Oxidative stress mediated by lipid metabolism contributes to high glucose-induced senescence in retinal pigment epithelium

Author’s Accepted Manuscript Oxidative stress mediated by lipid metabolism contributes to high glucose-induced senescence in retinal pigment epithelium Qingqiu Chen, Li Tang, Guang Xin, Shiyi Li, Limei Ma, Yao Xu, Manjiao Zhuang, Qiuyang Xiong, Zeliang Wei, Zhihua Xing, Hai Niu, Wen Huang

PII: DOI: Reference:

www.elsevier.com

S0891-5849(18)31335-2 https://doi.org/10.1016/j.freeradbiomed.2018.10.419 FRB13983

To appear in: Free Radical Biology and Medicine Received date: 4 August 2018 Revised date: 9 October 2018 Accepted date: 10 October 2018 Cite this article as: Qingqiu Chen, Li Tang, Guang Xin, Shiyi Li, Limei Ma, Yao Xu, Manjiao Zhuang, Qiuyang Xiong, Zeliang Wei, Zhihua Xing, Hai Niu and Wen Huang, Oxidative stress mediated by lipid metabolism contributes to high glucose-induced senescence in retinal pigment epithelium, Free Radical Biology and Medicine, https://doi.org/10.1016/j.freeradbiomed.2018.10.419 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting galley proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Oxidative stress mediated by lipid metabolism contributes to high glucose-induced senescence in retinal pigment epithelium Qingqiu Chena,1, Li Tangb1, Guang Xina1, Shiyi Lia, Limei Maa, Yao Xua, Manjiao Zhuanga, Qiuyang Xionga, Zeliang Weia, Zhihua Xinga, Hai Niua*, Wen Huanga* a

Laboratory of Ethnopharmacology, West China School of Pharmacy, West China Hospital, Sichuan

University, Chengdu, Sichuan, China. b

Department of Ophthalmology, West China Hospital, Sichuan University, Chengdu, Sichuan, China.

[email protected] [email protected] *

Corresponding author: Wen Huang, Dr. Wen Huang Keyuan Road 4 No.1, Gaopeng Avenue, Gaoxin

District, Chengdu, Sichuan, 610041, China.Fax: +86-028-85164073, Tel: +86-028-85164076 *

Corresponding authors: Hai Niu

1

These authors contributed equally to this work

Abstract Retinal pigment epithelium (RPE) dysfunction is thought to increase the risk of the development and progression of diabetic retinopathy (DR), the leading cause of blindness. However, the molecular mechanism behind high glucose-induced RPE cell damage is still blurred. We reported that ARPE-19 exposed to 25 mM glucose for 48 h did not induce apoptosis, but senescence validated by SA-β-Gal staining, p21 expression and cell cycle distribution. High glucose also increased oxidant species that exerted a pivotal role in senescence, which could be relieved by the treatment with antioxidant N-acetylcysteine (NAC). The accumulation of lipid droplets and the increase of lipid oxidation were also observed in ARPE-19 treated with high glucose. And the supplementation of free fatty acids (FFAs) indicated that lipid metabolism was associated with the generation of hydrogen peroxide (H2O2) and subsequent senescence in ARPE-19. PI3K/Akt/mTOR signaling pathway was shown to be responsible for the accumulation of intracellular lipids by regulating fatty acid synthesis, which in turn controlled senescence. Furthermore, high glucose induced autophagy in ARPE-19 with the treatment of glucose for 48 h, and autophagy inhibitor hydroxychloroquine (HCQ) or bafilomycin further aggravated the senescence, accompanying by an increase in oxidant species. Whereas, prolonged high glucose exposure inhibited autophagy and increased apoptotic cells. Experiments above provide evidence that lipid metabolism plays an important role in oxidative stressed senescence of RPE. Graphical Abstract:

Abbreviations DR, diabetic retinopathy; RPE, retinal pigment epithelium; ROS, reactive oxygen species; NAC, N-acetylcysteine; PPARα, peroxisome proliferator-activated receptor alpha; CPT1, carnitine palmitoyl transferase I; ACOX1, acyl-coenzyme A oxidase 1; FAS, fatty acid synthase; MMP, mitochondrial membrane potential; MTT, 3-(4,5-dimethyl-2-thia-zolyl)-2, 5-diphenyl-2H-tetrazolium bromide; FFAs, free fatty acids; HCQ, hydroxychloroquine; SA-β-gal, senescence-associated beta-galactosidase; MCU, mitochondrial calcium uniporter; H2O2, hydrogen peroxide. Key words Senescence, retinal pigment epithelium, reactive oxygen species, lipid oxidation

1. Introduction Diabetic retinopathy (DR) is the leading cause of blindness and severe vision impairment[1]. Retinal pigment epithelial (RPE), located between the neurosensory retina and the vascular choroids, forms the outer blood-retinal barrier to maintain the normal structural and functional integrity of retina. RPE is one of the most vulnerable cell populations of retina, and its damage has been reported to be involved in the development and progression of DR[2]. It is widely known that oxidative stress is a key mediator of DR, and the injury of RPE is also triggered by elevating the accumulation of reactive oxygen species (ROS)[3]. Nonetheless, the cellular mechanisms connecting ROS with the appearance of RPE pathological changes have not been fully elucidated. Cellular senescence, a status of irreversible growth arrest characterized by drastic cytomorphological and metabolic changes, has been known to be triggered by sub-lethal concentration of ROS in accordance with “the free radical theory of aging”[4, 5]. Previous studies also demonstrated that high glucose induces cellular senescence in many cell lines such as endothelial cells[6], glomerular mesangial cells[7], renal tubular epithelial cells[8]. Yet, no studies have reported that high glucose could induce senescence in RPE as far as we know. The generation of oxidative stress in eyes of patients with diabetes is known to result from an increase in the production of ROS, and RPE is particularly prone to oxidative damage from high oxygen tension of high metabolic activity, physiological phagocytosis as well as life-long light illumination[3] [9]. The development of DR is occult, and when the patient feels a significant decline in vision, the lesion is often advanced. Presently, there are no effective drug therapies for DR due to the unclear progression of the disease. Most studies of oxidative stress and DR have focused on elucidating mechanisms of oxidative stress-inducing apoptosis, but the widespread death of RPE cells is not usually seen in early DR[10, 11]. It is entirely possible that cellular senescence may be the result of RPE in response to oxidative stress in early DR. Some studies demonstrated that disruption of lipid homeostasis is related to age-related diseases[12, 13]. However, it is still controversial whether such changes are a cause or a consequence of aging. A recent report revealed that addition of specific lipids to the media is sufficient to obviously increase cellular senescence[14]. Moreover, a common feature of β-oxidation whether in peroxisomes or mitochondria is the potential to generate ROS[14]. Mitochondrial β-oxidation is responsible for the degradation of short, medium and long-chain fatty acids, which is associated with increased formation of superoxide and H 2O2[15, 16]. Peroxisomal β-oxidation of long chain, very long chain and branched chain fatty acid take a role in the progress of generating H2O2[17, 18]. Thus, lipid metabolism might be responsible for the increased oxidant species in senescent cells. Key enzymes in the fatty acid synthesis pathway are known to be regulated by glucose. And high glucose could increase lipid droplets in many cell types [19-22]. Recent report from Donato et al. demonstrated that several genes and non-coding regulatory RNA involved in regulation of lipid metabolism in oxidative stressed RPE cells exhibit expression alterations[23]. They also reported that elongation of very long chain fatty acids protein 4 (ELOVL4), fundamental for lipid biosynthesis, is involved in several visual pathologies like Stargardt disease and retinitis pigmentosa, especially under oxidative stress[24]. Both revealed a tight connection between lipid alterations and retinal diseases. However, it is still unclear whether high glucose induces lipid accumulation in RPE cells, meanwhile, the role of lipids in RPE dysfunction in DR remains ambiguous.

