels in 3-MA-treated cells (Figure 6A–B), which was corroborated by immunofluorescence staining (Figure 6C–D). These results indicate that autophagy can directly regulate the protein level of FASN in RPE cells. Figure 6.Defective autophagic-lysosomal degradation drives FASN accumulation. (A) Representative immunoblots of FASN in ARPE-19 cells treated with the autophagy inducer RAPA or the autophagy inhibitor 3-MA. (B) Quantification of FASN protein levels from (A) showing an inverse relationship between FASN accumulation and autophagy activity (n = 3). (C) Representative immunofluorescence images of FASN in ARPE-19 cells following pharmacological modulation of autophagy. Scale bar: 200 μm. (D) Quantification of relative FASN fluorescence intensity from (C) showing consistent results with WB analysis (n = 3 independent experiments; 5 fields per condition). (E) Real-time qPCR analysis showing that transcription of FASN was not significantly altered upon treatment with autophagy modulators (n = 6). (F) Representative immunoblots of LC3-II:LC3-I ratio, SQSTM1/p62, and FASN after treatment with 50 μM CQ for 0–24 h. (G) Quantification of (F) indicating that FASN accumulation lagged behind the blockage of autophagic flux. (H) Representative immunofluorescence images of MAP1LC3B/LC3B, SQSTM1/p62, and FASN after treatment with 50 μM CQ for 0–24 h. Scale bar: 100 μm. (I) Quantification analysis of (H) showing results consistent with WB analysis (n = 3 independent experiments; 5 fields per condition). Data are presented as mean ± SEM. Statistical significance was determined by one-way ANOVA; *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001; ns, not significant.The image A shows immunoblots of FASN and GAPDH in ARPE-19 cells treated with rapamycin (RAPA), 3-methyladenine (3-MA) and a control (HC). The image B shows a bar graph of relative FASN levels with RAPA, 3-MA and HC treatments, indicating significant differences. The x-axis is labeled with treatment types and the y-axis shows relative FASN level. The image C shows immunofluorescence images of FASN in ARPE-19 cells under different treatments, with a scale bar of 200 micrometers. The image D shows a graph of relative FASN intensity for the treatments, with the x-axis labeled with treatment types and the y-axis showing relative FASN intensity. The image E shows a graph of FASN mRNA expression level change, with the x-axis labeled with treatment types and the y-axis showing expression level change. The image F shows immunoblots of LC3-I, LC3-II, SQSTM1, FASN and GAPDH after treatment with 50 micromolar chloroquine (CQ) over time intervals. The image G shows a graph of relative levels of LC3-II:LC3-I ratio, SQSTM1 and FASN over time intervals, with the x-axis labeled with time intervals and the y-axis showing relative level. The image H shows immunofluorescence images of MAP1LC3B/LC3B, SQSTM1/p62 and FASN after treatment with 50 micromolar CQ over time intervals, with a scale bar of 100 micrometers. The image I shows a graph of relative intensity of LC3B, SQSTM1 and FASN over time intervals, with the x-axis labeled with time intervals and the y-axis showing relative intensity.Image: FASN levels & autophagy in ARPE-19 cells via immunoblots, graphs, fluorescence. Defective autophagic-lysosomal degradation drives FASN accumulation. (A) Representative immunoblots of FASN in ARPE-19 cells treated with the autophagy inducer RAPA or the autophagy inhibitor 3-MA. (B) Quantification of FASN protein levels from (A) showing an inverse relationship between FASN accumulation and autophagy activity (n = 3). (C) Representative immunofluorescence images of FASN in ARPE-19 cells following pharmacological modulation of autophagy. Scale bar: 200 μm. (D) Quantification of relative FASN fluorescence intensity from (C) showing consistent results with WB analysis (n = 3 independent experiments; 5 fields per condition). (E) Real-time qPCR analysis showing that transcription of FASN was not significantly altered upon treatment with autophagy modulators (n = 6). (F) Representative immunoblots of LC3-II:LC3-I ratio, SQSTM1/p62, and FASN after treatment with 50 μM CQ for 0–24 h. (G) Quantification of (F) indicating that FASN accumulation lagged behind the blockage of autophagic flux. (H) Representative immunofluorescence images of MAP1LC3B/LC3B, SQSTM1/p62, and FASN after treatment with 50 μM CQ for 0–24 h. Scale bar: 100 μm. (I) Quantification analysis of (H) showing results consistent with WB analysis (n = 3 independent experiments; 5 fields per condition). Data are presented as mean ± SEM. Statistical significance was determined by one-way ANOVA; *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001; ns, not significant. To determine whether autophagy influences FASN synthesis or degradation, we further found that FASN mRNA levels were unchanged upon autophagy modulation (Figure 6E), whereas its protein level was negatively correlated with autophagic activity. This indicates that FASN accumulation under early AMD conditions results from impaired autophagic degradation. Time-course experiments following chloroquine (CQ) treatment showed that autophagic flux blockade (elevated LC3-II:LC3-I ratio and SQSTM1/p62 accumulation) occurred within 3 h, whereas a significant increase in FASN protein levels was observed only at 24 h (Figure 6F-G). Immunofluorescence analysis corroborated these findings (Figure 6H-I), indicating that persistent autophagy disruption could precede and contribute to FASN accumulation. Furthermore, marked colocalization between FASN and SQSTM1/p62 suggested the involvement of SQSTM1/p62-mediated selective