Humanin G (HNG) protects age-related macular degeneration (AMD) transmitochondrial ARPE-19 cybrids from mitochondrial and cellular damage

Age-related macular degeneration (AMD) ranks third among the leading causes of visual impairment with a blindness prevalence rate of 8.7%. Despite several treatment regimens, such as anti-angiogenic drugs, laser therapy, and vitamin supplementation, being available for wet AMD, to date there are no FDA-approved therapies for dry AMD. Substantial evidence implicates mitochondrial damage and retinal pigment epithelium (RPE) cell death in the pathogenesis of AMD. However, the effects of AMD mitochondria and Humanin G (HNG), a more potent variant of the mitochondrial-derived peptide (MDP) Humanin, on retinal cell survival have not been elucidated. In this study, we characterized mitochondrial and cellular damage in transmitochondrial cybrid cell lines that contain identical nuclei but possess mitochondria from either AMD or age-matched normal (Older-normal (NL)) subjects. AMD cybrids showed (1) reduced levels of cell viability, lower mtDNA copy numbers, and downregulation of mitochondrial replication/transcription genes and antioxidant enzyme genes; and (2) elevated levels of genes related to apoptosis, autophagy and ER-stress along with increased mtDNA fragmentation and higher susceptibility to amyloid-β-induced toxicity compared to NL cybrids. In AMD cybrids, HNG protected the AMD mitochondria, reduced pro-apoptosis gene and protein levels, upregulated gp130 (a component of the HN receptor complex), and increased the protection against amyloid-β-induced damage. In summary, in cybrids, damaged AMD mitochondria mediate cell death that can be reversed by HNG treatment. Our results also provide evidence of Humanin playing a pivotal role in protecting cells with AMD mitochondria. In the future, it may be possible that AMD patient’s blood samples containing damaged mitochondria may be useful as biomarkers for this condition. In conclusion, HNG may be a potential therapeutic target for treatment of dry AMD, a debilitating eye disease that currently has no available treatment. Further studies are needed to establish HNG as a viable mitochondria-targeting therapy for dry AMD.

the Humanin peptide have a specific function. Serine at position 14 confers neuroprotection, 25 but its substitution with glycine generates a variant called Humanin G (HNG) that is 1000-fold more potent than its parent analog Humanin. 16 HNG is known to protect against cell death by preventing mitochondrial dysfunction. 26 Numerous studies highlight the cytoprotective role of HNG against Alzheimer's disease and cancer, 26,27 but none have shed light on the role of HNG in protecting AMD mitochondria or AMD ARPE-19 transmitochondrial cybrid cells. Our cybrids were created by fusing human ARPE-19 Rho0 (lacking mtDNA) cells with mitochondria from either AMD or agematched normal subjects. Previously we have shown that despite all cell lines having identical nuclei, cybrids with AMD mitochondria express significantly different levels of complement RNA and proteins compared to normal cybrids. 28 These changes can be attributed to variations in mtDNA, since all cybrids had similar nuclear content and were created using the same Rho0 ARPE-19 cells. Furthermore, our previous studies have shown that mtDNA variants in cybrids can also affect the expression of inflammation, angiogenesis, and signaling genes. 29 The present study demonstrates that cybrids containing AMD mitochondria have increased mtDNA fragmentation, impaired expression of mitochondrial replication/transcription genes, upregulation of pro-apoptotic genes and proteins, along with higher ROS levels and diminished cell viability compared to normal cybrids. Exposure to HNG reverses these events and protects the AMD mitochondria, promoting increased cellular longevity. Mechanistically, our results suggest that exogenous HNG acts via both intracellular (BAX) and extracellular (gp130) pathways and that retrograde signaling from AMD mitochondria to RPE cell nuclei contributes significantly to retinal cell death.