Autophagy, a process of lysosomal self-degradation that helps maintain homeostatic balance, is activated in response to many stress stimuli that are constantly present, including oxidative stress, in RPE[25]. The relationship between autophagy and senescence is a basic concern in the field of aging research, but still inconclusive. Most studies demonstrated that autophagy is positively correlated with senescence. For example, Dutta et al. reported that impaired mitochondrial autophagy can increase cardiac aging[26]. And Tai et al. reported that restoring autophagy activity could be a promising way to retard oxidative stress-induced cell senescence[27]. Meanwhile, autophagy is also important for the conversion of cell fate. Premature senescent tumor cells can be switched from senescence to apoptosis by autophagy inhibition[28]. Therefore, autophagy may play a significant role in RPE damage induced by high glucose. In the present study, we exposed ARPE-19 cells to 25 mM glucose to investigate the damage mechanism of RPE dysfunction in DR. We found that high glucose led to senescence of ARPE-19 cells but not apoptosis by the accumulation of ROS within 48 h. ARPE-19 treated with high glucose also presented accumulated lipid droplets,

and

subsequent

lipid

oxidation

advanced

by

up-regulating

the

key

enzymes

of

fatty acids oxidation made this ROS accumulation more serious. Accumulated lipid droplets were regulated by PI3K/AKT/mTOR pathway via increasing the expression of fatty acid synthase (FAS). Meanwhile, high glucose induced autophagy in ARPE-19 with the treatment of glucose for 48 h. And autophagy inhibitor HCQ or bafilomycin further aggravated the senescence with an increase of oxidant species, which indicated that autophagy played a cytoprotective role by resisting oxidative stress caused by high glucose. Whereas, long-term high glucose treatment reduced autophagy and increased apoptotic cells.

2. Materials and Methods 2.1 Reagents. Unless otherwise stated, all chemicals and plastics were from Sigma-Aldrich (St. Louis, MO, USA). 740 Y-P was purchased from APEXBIO (Houston, TX, USA). Bafilomycin A1 was purchased from Jiangsu Keygen Biotech Corp, Ltd (Nanjing, china). Rapamycin and hydroxychloroquine (HCQ) was purchased from Sigma–Aldrich (St. Louis, MO, USA). LY294002 was purchased from EMD-Calbiochem (La Jolla, CA, USA).

2.2 Cell culture The human RPE cell line ARPE-19 was recipient from Laboratory of Molecular Genetics of Eye Disease, West China Hospital. ARPE-19 was cultured with DMEM/low glucose medium (HyClone, Logan, Utah, USA) containing 5 mM glucose, supplemented with 10% fetal bovine serum (FBS) and 1% antibiotic/antimycotic in a humidified incubator of 5% CO2 at 37 °C.

2.3 Cell viability assay MTT (3-(4,5-dimethyl-2-thia-zolyl)-2, 5-diphenyl-2H-tetrazolium bromide; Sigma-Aldrich, St. Louis, MO, USA) assay was performed for cell viability. Suspended cells were seeded into 96-well plates at a density of 5×103 cells/well. After adherence overnight, cells were treated with different concentrations of glucose (5 or 25 mM) for 24 and 48 h respectively. At the endpoint, 0.5 mg/mL MTT was incubated for 3 h, and the SpectraMax M5 microplate reader (Molecular Devices, LLC, Sunnyvale, CA, USA) was used to record the absorbance at 490 nm.

2.4 Hoechst staining

Cells were washed with iced PBS for three times, then incubated with Hoechst (10 μg/ml) staining solution (Beyotime Institute of Biotechnology, Shanghai, China) for 10 min at dark. After rinsed with PBS, the images of dyed cells were captured under a fluorescence microscope (Zeiss, Axiovert 200, Germany).

2.5 Cell cycle and apoptotic assays. Cells were treated with various concentrations of glucose for 24 h or 48 h and harvested, then cells were washed with PBS for 3 times. As for cell cycle distribution, cells were finally fixed using 75% ethanol at 4° C overnight. After centrifugation at 2000rpm for 5 min, cells were stained using a cell cycle analysis kit (Jiangsu Keygen Biotech Corp, Ltd, Nanjing, China). Briefly, DNA was stained with propidium iodide (PI) solution (50 μg/mL) according to the manufacturer's instructions. The cell cycle distribution was analyzed using a flow cytometer (Becton Dickinson, USA). Finally, the ModFit LT was used for data analysis. As for cell apoptosis, cells treated and harvested, then washed 3 times with PBS followed by staining with Annexin V-FITC and PI at 37 °C for 15 min. The data were obtained from flow cytometer detection. Finally, the CytExpert 2.0 was used for data analysis.

2.6 Western blot analysis Cells were washed with iced PBS and then lysed in RIPA buffer containing proteinase inhibitors, and the protein content was measured with a BCA protein assay kit (Beyotime Institute of Biotechnology, Shanghai, China). Afterward, the equal amount of protein samples (20 μg) was loaded to SDS-polyacrylamide gel electrophoresis. The isolated proteins in gel were transferred to polyvinylidene fluoride membranes (Millipore, Bedford, MA, USA). After blocked in 5% non-fat milk in TBST (10 mM Tris-HCl, pH 8.0, 150 mM NaCl, 0.1% Tween-20 ) for 1 h at room temperature, the membranes were incubated respectively in primary anti-body overnight at 4 °C. Following rinsed thoroughly with TBST, and the membranes were incubated in a 1:1000 dilution of horseradish peroxidase conjugated secondary antibody for 1 hour at 37℃. Finally, the enhanced chemiluminescence (ECL) reagent (Pierce, Rockford, IL, USA) was added to magnify the HRP signals, which was detected with Immun-Star© Western C© Chemiluminescent Kit (Bio-Rad, California, USA) Western blotting detection systems. Finally, results were analyzed by Quantity One Software (Bio-Rad, California, USA). Anti-caspase3 (1:1000), anti-LC3 (1:1000), anti-p62 (1:1000), antiCathepsin B (1:1000), anti-phospho-P38 MAPK (1:1000), anti-P38 MAPK (1:1000), anti-β-actin (1:1000), from Cell Signaling Technology (CST, Danvers, MA, USA); Anti-phospho-AKT (1:1000) from Abcam (Cambridge, MA, USA); anti-p21 (1:1000), anti-MCU (1:1000), anti-FAS (1:1000), anti-PPARα (1:1000), anti- phospho-mTOR (1:1000), Anti-ACOX1 (1:1000) from Santa Cruz Biotechnology (Santa Cruz, CA, USA), anti-CPT1 (1:1000) from Cloud-Clone Corp (CCC, Houston, TX, USA).