autophagy in FASN degradation (Figure 5E–H). SQSTM1/p62 acts as both a substrate and a receptor for selective autophagy, linking substrates to autophagosomes and promoting the selective clearance of cargos. To explore this further, we performed molecular docking analysis, which demonstrated a potential direct interaction between FASN and SQSTM1/p62 (Figure S1). This implies that FASN accumulation in early AMD may be linked to impaired SQSTM1/p62-mediated selective autophagy. FASN suppresses autophagy via the AMPK-MTOR pathway To elucidate the role and mechanism of FASN in autophagy regulation, we performed lentivirus-mediated knockdown of FASN (shFASN) in ARPE-19 cells. WB analysis confirmed a significant reduction in FASN expression (Figure S2A-B). Cell counting kit-8 (CCK-8) assays showed no significant difference in cell viability between shFASN and healthy control (HC) groups within the first 24 h, although a decrease was observed thereafter (Figure S2C). Since all subsequent experiments in vitro were completed within this 24-hour window, the observed viability change did not substantially affect the results. Wound healing assays indicated that FASN knockdown reduced cell migration (Figure S2D-E). This effect may be attributed to impaired FASN-mediated fatty acid synthesis, which compromises the supply of energy and biomembrane precursors necessary for migration [32]. Additionally, FASN is recognized to promote cell migration by activating pro-migratory signaling pathways and modulating cellular adhesion and cytoskeletal dynamics [32,33]. We found that FASN knockdown consistently enhanced autophagic activity, as evidenced by an increased LC3-II:LC3-I ratio and decreased SQSTM1/p62 (Figure 7A). Furthermore, the expression of key autophagic markers, including BECN1, LAMP1, and CTSD (cathepsin D), was upregulated (Figure 7B). Similar pro-autophagic effects were observed following treatment with 20 μM C75 (a FASN inhibitor) (Figure S3A), which concentration-dependently inhibited FASN levels (Figure S3B-C). Notably, ATG5 (autophagy related 5) was upregulated only in C75-treated cells, but not in the shFASN group (Figure 7C–D). These findings indicate that both genetic and pharmacological inhibition of FASN promote autophagic flux. Immunofluorescence analysis of both ARPE-19 cells and primary human RPE (hRPE) cells showed that both shFASN and C75 treatment increased the levels of MAP1LC3B/LC3B and LAMP1, with shFASN group exhibiting a stronger effect. Moreover, the combination of both interventions synergistically enhanced autophagy activation (Figure 7E–J). Taken together, these results demonstrate that FASN inhibition activates the entire autophagic process and enhances lysosomal function in ARPE-19 cells. Figure 7.Genetic knockdown or pharmacological inhibition of FASN activates autophagy in RPE cells. (A, B) Assessment of SQSTM1/p62 protein levels, the LC3-II:LC3-I ratio, and staged autophagy markers (BECN1, ATG5, LAMP1, CTSD) after FASN knockdown (n = 4). (C, D) Analysis of autophagy marker expression following treatment with the FASN inhibitor C75 (n = 4). (E) Representative immunofluorescence images showing MAP1LC3B/LC3B and LAMP1 in ARPE-19 cells after FASN knockdown or C75 treatment. Scale bar: 100 μm. (F, G) Quantitative analysis of MAP1LC3B/LC3B and LAMP1 fluorescence intensity from (E) (n = 3 independent experiments, with 5 fields per condition). (H) Representative immunofluorescence images showing MAP1LC3B/LC3B and LAMP1 in hRPE cells after FASN knockdown or C75 treatment. Scale bar: 100 μm. (I, J) Quantitative analysis of MAP1LC3B/LC3B and LAMP1 fluorescence intensity from (H) (n = 3 independent experiments, with 5 fields per condition). Data are presented as mean ± SEM. Statistical comparisons between two groups were performed using an unpaired t-test; comparisons among multiple groups were analyzed by one-way ANOVA; *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001; ns, not significant.The image contains multiple panels showing the effects of FASN knockdown and C75 treatment on autophagy markers in ARPE-19 and hRPE cells. Panel A shows graphs of relative LC3-II:LC3-I ratio and SQSTM1 levels after shHC and shFASN treatments. Panel B displays graphs of BECN1, ATG5, LAMP1 and CTSD levels after shHC and shFASN treatments. Panel C shows graphs of LC3-II:LC3-I ratio and SQSTM1 levels after HC and C75 treatments. Panel D presents graphs of BECN1, ATG5, LAMP1 and CTSD levels after HC and C75 treatments. Panel E shows immunofluorescence images of MAP1LC3B/LC3B and LAMP1 in ARPE-19 cells under different treatments. Panel F and G provide quantitative analysis of LC3B and LAMP1 intensity in ARPE-19 cells. Panel H shows immunofluorescence images of MAP1LC3B/LC3B and LAMP1 in hRPE cells. Panel I and J provide quantitative analysis of LC3B and LAMP1 intensity in hRPE cells. Scale bars are 100 micrometers.FASN knockdown and C75 impact autophagy markers in ARPE-19 and hRPE cells via graphs/images. Genetic knockdown or pharmacological inhibition of FASN activates autophagy in RPE cells. (A, B) Assessment of SQSTM1/p62 protein levels, the LC3-II:LC3-I ratio, and staged autophagy markers (BECN1, ATG5, LAMP1, CTSD) after FASN knockdown (n = 4). (C, D) Analysis of autophagy marker expression following treatment with the FASN inhibitor C75 (n = 4). (E) Representative immunofluorescence images showing MAP1LC3B/LC3B and LAMP1 in ARPE-19 cells after FASN knockdown or C75 treatment. Scale bar: 100 μm. (F, G) Quantitative analysis of MAP1LC3B/LC3B and LAMP1 fluorescence intensity from (E) (n = 3 independent experiments, with 5 fields per condition). (H) Representative