AMD cybrids show decreased cellular viability and higher mt ROS production. Cell viability levels of normal and AMD cybrids were measured using both the MTT and Trypan blue dye exclusion assays. Using the MTT assay, AMD cybrids demonstrated significantly decreased cell viability at 24 h (18%), 48 h (17%), and 72 h (18%) compared to normal cybrids (Po0.05). At 96 h, normal and AMD cybrid values were similar to each other (P = 0.98) (Figure 2a). The trypan blue assay results at 72 h showed 11% decrease in cell viability in AMD cybrids compared to normal (P = 0.002, Figure 2b). The mitochondrial superoxide levels, as measured by the MitoSOX assay, were 80% higher in AMD cybrids compared to normal (P = 0.003, Figure 2c) (Supplementary Table S4). These findings demonstrate increased oxidative stress in AMD cybrids and suggest that the AMD mitochondria cannot support cell growth at the rates seen in normal cybrids. AMD cybrids show differential expression of apoptosis, autophagy, ER stress, and antioxidant markers. To determine if AMD mitochondria affect expression of cell death and cell stress markers, qRT-PCR and western blotting were performed. Compared to normal cybrids, the AMD cybrids showed significant upregulation of apoptosis genes (  Table S5). Western blot analyses showed that the AMD cybrids had significantly higher protein levels of Cleaved Caspase-3 (120.4%, Figure 3b Table S6).
Since all cybrids have identical nuclei, these results suggest that AMD mitochondria may mediate cell death/stress by upregulation of nuclear-encoded apoptosis, autophagy, and ER stress genes and downregulation of antioxidant genes.
HNG reduces mtDNA-mediated apoptosis in AMD cybrids. To test the effects of exogenous HNG (3.2 μM) on the numbers of live and apoptotic cells, cybrids were treated with YO-PRO-1 and propidium iodide stains, followed by flow cytometry analyses at 72 h.   Table S4). (c) MitoSOX assay at 72 h demonstrated higher mitochondrial ROS production in AMD cybrids compared to normal cybrids (80%, P = 0.003, n = 5) (Supplementary Table S4). Data are represented as mean ± S.E.M., normalized to normal cybrids (assigned value of 1  Table S7). No differences in either the apoptotic or live cell number were observed between untreated-and HNG-treated normal cybrids. Therefore, the AMD cybrids have higher levels of apoptotic and dead cells compared to cybrids with normal mitochondria. However, treatment with HNG can protect from apoptosis and elevate the levels of viable, healthy cells in AMD cybrids.
HNG modulates the protein levels of BAX, gp130 receptor, and Phospho-JAK2 in AMD cybrids. To assess mechanisms by which HNG protects AMD cybrids, the protein levels of BAX (acts intracellularly), gp130 (a transmembrane glycoprotein and cell surface receptor), and phosphorylated (Phospho)-JAK2 (a signaling molecule and a marker of activated JAK) were compared. As shown in Figure 6a, BAX was significantly elevated (145.9%, P = 0.04) in untreated AMD cybrids compared to untreated normal cybrids. In AMD cybrids, HNG decreased BAX protein levels by 74.42% compared to untreated-AMD cybrids (P = 0.002) (Supplementary Table S9).
HNG prevents loss of AMD mitochondria. Confocal microscopy of AMD cybrids stained with CellLight mtGFP fluorescent probe for 48 h showed higher levels of mitochondria in HNG-treated AMD cybrids (Figures 8a and b). Significant increase in fluorescent intensity (194.3%, Po0.001) was observed in HNG-treated AMD cybrids compared to untreated-AMD cybrids. No difference was observed between untreated-normal and HNG-treated normal cybrids (Supplementary Table S11). This finding indicates that HNG can increase mitochondria-targeted staining in AMD cybrids.

Discussion
In the current study, we used the transmitochondrial ARPE-19 cybrid model to identify and characterize the effects of AMD mitochondria on human retinal cells. Through a combination of cell-based and molecular biology assays, we found that cellular/mitochondrial damage occurred at multiple levels in AMD cybrids, and that HNG protected AMD cybrids from mtDNA mediated and amyloid-β-induced cellular stress.