2.7 SA-β-Gal assay SA-β-Gal activity was measured by a senescence associated β-galactosidase staining kit (Beyotime Institute of Biotechnology, Shanghai, China). Briefly, cells were washed in PBS and fixed in 0.5% glutaraldehyde for 15 min at room temperature. After washed in PBS, cells were incubated in SA-β-Gal stain solution according to the kit instructions. And cells were examined under a light microscope (uX71; Olympus Corp., Tokyo, Japan).

2.8 Detection of intracellular H2O2 level Intracellular H2O2 level was detected by 2’,7’-dichlorodihydrofluorescein diacetate (DCFDA; Jiancheng

Bioengineering Institute, Nanjing, China) as previously reported[29]. Detailly, cells were collected and washed three times with serum-free medium. Then cells were incubated with 5 μM DCFDA at 37 ℃ for 20 min in the dark. After washed with PBS, the fluorescence intensity was detected by fluorescence microplate reader (Aynergy Mx, BioTek, USA).

2.9 Detection of mitochondrial superoxide level For the mitochondrial superoxide, cells were collected and washed three times with serum-free medium. Then cells were loaded with MitoSOX™ Red Mitochondrial Superoxide Indicator (Thermo Fisher Scientific, USA) for 15 min, and the fluorescence intensity was detected by fluorescence microplate reader (Aynergy Mx, BioTek, USA).

2.10 Measurement of intracellular and mitochondrial Ca2+ Cells were washed with HBSS (8 g/LNaCl, 0.4 g/L KCl, 1 g/L glucose, 60 mg/L KH2PO4, 47.5 mg/L Na2HPO4, PH 7.2) for 3 times. Subsequently, cells were incubated in HBSS containing Fluo-4AM (5 μM) or Rhod-2AM(2 μM) for 30 min at 37 ℃ in the dark. After rinsed by HBSS for 3 times, the images were observed under the fluorescence microscope (Zeiss, Axiovert 200, Germany).

2.11 Mitochondrial membrane potential assay 5,5′,6,6′-Tetrachloro-1,1′,3,3′-tetraethylbenzimidazolylcarbocyanine iodide (JC-1) was used for measuring mitochondrial membrane potential (MMP). The JC-1 monomer predominating in depolarized mitochondria emits green fluorescence, whereas the oligomer (J-aggregate) forming in mitochondria with potentials more negative than −140 mV emits red fluorescence. Accordingly, the relative proportion of green and red fluorescence can be used to measure the proportion of mitochondrial depolarization. Cells were washed with PBS for 3 times, and incubated with 1μM JC-1 (Invitrogen Life Technology, Grand Island, NY, USA) for 8 min at 37℃ and rinsed by PBS for 3 times. Images were captured by fluorescence microscope (Zeiss, Axiovert 200, Germany).

2.12 Lysosome contents measurement with LysoTracker Red DND-99 Lysosome was stained with LysoTracker Red (Life Technologies, L7528) according to the manufacturer’s protocol. Briefly, LysoTracker Red was diluted to working concentration (1 μM) and incubated with cells at 37℃ for 15 min. when washed 3 times with PBS, the images were taken using a fluorescence microscope (Zeiss, Axiovert 200, Germany).

2.13 Oil Red O staining Cells were washed in PBS and fixed in 4% polyoxymethylene for 15 min at room temperature. After washed by PBS for 3 times, cells were incubated with Oil Red O solution (Sigma-Aldrich, St. Louis, MO, USA) for 30 min at 37℃. Then washed with PBS for 3 times, and examined under a light microscope (uX71; Olympus Corp, Tokyo, Japan).

2.14 Malondialdehyde (MDA) level Lipid peroxidation was determined by cell malondialdehyde (MDA) assay kit (Jiancheng Bioengineering Institute, Nanjing, China). Cells were washed with PBS. Then collected and ultrasonically break. After centrifuge, the supernatant was collected and used for MDA detection. MDA level was measured following

the kit instructions. and the SpectraMax M5 microplate reader (Molecular Devices, LLC, Sunnyvale, CA, USA) was used to record the absorbance at 532 nm.

2.15 Statistical analysis The quantitative data were all expressed as means ± SD, and statistical analysis was performed using GraphPad Prism 5. The significance between two groups was analyzed by Student’s t-test, the comparisons of three or more groups were made by one-way ANOVA with Bonferroni test, as appropriate. (*p < 0.05, **p < 0.01, and ***p < 0.001). All experiments were repeated at least three times independently.

3. Results 3.1 High glucose induces ARPE-19 cell senescence but not apoptosis Cell viability was examined at differential concentrations (5 or 25 mM) respectively after incubation for 24 h or 48 h. And a slight change on cell viability was observed by treating 25 mM glucose for 48 h (Fig. 1A). However, the result from Annexin-V/FITC double staining showed that ARPE-19 cells incubated with 25 mM glucose for 24 h or 48 h did not increase apoptotic cell death (Fig. 1B). Meanwhile, the expression of cleaved-Caspase-3, the marker of typical apoptosis, was found to have not changed (Fig. 1C). Supportively, the nuclear chromatin condensation and shrinkage were not observed in ARPE-19 treated with 5 or 25 mM glucose(Fig.1D). The results suggest that high glucose did not cause apoptosis of ARPE-19 in 48 h, which was consistent with reports from animal models of diabetes that the favorable evidence for apoptosis of retinal cells in mice is not detected during early diabetes[10, 11]. We found that the number of SA-β-Gal-positive cells significantly increased in a time-dependent manner (Fig.1E), indicating that ARPE-19 might be senescent when treated with 25 mM glucose for 48 h. Similarly, incubating ARPE-19 with 25 mM glucose resulted in increased expression of senescence-associated protein p21, the member of the cyclin-dependent kinase inhibitor (CKI) family (Fig.1F). High glucose also triggered the cell cycle arrest in ARPE-19, and the cells in G0/G1 phase gradually increased from 63.91% to 78.42% after incubation of ARPE-19 with 25 mM high glucose for 48 h. These results indicate that not apoptosis, but premature senescence is induced in ARPE-19 incubated with high concentration of glucose.

Fig. 1. High glucose induces ARPE-19 premature senescence rather than apoptosis. ARPE-19 were treated with different concentrations (5 or 25mM) of glucose after 24 or 48 h treatment. (A) Cell viability was detected by MTT assay. (B) Apoptosis detection by flow cytometry. (C) Expression of Caspase-3 measured by Western blot analysis. β-actin were used as internal standards. (D) The representative images of nucleus morphological changes by Hoechst stain. (E) The representative images of SA-β-Gal-positive cells. (F) Expression of p21 measured by Western blot analysis. β-actin were used as internal standards. (G)

Cell cycle distribution was analyzed by flow cytometry. Scale bar = 20 μm. Values from at least three independent experiments are shown as means ± SD. *p < 0.05, **p < 0.01 and ***p < 0.001.