immunofluorescence images showing MAP1LC3B/LC3B and LAMP1 in hRPE cells after FASN knockdown or C75 treatment. Scale bar: 100 μm. (I, J) Quantitative analysis of MAP1LC3B/LC3B and LAMP1 fluorescence intensity from (H) (n = 3 independent experiments, with 5 fields per condition). Data are presented as mean ± SEM. Statistical comparisons between two groups were performed using an unpaired t-test; comparisons among multiple groups were analyzed by one-way ANOVA; *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001; ns, not significant. Given the broad activation of autophagy upon FASN inhibition, we sought to determine whether this effect was mediated through the AMPK (adenosine 5’-monophosphate-activated protein kinase)-MTOR (mechanistic target of rapamycin kinase) pathway, a well-established regulatory axis of autophagy [34,35]. WB analysis revealed that FASN knockdown did not affect total AMPK protein levels but increased AMPK phosphorylation. Concurrently, both total and phosphorylated MTOR levels were significantly reduced (Figure 8A). These results suggest that FASN inhibition promotes autophagic flux by enhancing AMPK phosphorylation while suppressing MTOR signaling. Figure 8.FASN inhibition regulates autophagy in RPE cells via the AMPK-MTOR pathway. (A) Protein levels and phosphorylation status of AMPK and MTOR following FASN knockdown (n = 4). (B) Protein levels and phosphorylation status of AMPK and MTOR in FASN-knockdown cells treated with the AMPK inhibitors CC (n = 3). (C) Protein levels and phosphorylation status of AMPK and MTOR in FASN-knockdown cells treated with the MTOR activator MHY1485 (n = 3). Data are presented as mean ± SEM. Statistical comparisons between two groups were performed using an unpaired t-test; comparisons among multiple groups were analyzed by one-way ANOVA; *p < 0.05, **p < 0.01, ***p < 0.001; ns, not significant.The image A shows protein levels and phosphorylation status of AMPK and MTOR following FASN knockdown. The bar graphs display relative levels of AMPK, p-PRKAA over AMPK, MTOR and p-MTOR, with significant changes indicated. The image B shows protein levels and phosphorylation status of AMPK and MTOR in FASN-knockdown cells treated with AMPK inhibitors cc. The bar graphs illustrate relative levels of AMPK, p-PRKAA over AMPK, MTOR and p-MTOR, with statistical significance noted. The image C shows protein levels and phosphorylation status of AMPK and MTOR in FASN-knockdown cells treated with the MTOR activator MHY1485. The bar graphs present relative levels of AMPK, p-PRKAA over AMPK, MTOR and p-MTOR, highlighting significant differences. Each panel includes Western blot images and corresponding bar graphs for comparison.Three panels showing protein levels and phosphorylation of AMPK and MTOR in FASN-knockdown cells with treatments. FASN inhibition regulates autophagy in RPE cells via the AMPK-MTOR pathway. (A) Protein levels and phosphorylation status of AMPK and MTOR following FASN knockdown (n = 4). (B) Protein levels and phosphorylation status of AMPK and MTOR in FASN-knockdown cells treated with the AMPK inhibitors CC (n = 3). (C) Protein levels and phosphorylation status of AMPK and MTOR in FASN-knockdown cells treated with the MTOR activator MHY1485 (n = 3). Data are presented as mean ± SEM. Statistical comparisons between two groups were performed using an unpaired t-test; comparisons among multiple groups were analyzed by one-way ANOVA; *p < 0.05, **p < 0.01, ***p < 0.001; ns, not significant. To clarify the upstream-downstream relationship, we treated FASN-knockdown cells with the AMPK inhibitor Compound C (CC) and the MTOR activator MHY1485. CCK-8 assays first determined the optimal non-cytotoxic concentrations for 24-hour treatment as 2 µM for CC and 5 µM for MHY1485 (Figure S3D and G), which dose-dependently inhibited p-PRKAA/AMPK (Thr172) and activated p-MTOR (Ser2448), respectively (Figure S3E-F, H-I). CC effectively reduced the elevated p-PRKAA/AMPK in FASN-knockdown cells and restored the suppressed p-MTOR (Figure 8B). Conversely, MHY1485 did not alter the AMPK and p-PRKAA/AMPK levels but rescued the downregulated p-MTOR caused by FASN knockdown (Figure 8C). Collectively, inhibiting AMPK or rescuing MTOR activity reversed the core molecular alterations caused by FASN inhibition. These findings indicate that FASN contributes to autophagy impairment in early AMD via the FASN-AMPK-MTOR signaling axis. FASN inhibition enhances lysosomal lipid clearance To elucidate the role of FASN in lipid deposition and its interplay with autophagy, we pre-treated ARPE-19 cells with the FASN inhibitor C75 [36], prior to inducing neutral lipid accumulation with 400 μM oleic acid (OA) (Figure S3J). OA is a commonly used tool for inducing LDs formation and studying lipophagy due to its high efficacy and low cytotoxicity [37,38], which showed no significant cytotoxicity in RPE cells within 96 h (Figure S3K). Boron-dipyrromethene (BODIPY) staining showed rapid accumulation of LDs in both HC and C75-treated groups within 1 h. While lipid accumulation plateaued at around 6 h in HC cells, it reached saturation between 6 and 9 h in the C75 group (Figure 9A–D). These findings indicate that FASN inhibition alleviates intracellular lipid deposition, suggesting FASN as a key regulator of lipid metabolism in early AMD. Figure 9.FASN inhibition alleviates lipid accumulation and promotes lipophagy. (A) Representative images of BODIPY-stained LDs in ARPE-19 cells pretreated with or without the FASN inhibitor C75 and then challenged with OA. Scale bars: 40 μm and 200 μm. (B, C) Quantification of BODIPY fluorescence