Previous studies have shown that mtDNA-deficient (Rho0) ARPE-19 cells had diminished mitochondrial membrane potential, altered expression of genes related to lipid transport, inflammation, drusen deposits and various extracellular matrix materials compared to wild-type cells. 30 This indicates that the mitochondria play a role in those particular pathways. The present study used these same Rho0 ARPE-19 cells as the host cell to create the cybrids using mitochondria from either AMD or age-matched normal subjects. While the preparation of Rho0 cells may affect the nuclear (n) DNA, the cell has efficient, rapid nDNA repair mechanisms which the mtDNA lacks. With the introduction of the AMD or normal mitochondria into the uniform Rho0 ARPE-19 cell line, one can assume that alterations in molecular or biochemical behavior seen in the cybrids are due to the variations of the mtDNA rather than nDNA. To test consistency, we have created different cybrid batches using mitochondria from the same subject and the gene expression patterns, cell viability and membrane potentials were similar in the cell lines for the individuals.  Table S8). Data are represented as mean ± S.E.M., normalized to untreated-normal cybrids (assigned value of 1). The endpoint for all experiments was 72 h Therefore, our cybrid model represents a 'personalized' model that is reliable and uniform for each subject and can be used to screen mitochondria-targeting drugs. Furthermore, the additional significant changes in cell-death related genes observed in AMD cybrids could be attributed to unhealthy AMD mitochondria.
AMD pathology is characterized by RPE cell death, which precedes geographic atrophy and leads to eventual deterioration of the overlying photoreceptors. 31 In addition, RPE cell death induced by various stressors such as hydrogen peroxide (tBHP), BMP4, and hypoxia has been demonstrated in in vitro AMD models, [32][33][34] and in several retinal degeneration animal models. 35,36 Consistent with those findings, our results with Trypan blue and MTT cell viability assays demonstrated substantial increase in AMD cybrid cell death at 24h, 48h, and 72 h time points. Since all cybrids contained similar nuclei and differed only in mtDNA, these results indicated that mitochondria from AMD patients were dysfunctional and mediated cell death. RPE cells are highly metabolically active and contain large numbers of mitochondria, which are the primary source of endogenous ROS. 37 Our results revealed that AMD cybrids have: (1) elevated levels of mitochondrial superoxides, the predominant ROS found in the mitochondria, and (2) reduced expression of antioxidant genes, PRDX3 and SOD2. Together these data represent an imbalance between mitochondrial ROS production and antioxidant enzyme levels, which would lead to significantly higher oxidative stress within AMD cybrids compared to the controls. Our results are consistent with previous studies showing that damaged mitochondria and ageing lead to increased ROS and oxidative damage and reduced antioxidant capacity both in in vitro and in vivo models of retinal diseases. 38,39 Analyses of transcript profiles indicated that cybrids with mitochondria from AMD subjects have higher levels of apoptosis, autophagy and ER stress compared to cells possessing normal mitochondria, suggesting that the mechanism of cell death mediated by AMD mitochondria is multifactorial. Our recent study also showed AMD mitochondria altered the expression of complement pathway markers in AMD cybrids. 28 It seems clear that damaged, unhealthy AMD mitochondria transmit retrograde signals to mediate expression patterns of nuclear-encoded genes in cybrids but the exact mechanisms are not yet known. Possible pathways include changes in epigenetics mediated by mtDNA variants, 29,40,41 and mtDNA fragments that upregulate inflammatory genes' expression, via the STING (Stimulator of Interferon Genes) pathway. 42 To cope with perturbations in cellular homeostasis within retinal cells, apoptosis, unfolded protein response due to ER stress, and autophagy are triggered. [43][44][45] Additionally, apoptosis-and ER stressmediated cell death of RPE has been documented in various retinal layers in retinal degenerative diseases, including AMD. 46,47 However, to our knowledge, ours is the first study demonstrating that platelet-derived mitochondria from AMD   Table S10A). (b) Treatment of AMD cybrids with Amy-β Active (1-42) alone reduced cell viability by 41.8% (Po0.05, n = 3-4) compared to untreated-AMD cybrids. No difference in cell viability was observed between untreated-AMD cybrids and Amy-β SC(42-1)-treated AMD cybrids. Significant reduction in cell viability was observed in Amy-β Active (1-42)-treated AMD cybrids (94.9%, Po0.01, n = 3-4) compared to the Amy-β SC(42-1)-treated AMD cybrids. AMD cybrids treated with HNG alone had a 42.4% increase in cell viability compared to untreated-AMD cybrids (Po0.05, n = 3-4). Pre-treatment with HNG increased cell viability by 107.7% in Amy-β Active (1-42)-treated AMD cybrids (Po0.001, n = 3-4) compared to the AMD cybrids treated with active Amy-β Active (1-42) alone (Supplementary Table S10B). Data are represented as mean ± S.E.M., normalized to untreated-normal cybrids (assigned value of 1). The endpoint for all experiments was 72 h  Table S11). Scale bar = 20 μm. Data are represented as mean ± S.E.M., normalized to untreated-normal cybrids (assigned value of 1). The endpoint for all experiments was 72 h patients can significantly alter cell survival, reflecting similar events that occur in the AMD retinas.