3.2 Oxidant species mediates senescence of ARPE-19 induced by high glucose High concentrations of glucose，the characteristic of diabetes, can lead to excessive production of ROS [23]. Moreover, the generation of ROS in the cellular microenvironment causes stress-induced premature senescence[30]. Consistently, in ARPE-19, 25 mM glucose resulted in increasing intracellular H2O2 with time-dependence (Fig.2A). To further validate the role of oxidant species in high glucose-induced cellular senescence, we used the antioxidant N-acetylcysteine (NAC). As expected, H2O2 (Fig.2A) and senescence (Fig.2B) induced by high glucose were alleviated after co-incubating with NAC. ROS signaling activated p38, which was involved in the regulation of senescence. We found that the phosphorylation of p38 did increase (Fig.2F). Previous study reported that mitochondrial Ca2+ accumulation leads to a subsequent decrease in mitochondrial membrane potential (MMP), ROS accumulation and senescence[31]. Analogously, we found an increase in mitochondrial superoxide by using MitoSOX (Fig.2C,D). The MMP reduced in a time-dependent manner (Fig.2E), while the mitochondrial Ca2+ levels increased gradually in our study (Fig.2G). MCU (mitochondrial calcium uniporter) is the key component in mitochondrial Ca 2+ uptake. We

observed that the level of MCU was also elevated in ARPE-19 with 25 mM glucose exposure for 48 h (Fig.2F), which further confirmed that mitochondrial Ca2+ accumulated in senescent ARPE-19 cells induced by high glucose. Similarly, cytoplasmic Ca2+ level was elevated due to high glucose exposure (Fig.2H), which was consistent to the research conducted by Nada Abuarab[32]. These results indicate that high glucose increased intracellular Ca2+ levels, which are in turn transferred to mitochondria via up-regulated MCU, thereby induced MMP depolarization and subsequent oxidant species elevation, and ultimately led to ARPE-19 cell senescence.

Fig. 2 Oxidant species mediates Senescence of ARPE-19 induced by high glucose. (A) Intracellular H2O2 level was measured by DCFDA in ARPE-19 treated with 25 mM glucose alone for 0, 12, 24 or 48 h, or combined with NAC (500 μM) for 48h. (B) SA-β-Gal staining was used to detect the senescent state in ARPE-19 treated with 5 or 25 mM glucose alone, or combined with NAC (500 μM) for 48h. (C,D) Mitochondrial superoxide was measured by MitoSOX in ARPE-19 treated with 25 mM glucose alone for 0, 12, 24 or 48 h, or combined with NAC (500 μM) for 48 h. (E) The representative images of mitochondrial membrane potential determined by using JC-1 stain in ARPE-19 treated with 25 mM glucose for 0, 24 or 48 h. The cells with green-positive and red-negative fluorescence counted as depolarized cells. (F) Expression of MCU or p-p38 measured by Western blot analysis in ARPE-19 treated with 25 mM glucose for 0, 24 or 48 h. (G-H) The typical fluorescence microscopy images of Ca2+ fluorescence intensity stained by Fluo-4AM (G) or Fluo-2AM (H) in ARPE-19 treated with 25 mM glucose for 0, 24 or 48 h. Scale bar = 20 μm. Values from at least three independent experiments are shown as means ± SD. *p < 0.05, **p < 0.01 and ***p < 0.001.

3.3 Lipid oxidation mediates the accumulation of H2O2 Interestingly, accumulated lipids were detected by Oil red O staining in our study (Fig 3A). Mikolaj at el. reported that hepatocyte senescence promotes fat accumulation in vitro and in vivo by losing the ability to

metabolize fatty acids efficiently[33]. But the expression of carnitine palmitoyl transferase I (CPT1), the key enzyme of mitochondrial fatty acid oxidation, was up-regulated by 25 mM glucose treatment (Fig.3B), which supported that the cause of lipid accumulation within 48h of high glucose treatment was not metabolic slowdown. Considering that glucose could regulate fatty acid synthase (FAS), the key enzymes of fatty acid synthesis, we evaluated FAS level to explore whether the increase in fatty acid synthesis would be responsible for lipid accumulation. As expected, the expression of FAS obviously increased (Fig.3B), indicating that accumulation of intracellular lipids was caused by increased synthesis of fatty acids in ARPE-19 under high concentration of glucose. Tsushima et al. demonstrated that overload lipid induced mitochondrial ROS generation, which caused mitochondrial dysfunction, promoting further ROS leakage[34-36]. Therefore, we hypothesized that high glucose could also generate oxidant species through lipid oxidation to further enhance cell senescence in ARPE-19. Accordingly, we added free fatty acids (FFAs; free fatty acid mixture(Palmitic acid: oleate at 1:2) conjugated with bovine serum albumin (BSA) at 6:1) to aggravate the cellular lipid accumulation in ARPE-19 treated with 25 mM glucose to further verify the role of lipids in senescent cells induced by high glucose (Fig.3C). Indeed, the number of SA-β-gal-positive cells and the expression of p21 were significantly increased by the additional FFAs (Fig.3D,E). We found that high glucose increased the lipid oxidation product malondialdehyde (MDA) (Fig.3F), and that MDA was further increased by the additional FFAs, so as to the expression of peroxisome proliferator-activated receptor alpha(PPARα), a central regulator of β-oxidation, and its downstream CPT1 (the key enzyme of mitochondrial fatty acid oxidation) and peroxisomal acyl-coenzyme A oxidase 1 (ACOX1, the key enzyme of fatty acid oxidation in peroxisome) (Fig.3G). Moreover, the production of intracellular H2O2 induced by high glucose were significantly further increased with the addition of FFAs (Fig.3H), as well as the result of p-p38 level (Fig.3E). The results further validate our hypothesis that lipid oxidation mediates senescence in ARPE-19 incubated with high glucose by generating oxidant species.

Fig. 3 Lipid oxidation mediates the accumulation of H2O2. (A,B) ARPE-19 was treated with different concentrations of glucose (5 or 25mM) for 24 or 48 h. (A) The representative images of Oil Red O staining. (B) Expression of CPT1 and FAS measured by Western blot analysis.

(C-G) ARPE-19 were treated with 5 mM glucose or 25 mM glucose alone or combined

with FFAs (100 μM). (C) The representative images of Oil Red O staining. (D) The representative images of SA-β-Gal staining. (E) Expression of p21 or p-p38 measured by Western blot analysis. (F) MDA levels. (G) Expression of PPARα, CPT1 and ACOX measured by Western blot analysis. (H) Intercellular H2O2 level measured by DCFDA. Scale bar = 20 μm. Values from at least three independent experiments are shown as means ± SD. *p < 0.05, **p < 0.01 and ***p < 0.001.

3.4 High glucose induced senescence through fatty acid synthesis via the PI3K/AKT/mTOR pathway The activity of the PI3K/AKT/mTOR pathway is tightly controlled under normal physiological conditions but activated in senescent cells[37]. Another study reported that fatty acid synthesis is also regulated by PI3K/AKT/mTOR signaling[38]. However, the link between both cell processes has not been reported. We found the up-regulated expression of PI3K, p-AKT and p-mTOR in ARPE-19 treated with 25 mM glucose (Fig.4A). Next, we employed ARPE-19 with 740Y-P (the PI3K/AKT activator), LY294002 (the PI3K/AKT inhibitor) or rapamycin (the mTOR inhibitor) to further verify the role of PI3K/AKT/mTOR pathway. High glucose increased FAS expression, which was further elevated by 740Y-P but reversed by treatment with LY294002 or rapamycin (Fig.4A), supporting that PI3K/AKT/mTOR pathway mediates fatty acid synthesis. Consistently, the number of lipid droplets and SA-β-gal-positive cells was increased by high glucose, which was further increased by 740Y-P, while reversed by LY294002 or rapamycin (Fig.4B-C). The similar changes were also presented on intracellular H2O2 and the protein expression of p21 and phosphorylated form of p38 (Fig.4D,E). The results indicate that high glucose mediates fatty acid synthesis, oxidant species production and cell senescence through the activation of the PI3K/AKT/mTOR pathway in ARPE-19 cells. And to the best of our knowledge, it is the first time that fatty acid metabolism has been linked to cellular senescence.