intensity over time in untreated and C75-pretreated cells, respectively (n = 3 independent experiments, with 5 fields per condition). (D) Comparison of LDs accumulation between the two groups. (E) Representative images showing LAMP1 and BODIPY co-labeling following treating ARPE-19 cells with C75 for 24 h, OA for 6 h, or the combination (reflecting lysosomal degradation of LDs). Scale bar: 100 μm. (F) Representative images showing LAMP1 and BODIPY co-labeling in hRPE cells following the same procedure described above. Scale bar: 100 μm. (G-H) Quantitation of BODIPY fluorescence, LAMP1 fluorescence, and their colocalization in ARPE-19 cells and hRPE cells (n = 3 independent experiments, with 5 fields per condition). Data are presented as mean ± SEM. Multiple group comparisons were analyzed by one-way ANOVA; *p < 0.001, **p < 0.0001; ns, not significant.The image A shows BODIPY-stained lipid droplets in ARPE-19 cells over time, comparing untreated (HC) and C75-treated groups at 0, 1, 3, 6, 9 and 12 hours. Insets highlight cellular details. The image B shows a bar graph with x-axis labeled 'Time (h)' and y-axis labeled 'Mean BODIPY intensity', displaying data for HC cells. The image C shows a similar graph for C75-treated cells. The image D shows a combined graph for HC and C75 groups in hRPE cells. The image E shows representative images of ARPE-19 cells with LAMP1 and BODIPY co-labeling under different treatments: HC, C75, OA and C75 plus OA. The image F shows similar images for hRPE cells. The image G shows a bar graph for ARPE-19 cells with x-axis labeled 'Treatment' and y-axis labeled 'Relative level', comparing LAMP1, BODIPY and PCC levels. The image H shows a similar graph for hRPE cells. Scale bars are provided for reference in the images.Lipid changes in ARPE-19/hRPE cells with FASN inhibitor C75 and oleic acid shown in images/graphs. FASN inhibition alleviates lipid accumulation and promotes lipophagy. (A) Representative images of BODIPY-stained LDs in ARPE-19 cells pretreated with or without the FASN inhibitor C75 and then challenged with OA. Scale bars: 40 μm and 200 μm. (B, C) Quantification of BODIPY fluorescence intensity over time in untreated and C75-pretreated cells, respectively (n = 3 independent experiments, with 5 fields per condition). (D) Comparison of LDs accumulation between the two groups. (E) Representative images showing LAMP1 and BODIPY co-labeling following treating ARPE-19 cells with C75 for 24 h, OA for 6 h, or the combination (reflecting lysosomal degradation of LDs). Scale bar: 100 μm. (F) Representative images showing LAMP1 and BODIPY co-labeling in hRPE cells following the same procedure described above. Scale bar: 100 μm. (G-H) Quantitation of BODIPY fluorescence, LAMP1 fluorescence, and their colocalization in ARPE-19 cells and hRPE cells (n = 3 independent experiments, with 5 fields per condition). Data are presented as mean ± SEM. Multiple group comparisons were analyzed by one-way ANOVA; *p < 0.001, **p < 0.0001; ns, not significant. To explore the mechanism underlying this situation, we assessed the role of FASN in lipophagy (the autophagic degradation of lipids) by co-staining for LAMP1 and BODIPY in ARPE-19 cells (Figure 9E and G) and hRPE cells (Figure 9F and H). C75 treatment alone for 24 h did not affect basal lipid levels but increased LAMP1 expression. OA treatment for 6 h significantly promoted lipid accumulation. Notably, the combination of C75 and OA treatment synergistically enhanced LAMP1 levels and concurrently reduced lipid deposition. The increased colocalization of LAMP1 with BODIPY (yellow patches) in the C75+OA group indicated that FASN inhibition promoted lysosomal function and facilitated LDs degradation via lipophagy. These results demonstrate that FASN-mediated lysosomal dysfunction and autophagy impairment contribute to defective lipid clearance in RPE cells, which is a key mechanism driving the formation of pathological deposits in early AMD. Discussion This study identifies FASN as a key hub gene upregulated in early AMD and establishes its role in coupling impaired autophagy with lipid dysregulation in RPE cells. We demonstrate that FASN accumulation results from insufficient autophagic degradation and, in turn, suppresses AMPK-MTOR signaling to exacerbate lysosomal dysfunction and lipid retention – forming a previously unreported self-amplifying pathological circuit (Figure 10). Figure 10.Schematic diagram illustrating the pathogenic feedback loop involving FASN accumulation, lipid deposition, and autophagy impairment in early AMD pathogenesis. Impaired autophagy in RPE cells results in the accumulation of LDs and FASN, due to inadequate selective autophagic degradation of lipid-rich POS and FASN. Elevated FASN further suppresses AMPK phosphorylation and promotes MTOR activation, thereby exacerbating autophagy impairment. The current study reveals a vicious cycle linking FASN accumulation, lipid deposition, and autophagy dysfunction in the pathogenesis of early AMD.The diagram outlines a feedback loop in early AMD involving FASN buildup, lipid deposition and impaired autophagy. Photoreceptor outer segments (POS) are at the top, with a nucleus and cellular structures below. Central to the diagram is autophagy impairment, showing the transition from phagophore to autophagosome to autolysosome, labeled with LC3-II, SQSTM1 and FASN. To the right, lipid droplets and FASN accumulation are depicted. Pathways indicate AMPK suppression and MTOR activation, worsening autophagy issues. FASN builds up due to poor autophagic degradation, creating a self-perpetuating pathological loop. At the