Feher et al. used morphometric analyses of transmission electron micrographs to show that mitochondria of human primary RPE cells from AMD eyes were decreased in number and size, had disrupted cristae, disrupted mitochondrial membranes, and extensive reduction in matrix density compared to the controls. 48 In our study, multiple techniques were used to demonstrate that AMD mitochondria are damaged compared to age-matched normal. AMD cybrids exhibited a significant decrease in mitochondrial GFP fluorescence, lower mtDNA copy numbers and increased mtDNA fragmentation compared to normal cybrids. Furthermore, AMD cybrids had downregulation of TFAM, POLG, POLRMT, and TFB2M genes, suggesting impaired mitochondrial transcription/replication. Cytoprotective MDPs, such as Humanin and SHLPs, are encoded in the mitochondrial genome at the 16S rRNA gene. 15 Real-time PCR with primers spanning 503-2484 bps mtDNA regions showed increased numbers of mtDNA lesions in this region of the AMD mitochondrial genome. Our results are consistent with Ferrington et al.'s that demonstrated elevated levels of mtDNA lesions in specific regions, including MDP-coding regions of the mitochondrial genome, in AMD primary human RPE cells. 49 Other studies have also established a link between human RPE mtDNA damage and pathologic ageing in AMD. 7 It may be that mtDNA fragmentation disrupts the integrity of mitochondria and the threshold levels of MDPs essential for maintaining mitochondrial/cellular health and longevity.
In order to rescue and restore homeostasis to cells, we hypothesized that cellular and mitochondrial damage observed in AMD cybrids could be prevented by treatment with HNG, a more potent variant of Humanin. To that end, AMD and normal cybrids were treated with 3.2 μM HNG, co-stained with YO-PRO1/Propidium iodide and the apoptotic cell percentage measured using flow cytometry. As predicted, treatment of AMD cybrids with HNG led to a drastic decrease in numbers of apoptotic cells, confirming that HNG protected AMD cybrids against mitochondria-mediated cell death. These results are similar to Hinton et al.'s study showing Humanin was capable of protecting primary human RPE cells against oxidative stress-induced and ER stress-induced cell death, 24 using tert-Butyl hydroperoxide and tunicamycin, respectively.