Fig. 4 High glucose induces fatty acid synthesis via the PI3K/AKT/mTOR pathway. ARPE-19 were treated with 5 mM glucose or 25 mM glucose alone or combined with LY294002 (1 nM), 740Y-P (6 μM) or rapamycin (100 nM) for 48 h. (A) The expression of PI3K, p-AKT, p-mTOR and FAS. (B) The representative images of Oil Red staining. (C) The representative images of SA-β-Gal staining. (D) The expression of p-p38, p38, and p21. (E) Intercellular H2O2 level measured by DCFDA. Scale bar = 20 μm. Values from at least three independent experiments are shown as means ± SD. *p < 0.05, **p < 0.01 and ***p < 0.001.

3.5 High glucose induced autophagy in ARPE-19 cells Given that autophagy has been reported in relation to cellular senescence[26, 27], we next detected the status of autophagy in ARPE-19 treated with high glucose for 48 h. The results showed that the ratio of LC3II/LC3I increased in a time-dependent manner after high glucose treatment, while the accumulation of p62, one of the selective substrates for autophagy, decreased (Fig.5A). LysoTracker Red staining revealed that lysosomes and autolysosomes structures

also increased over time (Fig.5B). These data suggest that

autophagy was induced in ARPE-19 treated with high glucose for 48 h. Moreover, the result of the additional hydroxychloroquine (HCQ), an autophagy inhibitor that interrupts the binding of autophagosome and lysosome by altering lysosomal pH, further verified the autophagy flux was unobstructed (Fig.5C). Therefore, these results indicate that high glucose activated autophagy in ARPE-19, which is in accordance with previous reports that ARPE-19 undergoes autophagy with high glucose treatment for 48 h, and mTOR signaling may not play a major role in ARPE-19 autophagy[39]. Next, we examined the effect of inhibitors of autophagy on senescence. Indeed, as measured by SA-β-gal staining (Fig.5D) and the expression of p21 (Fig.5E), the autophagy inhibitors bafilomycin A1 and HCQ significantly increased the senescence number induced by high glucose in ARPE-19. The similar changes

were also presented on p-p38 and intercellular H2O2 (Fig.5F). Our results show that the protective autophagy is induced during the senescence process in ARPE-19 within 48 h of high glucose exposure.

Fig. 5 High glucose induces autophagy in ARPE-19 cells. (A,B) ARPE-19 were treated with 5 or 25 mM glucose for 24 or 48 h. (A) The expression of LC3, CB and p62. (B) The representative images of LysoTracker Red DND-99 staining. (C) The expression of LC3, CB and p62 in ARPE-19 treated with 5 mM glucose or 25 mM glucose alone or combined with HCQ (10 μM) for 48 h. (D-F) ARPE-19 were treated with 5 mM glucose or 25 mM glucose alone or combined with HCQ (10 μM), bafilomycin (1 nM). (D) The representative images of SA-β-Gal staining. (E) The expression of p-p38, p38 and p21. (F) Intercellular H2O2 level measured by DCFDA. Scale bar = 20 μm. Values from at least three independent experiments are shown as means ± SD. *p < 0.05, **p < 0.01 and ***p < 0.001.

3.6 Autophagy was inhibited and apoptosis was induced in ARPE-19 under long-term high glucose exposure We next investigated whether the mechanism of RPE cell damaged under prolonged high glucose exposure was the same as that of short-term high-glucose exposure. ARPE-19 cells were treated with 25 mM glucose for 7 d or even 14 d. Consistently, whether treated for 7 d or 14 d, the cells treated with high glucose still showed an increase in the number of SA-β-gal-positive cells, increased lipid droplets and accumulated ROS (Fig.6A-C). But surprisingly, the accumulated p62 was observed, suggesting that autophagy may be damaged (Fig.6D). We also found that the ratio of LC3 II to LC3 I was higher in the 7th day of high glucose treatment than that in the control group, but significantly lower at the 14th day (Fig.6D), indicating that high glucose induced autophagy in a short time, but then the autophagic flow was blocked, and finally autophagosome formation was inhibited. Meanwhile, the level of cleaved-Caspase-3 was significantly raised

during 14 d of high glucose treatment (Fig.6E), which was in agreement with study that autophagy inhibition could switch stress-induced senescence to apoptosis[28]. Collectively, these results indicate that the state of autophagy in ARPE-19 cells from activation turns to inhibition with prolonged high glucose exposure, which turns cell fate to become apoptotic.

Fig. 6 Long term high glucose exposure induces apoptosis, accompanying with autophagy inhibition in ARPE-19. ARPE-19 were treated with different concentrations (5 or 25mM) of glucose for 7 or 14 d. (A) Cell senescence measured by SA-β-Gal staining. (B) Lipid droplets detected by Oil Red O staining. (C) Intercellular H2O2 level in ARPE-19 measured by flow cytometry. (D) Western blot analysis of p62 and LC3 expression. β-actin were used as internal standards. (E) Western blot analysis of Caspase-3 expression. β-actin were used as internal standards. Scale bar = 20 μm. Values from at least three independent experiments are shown as means ± SD. *p < 0.05, **p < 0.01 and ***p < 0.001.

4. Discussion DR is the most common complication of diabetes, and represents a significant socioeconomic cost for healthcare systems worldwide. It is well established that RPE damage is highly correlated with the incidence of DR. But the molecular mechanism behind high glucose-induced RPE cell damage remains unclear. Here, we found that RPE was senescent due to increased ROS production under high glucose treatment. High glucose also induced lipid accumulation and lipid oxidation in RPE cells, which contributed to excessive accumulation of ROS and subsequent stress-induced premature senescence. Antioxidant NAC relieves