bottom, drusen and BLinD accumulate above BrM, with mitochondria and FASN labeled. Arrows show interactions, highlighting the cycle of lipid dysregulation and autophagy dysfunction in RPE cells.Diagram showing FASN accumulation, lipid deposition and autophagy impairment in early AMD pathogenesis. Schematic diagram illustrating the pathogenic feedback loop involving FASN accumulation, lipid deposition, and autophagy impairment in early AMD pathogenesis. Impaired autophagy in RPE cells results in the accumulation of LDs and FASN, due to inadequate selective autophagic degradation of lipid-rich POS and FASN. Elevated FASN further suppresses AMPK phosphorylation and promotes MTOR activation, thereby exacerbating autophagy impairment. The current study reveals a vicious cycle linking FASN accumulation, lipid deposition, and autophagy dysfunction in the pathogenesis of early AMD. AMD is an aging-associated degenerative macular disorder [39], with approximately 85% of patients presenting early or intermediate disease characterized by soft drusen underneath the RPE [40]. Despite increasing recognition of lipid metabolism and autophagy as central pathogenic nodes of early or intermediate stages, their mechanistic interplay remains unclear [41]. Previous genetic studies have identified various lipid-related genes, including APOE, ABCA1, LIPC, CETP, and LPL as AMD risk factors [11]. Among these, APOE ε4 is protective while APOE ε2 increases risk [42]. ABCA1 variants elevate AMD prevalence [43], while certain LIPC variants may protect against AMD [44]. Beyond these risk genes, non-canonical DEGs such as OLR1, SIK1, and F3 may promote AMD progression by modulating inflammatory pathways [45], supporting that AMD arises from age-related damage accumulation [46]. Our transcriptomic analysis of human macular tissue revealed significant enrichment of lipid biosynthesis pathways and identified FASN, ACAT2, GFAP, and other lipid-associated genes among early AMD DEGs. These findings reinforce aberrant lipid homeostasis as a core pathological feature of AMD. Although these lipids are thought to originate primarily from RPE cells [19], the underlying mechanisms remain elusive – with impaired autophagy emerging as a key driver [16]. As specialized phagocytic cells with high acidic lipase activity, RPE cells rely on autophagy-lipid metabolism crosstalk to maintain homeostasis [19]. Previous studies have suggested that metformin and trehalose ameliorate AMD phenotypes by promoting ATM-AMPK-driven autophagy [47]. Elevated MTOR activity conversely disrupts autophagy, causing RPE dysfunction and AMD-like pathology [48]. Moreover, selective autophagy also plays a pivotal role in AMD progression. Dysfunctional lipophagy contributes to drusen formation [22], while impaired mitophagy increases oxidative stress and epithelial-mesenchymal transition in RPE cells [49]. Furthermore, chaperone-mediated autophagy has emerged as a novel therapeutic target in AMD recently [50]. Lysosomal dysfunction accompanied by enhanced secretory autophagy may facilitate drusen biogenesis [24,25,51]. We show that autophagic flux is blocked in early AMD and mouse models, with DEGs enriched in phagosome-related pathways. Notably, we link this defect to FASN: its accumulation stems from disrupted autophagic degradation, and accumulated FASN further suppresses autophagy. FASN, the rate-limiting enzyme in de novo long-chain fatty acid synthesis [52], regulates autophagy in a context-dependent manner. Studies indicate that FASN promotes autophagosome-lysosome fusion by facilitating the assembly of the STX17-SNAP29-VAMP8 complex in epithelial ovarian cancer [53], activates autophagy (elevating LC3-II:LC3-I ratio and BECN1, and reducing SQSTM1/p62) to facilitate tumor progression in osteosarcoma [54], while sustains MTOR signaling to suppress autophagy in colorectal cancer [55]. In ocular disease, although aberrant FASN expression has been implicated in diabetic retinopathy and pathological retinal angiogenesis [56,57], its role in AMD, particularly in RPE autophagy and lipid metabolism, remains poorly defined. A previous study reported elevated FASN transcript levels in a sodium iodate-induced dry AMD mouse model [58], but its role in sub-RPE lipid deposition was unexplored. Our study identified FASN as a core gene in early AMD, consistently upregulated in human AMD macular, apoe−/−-HFD mice, and lipid-loaded RPE cells. Mechanistically, lysosomal dysfunction led to FASN accumulation, which suppressed AMPK-MTOR signaling and impaired lipid clearance. Conversely, FASN inhibition promoted lysosomal lipid degradation and ameliorated lipid accumulation, establishing FASN as a molecular bridge between lipid dysregulation and autophagy failure in early AMD. A strength of this study is the use of more pathophysiologically relevant models for investigating fundus lipid deposition in early AMD. Although aged apoe−/− mice display early AMD-like features [47,59] and prolonged HFD feeding induces LDs accumulation in the RPE of WT mice [60,61], the combination of aging, APOE knockout, and HFD in our aged apoe−/−-HFD model exacerbated pathology, evidenced by more severe BrM thickening, abnormal electroretinography (ERG) findings (reduced responses in rod photoreceptor, post-photoreceptor, and oscillatory potentials) [22], and distinct retinal structural abnormalities [62]. This model recapitulates crucial features of early AMD, such as drusen-like deposits, RPE damage, and autophagy dysfunction. Despite its practical advantages, it fails to capture the chronic, multifactorial nature of human AMD. Ocular-specific FASN