Having confirmed the cytoprotective role of HNG on quiescent cybrids containing AMD mitochondria, we next determined the effects of HNG on cybrids stressed by exposure to amyloid-β peptides. Significant cytotoxicity was demonstrated in normal and AMD cybrids exposed to the active amyloid-β 1-42 peptide. However, pretreatment with HNG protected both AMD and normal cybrids from amyloid-β 1-42induced toxicity. These findings are significant because amyloid-β is a key constituent of drusen, which are sub-RPE deposits composed primarily of lipid and proteins. 50 Previous studies have demonstrated that cleavage of APP (amyloid precursor protein) by proteases produces the long forms of amyloid-β (i.e.,amyloid-β 42 ) which are fibrillogenic, plaqueforming and associated with neurodegenerative disease states, including Alzheimer's disease. 51,52 Our findings agree with studies reporting the cytoprotective roles of Humanin and HNG against amyloid-β-induced cell death in numerous agingrelated diseases. 53 For instance, Hashimoto et al. demonstrated that Humanin inhibits neuronal cell death induced by amyloid-β, mutant APP, and presenilin. 54,55 Zhang et al. showed that administration of HNG attenuates neuro-inflammation, cognitive deficits, and amyloid-β burden in a transgenic mouse model of Alzheimer's disease. 56 Humanin and/or HNG have also been shown to offer protection in cardiovascular, metabolic, inflammation, and cancer disease models. 20,57,58 One mechanism by which HNG protects AMD cybrid cells from death may be through stabilization of mitochondria and prevention of mitochondrial death. HNG-treated AMD cybrids had 197.1% increase in mtGFP fluorescence compared to untreated-AMD cybrids, demonstrating that the AMD mitochondria can be rescued by HNG pretreatment. This finding agrees with a previous study showing Humanin protects mitochondria in primary RPE cells. 24 Next, we investigated the mechanisms by which HNG exerts its cytoprotective effects. Humanin affects cells via both intracellular and extracellular mechanisms. 13 Humanin can bind BAX protein, which prevents the coupling of tBid with BAX, and the release of cytochrome c, thereby inhibiting apoptosis. 59 In untreated-AMD cybrids, the BAX protein levels were significantly elevated by 146% but declined drastically by 74% after HNG treatment. These findings indicate that HNG exerts its cytoprotective action via the intracellular mechanism involving BAX protein. With respect to the extracellular mechanism, Humanin is known to bind to a trimeric CNTFR/ gp130/WSX-1 receptor complex, inducing activation of JAK-STAT pathway and transcription of downstream genes. 60 In the current study, untreated-AMD cybrids had significantly reduced gp130 protein compared to untreatednormal. However, addition of HNG led to 61.87% increase in gp130 protein levels compared to untreated-AMD cybrids, thus restoring levels to that seen in normal cybrids. This suggests that AMD cybrids may have either a low abundance of the trimeric receptor or perhaps a faulty receptor complex which does not allow Humanin to bind and exert its cytoprotective effects. We speculate that gp130 and the trimeric receptor complex is associated with HNG-induced rescue of AMD cybrids. Furthermore, the initially low levels of phospho-JAK2 protein were restored with HNG in normal and AMD cybrids, suggesting that restoring gp130 levels may mediate activation of JAK and promotes cytoprotective effects.
To conclude, this novel study demonstrates that mitochondria from AMD patients are severely damaged, and highlights the protective role of HNG against mitochondria-mediated and amyloid-β-induced cell death in AMD transmitochondrial ARPE-19 cybrids. Our findings support the hypothesis that HNG is a significant cell survival factor and should be considered as a potential therapy for the treatment of dry AMD. These results also highlight the potential beneficial properties of Humanin and other MDPs that have been characterized in non-ocular disease models but have not yet been investigated in AMD and other age-related eye diseases.

Materials and Methods
Human subjects. Research involving human participants was approved by the Institutional review board of the University of California, Irvine (#2003-3131). Written informed consent was obtained from all subjects (Supplementary Table S1) and clinical investigations were conducted according to the tenets of Declaration of Helsinki for research involving human subjects. We found no significant difference (P = 0.07) in ages of normal (72.60 ± 1.7 years) and AMD (77.60 ± 1.6 years) subjects (N = 5).