senescence through scavenging ROS. Meanwhile, PI3K/AKT/mTOR pathway was involved in high glucose-induced senescence by regulating fatty acid synthesis leading to lipid accumulation. Additionally, early exposure to high glucose increased autophagy to resist oxidative stress, while the long-term treatment of high glucose inhibited autophagy and increased apoptotic cells. It is widely accepted that ROS constitutes the primary cause for cellular senescence[4]. ROS also plays a significant role in the development and progression of diabetes and its complications as reported [40]. Our study proposed that high glucose could induce RPE senescence through elevating ROS accumulation. The p38 signaling, which is activated in senescent cells in response to ROS to regulate cells growth, proliferation, death and differentiation as reported[41], was also found to be increased in high glucose-treated RPE cells consistent with the results of the senescence-associated protein p21. Mitochondria are the main source of intracellular ROS as the electron transport consumes about 85% of the oxygen that the cell uses[42]. And during biological aging, aberrations in the control of cell death, impaired metabolism and perturbed redox homeostasis have been attributed to mitochondrial dysfunction[43]. As described before[31, 44, 45], mitochondrial calcium accumulation and mitochondrial depolarization led to mitochondrial dysfunction, resulting in the generation of ROS. Consistently, we observed that calcium release from ER to mitochondria via MCU was involved in glucose-elicited RPE cell senescence, which could be partially reversed by NAC. Glucose could regulate the key enzymes of lipid acid synthesis. Our findings verified that 25 mM glucose could increase lipid droplets levels through promoting fatty acid synthesis by FAS in RPE, in agreement of the studies that sustained hyperglycemia in diabetes promotes fatty acid synthesis and TG accumulation to cause ectopic lipid accumulation in nonadipose tissues[46]. And this process, termed lipotoxicity, takes a major role in the pathogenesis of diabetic complications[47, 48]. Dysregulation of lipid metabolism may disrupt normal cell processes. Studies have reported that altered lipid metabolism is observed during age-related diseases[49-51]. However, whether such changes are a cause or a consequence of aging are not fully understood. Mikolaj at el. reported that senescent hepatocyte promoted fat accumulation due to the loss of the ability of fatty acid metabolism. However, the expression of CPT1 was up-regulated by high glucose treatment, indicating that the cause of lipid accumulation was not metabolic slowdown caused by senescence. And after adding additional FFAs to increase intracellular lipid accumulation, we detected an increase in cellular senescence. Overall, it is clear that the increased lipid under high glucose treatment resulted in RPE senescence. Additionally, a common feature of β-oxidation whether in peroxisomes or mitochondria is the potential to generate ROS[14]. Yet, the ROS-producer mitochondria is particularly susceptible to locally mediated oxidative damage to lipids, and oxidative damage from lipid peroxidation products may promote further ROS leakage in a vicious cycle of oxidative damage[34, 35]. The key enzymes involved in fatty acid oxidation, such as CPT1, a key enzyme of fatty acid β oxidation in mitochondria, and ACOX, a key enzyme of β-oxidation in peroxisomes, are influenced by PPARα, which were all detected to be up-regulated in sync with ROS in RPE treated with high glucose alone or co-treated with FFAs in our study, further supporting the imbalance of lipid metabolism is responsible for the increased ROS and subsequent senescence we detected. PI3K and Akt are signaling kinases involved in cell survival and proliferation. Astle et al. demonstrated that activation of PI3K/AKT signaling induced cell senescence in dependent of mTOR[52]. Correia-Melo et al. also demonstrated that the Akt and mTORC1 phosphorylation cascade contributed to ROS-mediated cellular

senescence in vitro and in vivo[53]. Nevertheless, another study reported that PI3K/Akt/mTOR signaling activated the transcription of major lipogenic genes FAS and ACC to promote de novo lipogenesis[38]. Our study links cellular senescence with fatty acid synthesis through mediating PI3K/AKT/mTOR, that high glucose-activated PI3K/Akt/mTOR signaling induces intracellular lipid accumulation by promoting fatty acid synthesis in RPE, thereby increasing fatty acids oxidation and subsequent senescence. And to the best of our knowledge, this is the first time to link fatty acid synthesis with cellular senescence. Autophagy is activated in response to many stress stimuli that are constantly present, including oxidative stress, in RPE. However, impaired autophagy or lysosomal degradation in the RPE is reported to lead to an accumulation of damaged organelles, toxic proteins including lipofuscin, and extracellular drusen deposits, all of which can contribute to RPE dysfunction or death[54]. And it has been extensively reported that autophagy is involved in cell senescence, although the functional relationship is still inconclusive. In present study, we found that 25 mM glucose activated autophagy within 48 h. Furthermore, the result of autophagy inhibition by HCQ or bafilomycin showed that high glucose-induced autophagy is protective. Notably, in contrast to some reports that autophagy was negatively regulated by mTOR[55], the PI3K/AKT/mTOR signaling was activated in our study as well as autophagy. A possible explanation was that mTOR signaling is not a major signaling involved in RPE autophagy as reported by Ale et al.[39]. Furthermore, we examined the status of autophagy under long-term exposure (7 or 14 d) to high glucose in RPE. Surprisingly, inhibited autophagy and increased apoptosis were observed under high glucose exposure during 14 d, which was in line with previous report that autophagy inhibition could switch stress-induced senescence to apoptosis[28]. Thus, it is likely that changes in autophagy begin as a consequence of oxidative damage, but later the autophagic process becomes impaired and thus contributes to the progression of RPE degeneration[54]. Collectively, the generation of oxidant species increased through altering lipid metabolism induced by 25 mM glucose, and then resulted in oxidative stress-induced senescence of RPE, which eventually participated in the development of DR. Furthermore, high glucose in short-term exposure was able to increase autophagy to protect RPE from oxidative stress-induced damage, while long-term high glucose treatment reduced autophagy and increased apoptotic cells. Our study is the first to link lipid metabolism to oxidative stress-induced premature senescence in RPE of DR and provides new insights to help understand the mechanisms of high glucose-induced RPE damage. Overall, our findings would provide more understandings and treatment strategies for DR.

Acknowledgments We thank Bin Zhou for using ChemiDoc™ - XRS imaging system, Jingyao Zhang and Fulai Xue for using fluorescence microplate reader.

Sources of funding This work was supported by the National Natural Science Foundation of China (No: 81673710 and No: 81803866) awarded to Wen Huang and Guang Xin respectively.

Conflict of interest statement No conflicts of interest to declare.