overexpression models may provide a valuable next step. Similarly, robust cellular models that recapitulate RPE lipid accumulation remain scarce. Long-term culture of primary porcine RPE cells can generate drusen-like deposits containing human-relevant components [63], but its extended timeline and technical complexity limit widespread adoption. More tractable cell models are therefore urgently needed. Given the complex interplay of risk factors in AMD pathogenesis, developing customized models tailored to specific mechanistic questions is particularly critical. In this study, we established a lipid-overload RPE cell model using OA, a major component of POS and lipofuscin [64]. OA efficiently induces LDs formation with lower lipotoxicity, endoplasmic reticulum stress, apoptosis, and inflammation than saturated fatty acids such as palmitic acid and stearic acid [65–68], enabling targeted interrogation of the autophagy-lipophagy axis [22,37,38]. We found that OA rapidly induced LDs formation in both ARPE-19 and primary hRPE cells, and FASN inhibition enhanced lipophagic clearance in this system, supporting its suitability for lipophagy studies. Several limitations should be acknowledged. First, FASN protein expression in human AMD macular tissue requires direct validation. Second, C75 is not a FASN-specific inhibitor, thus findings relying solely on its use should be interpreted cautiously. Although combining FASN genetic knockdown with pharmacological inhibition partially offsets this limitation, employing the Cas9-SAM system or creating ocular-specific FASN-overexpressing models remains the essential next step. Third, FASN is a crucial metabolic enzyme, systemic inhibition of FASN is clinically impractical for AMD, a chronic ocular disease. Local delivery strategies merit exploration. Finally, while we reveal a FASN-autophagy feedback loop and suggest that autophagy impairment precedes FASN accumulation, the precise temporal sequence remains to be resolved through dynamic tracing. However, the multifactorial nature of AMD pathogenesis makes it challenging to definitively establish the absolute sequence or dominance of events in this loop. In conclusion, this work establishes FASN as a critical molecular node linking impaired autophagy and lipid accumulation in early AMD. Targeting the FASN-AMPK-MTOR axis may offer a promising therapeutic strategy to restore RPE lipid homeostasis and halt disease progression at its earliest stages. Materials and methods Transcriptomic data collection Transcriptomic data (GSE29801) were obtained from the Gene Expression Omnibus database [69]. RNA expression profiling was performed using the GPL4133 platform. The dataset includes human RPE-choroid complexes (n = 177) and retina samples (n = 118) from macular or extramacular regions, covering 8 phenotypic categories: no ocular abnormality (normal); hard drusen only (<63 μm, Pre-AMD, MD1); soft drusen (>63 μm) without macular pigmentary changes (Sub-clinical pre-AMD, MD2); clinically diagnosed AMD; dry AMD; choroidal neovascularization (CNV); geographic atrophy (GA); and GA with CNV (GA/CNV). Given that AMD primarily affects the macular region and that this study focuses on its early stage, we analyzed only macular samples from both the RPE-choroid complex and the retina obtained from normal controls and early AMD phenotypes, including MD2, clinically diagnosed AMD, and dry AMD. The final analytical cohort consisted of 50 normal RPE-choroid samples, 26 early AMD RPE-choroid samples, 28 normal retina samples, and 18 early AMD retina samples. DEGs identification and enrichment analysis DEGs between patients and normal samples were identified using thresholds of fold change ≥ 1.2 or ≤0.833 and p-value < 0.05. GO and KEGG enrichment analyses were performed to determine the biological functions of the DEGs. The GO analysis covered three categories, biological process, cellular component, and molecular function, to provide functional annotations. KEGG pathway enrichment analysis was applied to identify pathways involved in the pathogenesis of early AMD. PPI network analysis and hub gene identification A PPI network of DEGs was constructed using the STRING database with a medium confidence threshold (0.4) and visualized in Cytoscape 3.9.1. Highly interconnected modules were identified using the MCODE plugin. Functional enrichment analysis of GO terms among the DEGs was performed and visualized with the ClueGO and CluePedia plugins. Hub genes were screened using the cytoHubba plugin. To enhance the sensitivity and specificity of hub gene identification, the top 10 genes from each of the 12 topological algorithms available in cytoHubba were compiled, and overlapping genes across these lists were defined as key hub genes in early AMD. Animal experiments Eight-month-old apoe−/− mice along with age- and sex-matched WT mice were purchased from Shanghai Model Organisms Center (Shanghai, China). All mice were housed under specific pathogen-free conditions at 22°C with a 12-hour light/dark cycle. Fifteen apoe−/− mice were fed a high-fat diet (Research Diets, New Brunswick, NJ, USA) for 3 months, whereas 15 WT mice received standard chow for the same period. After the dietary intervention, all mice were subjected to phenotypic evaluation, including morphological photography, body weight measurement, and fundus imaging. All experimental procedures complied with the ARRIVE guidelines and the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research, and were approved by the Ethics Committee of the First Affiliated Hospital of Chongqing Medical University. CFP and OCT RPE abnormalities and soft