Transmitochondrial cybrids. Transmitochondrial cybrids were created by fusing mitochondrial DNA-deficient APRE-19 (Rho0) cell line with platelets isolated from either AMD patients or age-matched normal subjects. Peripheral blood (10 ml) was collected via venipuncture in tubes containing 10 mM EDTA. Blood platelets were isolated by a series of centrifugation steps, in tubes containing 3.2% sodium citrate, and final pellets were suspended in Tris buffered saline. To create the ARPE-19 Rho0 cells, ARPE-19 cells were serially passaged into medium containing 50 ng/ml ethidium bromide. 30 Cybrids were created by polyethylene glycol fusion of platelets with Rho0 ARPE-19 cells in medium containing uridine. After growing the cybrid cells in uridine-containing medium for 2 weeks, the medium was replaced with regular culture medium. Mitochondrial DNA haplogroups profiles of each cybrid were confirmed using PCR, restriction enzyme digestion, and sequencing. The SNPs defining the H haplogroup were T7028C and G73A. All transmitochondrial cybrids were grown in DMEM/Ham's F12 1:1 (Invitrogen-Gibco, Gaithersburg, MD, USA) cell culture medium containing 24 mM sodium bicarbonate, 10% dialyzed fetal bovine serum, and 1.0 mM sodium pyruvate. Passage 5 cybrids were used for all experiments. Age-matched normal cybrids served as controls.
Treatment with Humanin G (HNG). Lyophilized HNG (Anaspec, Fremont, CA, USA) was reconstituted in water to obtain a stock solution that was subsequently dissolved in culture media to obtain a 3.2 μM HNG working solution.
In this study, all cybrids were treated with 3.2 μM HNG. At the beginning of this study we tested a wide range of concentrations (based on the published Humanin literature) ranging from 0.8 to 10 μM to determine a low optimum working concentration of HNG in our ARPE-19 cybrid cells. The HNG concentrations of 0.8 and 1.6 μM did not show any protective effects. Our preliminary studies demonstrated that 3.2 μM was the lowest concentration that showed significant protective effects in the cybrid cells. Therefore, we used 3.2 μM as our final working concentration for all experiments with HNG.
MTT assay. The MTT assay is a colorimetric test that utilizes the phenomenon of reduction of tetrazolium salts to measure cell viability. Healthy actively metabolizing cells contain NAD(P)H-dependent cellular oxidoreductase enzymes that reflect the number of viable cells present. These enzymes reduce the water-soluble yellow tetrazolium dye MTT [3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide] to its insoluble formazan, which has a purple color. The formazan crystals are then solubilized by adding dimethyl sulfoxide (DMSO) and the colorimetric signal, which is proportional to the number of viable cells, is determined by measurement of optical density at 570 nm. MTT assays were performed in the dark since the MTT reagent is sensitive to light.
AMD and normal cybrids (n = 4-5) plated in 96-well tissue culture plates were treated with MTT solution (Cat.# 30006, Biotium, Fremont, CA, USA) at 24, 48, 72, and 96 h time points. MTT-treated cells were incubated at 37°C for 1 h, followed by addition of DMSO. Absorbance was measured on a spectrophotometer at 570 nm and background absorbance measured at 630 nm. Background absorbance was subtracted from signal absorbance to obtain normalized absorbance values. The colorimetric signal obtained was proportional to the cell number.
Trypan Blue dye exclusion assay. Trypan blue is a diazo dye that selectively stains dead cells blue. Trypan blue dye exclusion test is based on the principle that live cells that have intact cell membranes are impervious to Trypan blue dye; therefore viable cells are excluded from staining blue. Since in our study, cell growth of cybrids peaked at 72 h, AMD and normal cybrids (n = 5) were plated in six-well plates, trypsinized at 72 h, and transferred to a counting vial. Cell count in each sample was determined using a cell counter (ViCell, Beckman Coulter, Brea, CA, USA) that uses Trypan blue dye to determine the number of viable cells. A total of 50 images were captured and averaged per sample.