References [1] J.W. Yau, S.L. Rogers, R. Kawasaki, E.L. Lamoureux, J.W. Kowalski, T. Bek, S.J. Chen, J.M. Dekker, A. Fletcher, J. Grauslund, Global prevalence and major risk factors of diabetic retinopathy, Diabetes Care, 35 (2012) 556-564. [2] R. Simó, M. Villarroel, L. Corraliza, C. Hernández, M. Garciaramírez, The retinal pigment epithelium: something more than a constituent of the blood-retinal barrier--implications for the pathogenesis of diabetic retinopathy, J Biomed Biotechnol., 2010 (2014) 190724. [3] F.Q. Liang, B.F. Godley, Oxidative stress-induced mitochondrial DNA damage in human retinal pigment epithelial cells: a possible mechanism for RPE aging and age-related macular degeneration, Exp. Eye Res., 76 (2003) 397-403. [4] D. Harman, Aging: a theory based on free radical and radiation chemistry, J Gerontol, 11 (1956) 298-300. [5] R.S. Balaban, S. Nemoto, T. Finkel, Mitochondria, oxidants, and aging, Cell, 120 (2005) 483-495. [6] T. Yokoi, K. Fukuo, O. Yasuda, M. Hotta, J. Miyazaki, Y. Takemura, H. Kawamoto, H. Ichijo, T. Ogihara, Apoptosis signal-regulating kinase 1 mediates cellular senescence induced by high glucose in endothelial cells, Diabetes, 55 (2006) 1660-1665. [7] X. Zhang, X. Chen, D. Wu, W. Liu, J. Wang, Z. Feng, G. Cai, B. Fu, Q. Hong, J. Du, Downregulation of connexin 43 expression by high glucose induces senescence in glomerular mesangial cells, J Am Soc Nephrol., 17 (2006) 1532-1542.. [8] K. Chen, H. Dai, J. Yuan, J. Chen, L. Lin, W. Zhang, L. Wang, J. Zhang, K. Li, Y. He, Optineurin-mediated mitophagy protects renal tubular epithelial cells against accelerated senescence in diabetic nephropathy, Cell Death Dis., 9 (2018) 105-123. [9] C. Zhao, D. Yasumura, X. Li, M. Matthes, M. Lloyd, G. Nielsen, K. Ahern, M. Snyder, D. Bok, J.L. Dunaief, mTOR-mediated dedifferentiation of the retinal pigment epithelium initiates photoreceptor degeneration in mice, J Clin Invest., 121 (2011) 369-383. [10] I. Piano, E. Novelli, S.L. Della, E. Strettoi, L. Cervetto, C. Gargini, Involvement of autophagic pathway in the progression of retinal degeneration in a mouse model of diabetes, Front Cell Neurosci., 10 (2016) 42. [11] É. A, A. Szabó, O. Kántor, C. Dávid, P. Szalay, K. Szabó, Á. Szél, J. Németh, Á. Lukáts, Pathologic alterations of the outer retina in streptozotocin-induced diabetes, Invest Ophthalmol Vis Sci., 55 (2014) 3686-3699. [12] L.F. Dmitriev, V.N. Titov, Lipid peroxidation in relation to ageing and the role of endogenous aldehydes in diabetes and other age-related diseases, Ageing Res Rev., 9 (2010) 200-210. [13] J.A. van Diepen, J.F. Berbée, L.M. Havekes, P.C. Rensen, Interactions between inflammation and lipid metabolism: relevance for efficacy of anti-inflammatory drugs in the treatment of atherosclerosis, Atherosclerosis, 228 (2013) 306-315. [14] A.C. Flor, D. Wolfgeher, D. Wu, S.J. Kron, A signature of enhanced lipid metabolism, lipid peroxidation

and aldehyde stress in therapy-induced senescence, Cell Death Discov., 3 (2017) 17075. [15] J. St-Pierre, J.A. Buckingham, S.J. Roebuck, M.D. Brand, Topology of superoxide production from different sites in the mitochondrial electron transport chain, J Biol Chem., 277 (2002) 44784-44790. [16] E.B. Tahara, F.D. Navarete, A.J. Kowaltowski, Tissue-, substrate-, and site-specific characteristics of mitochondrial reactive oxygen species generation, Free Radic Biol Med, 46 (2009) 1283-1297. [17] J.K. Reddy, T. Hashimoto, Peroxisomal beta-oxidation and peroxisome proliferator-activated receptor alpha: an adaptive metabolic system, Annu Rev Nutr., 21 (2001) 193-230. [18] C. Quijano, M. Trujillo, L. Castro, A. Trostchansky, Interplay between oxidant species and energy metabolism, Redox Biol., 8 (2016) 28-42. [19] J. Hao, S. Liu, Z. Song, Q. Liu, L. Xin, H. Chen, Y. Niu, H. Duan, PI3K/Akt pathway mediates high glucose-induced lipogenesis and extracellular matrix accumulation in HKC cells through regulation of SREBP-1 and TGF-β1, Histochem Cell Biol., 135 (2011) 173-181. [20] Y. Zhang, H. Takemori, C. Wang, J. Fu, M. Xu, L. Xiong, N. Li, X. Wen, Role of salt inducible kinase 1 in high glucose-induced lipid accumulation in HepG2 cells and metformin intervention, Life Sci., 173 (2017) 107-115. [21] S. Zhao, L. Zhu, H. Duan, S. Liu, Q. Liu, W. Liu, J. Hao, PI3K/Akt pathway mediates high glucose‐ induced lipid accumulation in human renal proximal tubular cells via spliced XBP‐ 1, J Cell Biochem., 113 (2012) 3288-3298. [22] Y. He, J. Yang, H. Li, H. Shao, C. Wei, Y. Wang, M. Li, C. Xu, Exogenous spermine ameliorates high glucose-induced cardiomyocytic apoptosis via decreasing reactive oxygen species accumulation through inhibiting p38/JNK and JAK2 pathways, Int J Clin Exp Pathol., 8 (2015) 15537. [23] L. Donato, C. Scimone, C. Rinaldi, P. Aragona, S. Briuglia, A. D'Ascola, R. D'Angelo, A. Sidoti, Stargardt phenotype associated with two ELOVL4 promoter variants and ELOVL4 downregulation: new possible perspective to etiopathogenesis?, Invest Ophthalmol Vis Sci., 59 (2018) 843-857. [24] L. Donato, P. Bramanti, C. Scimone, C. Rinaldi, R. D'Angelo, A. Sidoti, miRNA expression profile of retinal pigment epithelial cells under oxidative stress conditions, FEBS open bio., 8 (2018) 219-233. [25] A. Salminen, K. Kaarniranta, Regulation of the aging process by autophagy, Trends Mol Med., 15 (2009) 217-224. [26] D. Dutta, R. Calvani, R. Bernabei, C. Leeuwenburgh, E. Marzetti, Contribution of impaired mitochondrial autophagy to cardiac aging: mechanisms and therapeutic opportunities, Circ Res., 110 (2012) 1125-1138. [27] H. Tai, W. Zhe, G. Hui, X. Han, Z. Jiao, X. Wang, X. Wei, D. Yi, H. Ning, J. Qin, Autophagy impairment with lysosomal and mitochondrial dysfunction is an important characteristic of oxidative stress-induced senescence, Autophagy, 13 (2016) 99-113. [28] J.W. Zhang, S.S. Zhang, J.R. Song, K. Sun, C. Zong, Q.D. Zhao, W.T. Liu, R. Li, M.C. Wu, L.X. Wei, Autophagy inhibition switches low-dose camptothecin-induced premature senescence to apoptosis in human