drusen-like deposits were evaluated using CFP and OCT. Mice were anesthetized, and pupils were dilated with 1% atropine sulfate. Fundus imaging and structural assessments were conducted using the phoenix micron IV retinal imaging microscope and a spectral-domain optical coherence tomography system (both from Phoenix Research Labs, Pleasanton, CA, USA). RPE flat mount staining RPE flat mount staining was performed to evaluate the integrity of RPE cell junctions. Briefly, enucleated eyes were fixed in 4% paraformaldehyde at 4°C for 12 h. The RPE-choroid complexes were then isolated, radially incised, and flattened. After permeabilization and blocking with 0.25% Triton X-100 (Beyotime Biotechnology, P0096) and 5% bovine serum albumin (Saiguo Biotechnology, 4240GR025) for 2 h, the tissues were incubated with TJP1/ZO-1 recombinant monoclonal antibody (Proteintech, 82,870–7-RR; 1:1000) overnight at 4°C. Following washes, the tissues were incubated with a secondary antibody, goat anti-rabbit IgG H&L (Alexa Fluor® 488; Abcam, ab150077; 1:500) for 2 h at 25°C in the dark. Finally, the tissues were mounted using the antifade mounting medium with DAPI (Beyotime Biotechnology, P0131) and visualized under an inverted fluorescence microscope (Leica, Wetzlar, Germany). For quantitative analysis, three random non-overlapping fields of view were imaged from the central region of each RPE flat-mount (3 flat mounts per group). The relative cell area was then quantified using ImageJ software. TEM TEM was employed to examine LDs and autophagy-related structures within RPE cells, and to measure the thickness of BrM. Posterior eye segments were fixed in 2.5% glutaraldehyde at 4°C for 12 h, then post-fixed in 1% osmium tetroxide for 1–2 h. Subsequently, the samples were dehydrated through a graded series of ethanol and acetone, and permeabilized with a mixture of Spurr’s resin (Sigma-Aldrich, EM0300) and acetone. Following this, the tissues were stained with uranyl acetate and lead citrate, and examined under a TEM system (Hitachi HT7800, Tokyo, Japan) for imaging and BrM thickness measurement. hRPE cell isolation Human RPE cells were isolated from 6 donor eyeballs (4 males, 2 females; no AMD history) obtained from the Chongqing Eye Bank (Ophthalmology Medical Center, The First Affiliated Hospital of Chongqing Medical University) within 6 h postmortem (Table S6). After immersion in 20% penicillin-streptomycin for 10 min, the anterior segment, vitreous, and neural retina were removed. Eyecups were filled with 0.25% trypsin-EDTA and incubated at 37°C for 20 min. Dissociated cells were collected and neutralized with cell culture medium. This digestion and neutralization cycle was repeated 3 times. Cells were pelleted by centrifugation at 300 ×g for 10 min and cultured in medium containing 20% fetal bovine serum (FBS). After 5–6 days, hRPE cells exhibited typical cobblestone morphology with abundant pigment granules. Cells were passaged 2–3 times until significant pigment loss was observed, then used for immunofluorescence. The study complied with the Declaration of Helsinki and was approved by the Ethics Committee of Chongqing Medical University (2023–286; IACUC-CQMU-2024–0884). Cell culture and treatment ARPE-19 cells were obtained from the American Type Culture Collection (CRL-2302), and were cultured according to established protocols. Briefly, cells were seeded in DMEM/F-12 medium (Gibco, C11330500BT) with 10% FBS and 1% penicillin-streptomycin and allowed to adhere for 12 h before treatments. To modulate autophagy, cells were treated with 100 nM RAPA (MedChemExpress, HY-10219) or 10 mM 3-MA (Sigma-Aldrich, M9281) for 12 h, and treated with 10 μM CQ (MedChemExpress, HY-17589A) for 0–24 h. To induce intracellular lipid accumulation, cells were exposed to 400 μM OA (Sigma-Aldrich, O7501) for 6 h. To suppress FASN expression, cells were transfected with FASN-targeting lentivirus particles (constructed by Zolgene Biotechnology, Fuzhou, China) at a multiplicity of infection of 25 for 72 h, cells transfected with empty lentiviral particles served as the control. For pharmacological interventions, FASN activity was inhibited with 20 μM C75 (MedChemExpress, HY-12364) for 24 h. AMPK activity was suppressed using 2 μM CC (MedChemExpress, HY-13418A) for 24 h, and MTOR was activated by 5 μM MHY1485 (MedChemExpress, HY-B0795) for 24 h. For all pharmacological treatments, corresponding control groups received an equal volume of dimethyl sulfoxide (DMSO; Solarbio Technology, D8371) to exclude potential solvent-related effects. Total RNA extraction and real-time quantitative polymerase chain reaction (RT-qPCR) Total RNA was extracted from ARPE-19 cells (5 × 106) using TRIzol (Invitrogen, 15596018CN). Reverse transcription of the isolated RNAs was performed using RT Master Mix for qPCR (gDNA digester plus; MedChemExpress, HY-K0511A). RT-qPCR amplification was conducted using SYBR Green qPCR Master Mix (Accurate Biology, AG11762) and detected with an Applied Biosystems 7500 Real-Time PCR System (Thermo Fisher Scientific). Primer sequences are provided in Table S7. Molecular docking analysis To investigate the potential interaction between FASN and SQSTM1/p62, molecular docking analysis was conducted. The three-dimensional structures of both proteins were retrieved from the Protein Data Bank. Docking simulations were performed using AutoDock Vina to predict the binding affinity and orientation. The resultant docking poses were then visualized and analyzed using PyMOL. CCK-8 assay Cell viability following FASN genetic suppression or pharmacological treatment was assessed using the CCK-8 