colorectal cancer cells, Biochem Pharmacol., 90 (2014) 265-275. [29] R. Gonnella, S. Yadav, M.S. Gilardini Montani, M. Granato, R. Santarelli, A. Garufi, G. D'Orazi, A. Faggioni, M. Cirone, Oxidant species are involved in T/B-mediated ERK1/2 phosphorylation that activates p53-p21 axis to promote KSHV lytic cycle in PEL cells, Free Radic Biol Med., 112 (2017) 327-335. [30] G. Salazar, J. Huang, R.G. Feresin, Y. Zhao, K.K. Griendling, Zinc regulates Nox1 expression through a NF-κB and mitochondrial ROS dependent mechanism to induce senescence of vascular smooth muscle cells, Free Radic Biol Med., 108 (2017) 225-235. [31] C. Wiel, H. Lalletdaher, D. Gitenay, B. Gras, B.L. Calvé, A. Augert, M. Ferrand, N. Prevarskaya, H. Simonnet, D. Vindrieux, Endoplasmic reticulum calcium release through ITPR2 channels leads to mitochondrial calcium accumulation and senescence, Nat Commun., 5 (2014) 3792. [32] N. Abuarab, T.S. Munsey, L.H. Jiang, J. Li, A. Sivaprasadarao, High glucose-induced ROS activates TRPM2 to trigger lysosomal membrane permeabilization and Zn2+-mediated mitochondrial fission, Sci Signal., 10 (2017) eaal4161. [33] M. Ogrodnik, S. Miwa, T. Tchkonia, D. Tiniakos, C.L. Wilson, A. Lahat, C.P. Day, A. Burt, A. Palmer, Q.M. Anstee, Cellular senescence drives age-dependent hepatic steatosis, Nat Commun., 8 (2017) 15691. [34] J. Yao, R.D. Brinton, Targeting mitochondrial bioenergetics for alzheimer’s prevention and treatment, Curr Pharm Des., 17 (2011) 3474. [35] O.S. Ademowo, H.K.I. Dias, D.G.A. Burton, H.R. Griffiths, Lipid (per) oxidation in mitochondria: an emerging target in the ageing process?, Biogerontology, 18 (2017) 1-21. [36] K. Tsushima, H. Bugger, A.R. Wende, J. Soto, G.A. Jenson, A.R. Tor, R. McGlauflin, H.C. Kenny, Y. Zhang, R. Souvenir, X.X. Hu, C.L. Sloan, R.O. Pereira, V.A. Lira, K.W. Spitzer, T.L. Sharp, K.I. Shoghi, G.C. Sparagna, E.A. Rog-Zielinska, P. Kohl, O. Khalimonchuk, J.E. Schaffer, E.D. Abel, Mitochondrial reactive oxygen species in lipotoxic hearts induce post-translational modifications of AKAP121, DRP1, and OPA1 that promote mitochondrial fission, Circ Res., 122 (2018) 58-73. [37] Y. Xu, N. Li, R. Xiang, P. Sun, Emerging roles of the p38 MAPK and PI3K/AKT/mTOR pathways in oncogene-induced senescence, Trends Biochem Sci., 39 (2014) 268-276. [38] J. Li, Q. Huang, X. Long, J. Zhang, X. Huang, J. Aa, H. Yang, Z. Chen, J. Xing, CD147 reprograms fatty acid metabolism in hepatocellular carcinoma cells through Akt/mTOR/SREBP1c and P38/PPARalpha pathways, J Hepatol., 63 (2015) 1378-1389. [39] J. Yao, Z.F. Tao, C.P. Li, X.M. Li, G.F. Cao, Q. Jiang, B. Yan, Regulation of autophagy by high glucose in human retinal pigment epithelium, Cell Physiol Biochem., 33 (2014) 107-116. [40] R.A. Defronzo, Glucose intolerance and aging, Diabetes Care, 4 (1981) 493-501. [41] H. Jung, O.K. Dong, J.E. Byun, W.S. Kim, J.K. Mi, H.Y. Song, Y.K. Kim, D.K. Kang, Y.J. Park, T.D. Kim, Thioredoxin-interacting protein regulates haematopoietic stem cell ageing and rejuvenation by inhibiting p38 kinase activity, Nat Commun., 7 (2016) 13674. [42] G. Vitale, S. Salvioli, C. Franceschi, Oxidative stress and the ageing endocrine system, Nat Rev

Endocrinol., 9 (2013) 228-240. [43] F. Yin, H. Sancheti, Z. Liu, E. Cadenas, Mitochondrial function in ageing: coordination with signalling and transcriptional pathways, J Physiol., 594 (2016) 2025-2042. [44] C. Leikam, A. Hufnagel, M. Schartl, S. Meierjohann, Oncogene activation in melanocytes links reactive oxygen to multinucleated phenotype and senescence, Oncogene, 27 (2008) 7070-7082. [45] P. Rai, T.T. Onder, J.J. Young, J.L. Mcfaline, B. Pang, P.C. Dedon, R.A. Weinberg, Continuous elimination of oxidized nucleotides is necessary to prevent rapid onset of cellular senescence, Proc Natl Acad Sci USA., 106 (2009) 169-174. [46] M. Herman-Edelstein, P. Scherzer, A. Tobar, M. Levi, U. Gafter, Altered renal lipid metabolism and renal lipid accumulation in human diabetic nephropathy, J Lipid Res., 55 (2014) 561-572. [47] R. Wu, H.C. Chang, A. Khechaduri, K. Chawla, M. Tran, X. Chai, C. Wagg, M. Ghanefar, X. Jiang, M. Bayeva, Cardiac-specific ablation of ARNT leads to lipotoxicity and cardiomyopathy, J Clin Invest., 124 (2014) 4795-4806. [48] A. Mehlem, C.E. Hagberg, L. Muhl, U. Eriksson, A. Falkevall, Imaging of neutral lipids by oil red O for analyzing the metabolic status in health and disease, Nat Protoc., 8 (2013) 1149-1154. [49] H. Abouassi, M.A. Connelly, L.A. Bateman, K.N. Tune, J.L. Huebner, V.B. Kraus, D.A. Winegar, J.D. Otvos, W.E. Kraus, K.M. Huffman, Does a lack of physical activity explain the rheumatoid arthritis lipid profile?, Lipids Health Dis., 16 (2017) 39. [50] J.H. Hong, J.W. Kang, D.K. Kim, S.H. Baik, K.H. Kim, S.R. Shanta, J.H. Jung, I. Mook-Jung, K.P. Kim, Global changes of phospholipids identified by MALDI imaging mass spectrometry in a mouse model of Alzheimer's disease, J Lipid Res., 57 (2015) 36-45. [51] C. Pararasa, C.J. Bailey, H.R. Griffiths, Ageing, adipose tissue, fatty acids and inflammation, Biogerontology, 16 (2015) 235-248. [52] M.V. Astle, K.M. Hannan, P.Y. Ng, R.S. Lee, A.J. George, A.K. Hsu, Y. Haupt, R.D. Hannan, R.B. Pearson, AKT induces senescence in human cells via mTORC1 and p53 in the absence of DNA damage: implications for targeting mTOR during malignancy, Oncogene, 31 (2012) 1949-1962. [53] C. Correiamelo, F.D. Marques, R. Anderson, G. Hewitt, R. Hewitt, J. Cole, B.M. Carroll, S. Miwa, J. Birch, A. Merz, Mitochondria are required for pro‐ ageing features of the senescent phenotype, Embo Journal., 35 (2016) 724-742. [54] K. Kai, S. Debasish, B. Janusz, K. Anu, V. Zoltán, S. Antero, M.E. Boulton, P. Goran, Autophagy and autophagy dysregulation leads to retinal pigment epithelium dysfunction and development of age-related macular degeneration, Autophagy, 9 (2013) 973-984. [55] J.M. Rosenbluth, J.A. Pietenpol, mTOR regulates autophagy-associated genes downstream of p73, Autophagy, 5 (2009) 114-116.

Highlights

1. Early high glucose exposure induces RPE cell senescence rather than apoptosis.

2. Fatty acid synthesis induced by high glucose through PI3K/AKT/mTOR pathway leads to lipid accumulation.

3. Lipid oxidation contributes to high glucose-induced RPE senescence by accumulating oxidant species.

4. Early high glucose treatment induces protective autophagy, while prolonged high glucose exposure inhibits autophagy.