assay. FASN transfected and control cells were seeded into 96-well plates and cultured for 3–48 h, followed by incubation with 10 μL of CCK-8 reagent (Beyotime Biotechnology, C0037; 5 mg/mL) for 2.5 h in the dark. Absorbance at 450 nm was measured using an enzyme-linked immunosorbent assay reader (Thermo Varioskan Flash, USA). To determine the optimal treatment dose of modulators, cells were exposed to a concentration gradient of each modulator or DMSO for 24 h, and the highest concentration maintaining > 80% cell viability was selected. Efficacy of each chosen dose was further verified by WB analysis. Wound healing assay Cell migration was assessed using a wound healing assay following FASN inhibition. Transfected and control cells were cultured in 6-well plates to >90% confluence, scratched with a sterile pipette tip, and incubated for 12–24 h. Wound closure was imaged at regular intervals under an inverted phase-contrast microscope (Leica, Wetzlar, Germany). The remaining wound area was measured using ImageJ software, and the migration rate was calculated to quantify migration capacity. LDs staining Neutral lipids in ARPE-19 cells were stained using BODIPY™ 493/503 (Thermo Fisher, D3922). Cells cultured on 14-mm-diameter glass coverslips (70% confluence) were fixed with 4% paraformaldehyde at 25°C for 30 min, incubated with 2 μM BODIPY™ 493/503 (Thermo Fisher, D3922) for 15 min protected from light, and mounted with 20 μL of antifade mounting medium with DAPI (Beyotime Biotechnology, P0131). Fluorescence images were acquired using an inverted fluorescence microscope (Leica, Wetzlar, Germany). Immunofluorescence Tissue immunofluorescence was employed to co-stain FASN with 3 autophagy markers (MAP1LC3B/LC3B, SQSTM1/p62, and LAMP1), to investigate the relationship between FASN expression and autophagic activity in early AMD. Cellular immunofluorescence was conducted to examine lipid deposition, autophagic activity, and lipophagy levels in ARPE-19 and hRPE cells following different treatment conditions. For tissue staining, paraffin-embedded mouse ocular sections were deparaffinized, rehydrated, and subjected to antigen retrieval (1 mM EDTA buffer, pH 9.0, 95°C, 20 min). Endogenous peroxidase activity was quenched with 3% hydrogen peroxide for 10 min, followed by blocking with goat serum (Beyotime Biotechnology, C0265) at 25°C for 30 min. Subsequently, sections were incubated with primary antibodies at 4°C overnight, followed by secondary antibody incubation at 25°C for 30 min after washing. Finally, fluorescent signals were visualized using an inverted fluorescence microscope (Leica, Wetzlar, Germany). Cellular immunofluorescence followed a similar procedure: cells were fixed with either 4% paraformaldehyde or ice-cold methanol, permeabilized with 0.1% Triton X-100 as needed, then blocked, immunostained, and imaged similarly. Quantitative images were acquired under identical microscope settings (including laser power, gain, and exposure time) within the same experiment session. Three (for tissue) or five (for cell) non-overlapping fields of view were imaged from the central region of each biological replicate sample (n = 3). Relative fluorescence intensity was measured using ImageJ software to quantify protein expression levels. Colocalization was assessed by Pearson’s correlation coefficient (PCC). The following antibodies were used: anti-FASN (Proteintech, 10,624–2-AP; 1:1000), anti-MAP1LC3B/LC3B (Abcam, ab192890; 1:500), anti-LAMP1 (Abcam, ab208943; 1:500) (Abcam, ab278043; 1:100), anti-SQSTM1/p62 (Afantibody, AFRM0235; 1:500), goat anti-mouse IgG H&L (Alexa Fluor® 488; Abcam, ab150113; 1:500), and goat anti-rabbit IgG H&L (Alexa Fluor® 594; Abcam, ab150080; 1:500). WB WB was performed to quantify the expression levels of FASN and autophagy-related proteins. In brief, tissue and cell samples were lysed using RIPA buffer, and total protein concentration was determined using a BCA assay kit (Beyotime Biotechnology, P0010). Then, standard WB procedures were conducted, including protein denaturation, gel loading, electrophoresis, membrane transfer, blocking, antibody incubation, and signal development. Band intensities were quantified using ImageJ software and normalized to GAPDH. Background subtraction was applied uniformly across all lanes. Data from 3 or 4 independent biological replicates were utilized for statistical analysis. The following antibodies were used in this study: anti-GAPDH (AF7021; 1:10000) and anti-p-PRKAA/AMPK (AF3423; 1:1000) from Affinity Biosciences; anti-FASN (10624–2-AP; 1:5000), anti-CTSD (83743–2-RR; 1:5000), anti-MTOR (81670–1-RR; 1:3000), anti-p-MTOR (80596–1-RR; 1:5000), anti-PRKAA/AMPK (80209–5-RR; 1:3000), and Multi-rAb™ HRP-goat anti-rabbit recombinant secondary antibody (H+L) (RGAR001; 1:10000) from Proteintech; anti-MAP1LC3B/LC3B (ab192890; 1:2000), anti-SQSTM1/p62 (ab109012; 1:1000), anti-BECN1 (ab207612; 1:2000), anti-LAMP1 (ab208943/ab278043; 1:1000), and anti-ATG5 (ab108327; 1:2000) from Abcam. Statistical analysis Statistical analyses and graph generation were performed using Prism 8.0.2 (GraphPad Software, CA, USA). Data are presented as mean ± standard error of the mean from at least 3 independent biological replicates. Differences between 2 groups were analyzed by unpaired Student’s t-test, and comparisons among multiple groups were assessed by one-way analysis of variance (ANOVA). A p-value < 0.05 was considered statistically significant. *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001, ns, not significant. Supplementary Material Supplementary_figures_and_legends_R3.docx Supplementary tables R2.xlsx