MitoSOX assay. The MitoSOX Red reagent is a live-cell permeant fluorogenic dye that allows highly selective detection of mitochondrial superoxides in live cells. Once the MitoSOX Red reagent is oxidized by superoxides, it exhibits red fluorescence that can be quantified. AMD and normal cybrids (n = 5) were plated in 24-well tissue culture plates for 72 h. The 5 mM MitoSOX reagent (Cat.# M36008, Invitrogen, Grand Island, NY, USA) stock solution was diluted with HBSS (Hank's balanced salt solution) buffer to make a 5 μM MitoSOX reagent working solution. Cells were treated with 5 μM MitoSOX reagent and incubated for 10 min at 37°C. Cells were then washed with HBSS buffer, and fluorescence was measured at excitation/emission maxima of 510/580 nm.
Confocal microscopy. To estimate the number of mitochondria in normal and AMD cybrid cells, CellLight mtGFP staining was performed. The CellLight Mitochondrial GFP stain (Cat. # C10582, Thermo Fisher Scientific, Waltham, MA, USA) is a baculovirus fusion construct having a mammalian promoter and a leader sequence of E1 alpha pyruvate dehydrogenase fused to GFP. Therefore, it specifically targets mitochondria and mtGFP fluorescence can be quantified. Normal and AMD cybrids (n = 5) were plated in four-well tissue culture chamber slides for 72 h. At 48 h, CellLight stain was added to the medium at a concentration based on the cell density at the end of incubation period. The cells in chamber slides were incubated overnight at 37°C. The cells were washed with 1 × TBS (Tris-buffered saline), fixed in paraformaldehyde, washed again with 1 × TBS, mounted in DAPI, coverslipped, and stored at 4°C. Confocal z-stack images were captured using the LSM-700 Confocal microscope (Zeiss, Thornwood, NY, USA). All confocal images were quantified using the inbuilt ZEN 2 lite software.
Mitochondrial copy number. Normal and AMD cybrids (n = 5) were grown in six-well plates, and total DNA was isolated at 72 h after plating. Quantitative realtime PCR was performed using TaqMan gene expression assays for 18S and MT-ND2 (Cat. # 4331182, Thermo Fisher Scientific) genes and TaqMan gene expression master mix (Cat. # 4369016, Thermo Fisher Scientific). Relative mtDNA copy numbers were determined using delta Cts. All samples were run in triplicates.
Mitochondrial DNA fragmentation assay. To examine mitochondrial lesions in cybrid DNA, real-time PCR was performed with total DNA isolated at 72 h from normal (n = 5) and AMD cybrids (n = 5) using primers covering the MT-12S and MT-16S region of the mtDNA (Forward primer: 5′-ATCCTACCCAGCACAC AC-3′; Reverse primer: 5′-GATTTGCCGAGTTCCTTTTACT)-3′, and the Accuprime TaqMan DNA Polymerase System (Cat. # 12339016, Thermo Fisher Scientific). PCR products were run on 1% agarose gels and imaged under UV light. Number of mtDNA lesions were counted and compared between normal and AMD groups.
qRT-PCR. qRT-PCR was performed using StepOnePlus Real-Time PCR system (Applied Biosystems, Carlsbad, CA, USA) to study the expression of apoptosis, autophagy, ER stress, and antioxidant genes. Each sample was run in triplicate. The QuantiTect Primer Assays (Qiagen) (Supplementary Table S2) and Power SYBR Green PCR Master Mix (Cat. # 4367659, Life Technologies, Grand Island, NY, USA) were used. Specific housekeeper genes used are listed in Supplementary Table S2. qRT-PCR data were analyzed using ΔΔCt method. ΔCt was the difference between the Cts (threshold cycles) of the target gene and Cts of the housekeeper gene (reference gene). ΔΔCt was calculated by subtracting ΔCt of the AMD group from ΔCt of the normal group. Fold change was calculated using the following formula: Fold change = 2 ΔΔCt .

Western blotting analyses
Protein extraction: Normal and AMD cybrid cells (n = 3-5) were plated in sixwell plates for 72 h. Cells were lysed using RIPA buffer (Cat. # 89900, Life Technologies), supernatants transferred to a new microfuge tube and concentrations of proteins were measured using Bio-Rad Dc protein assay system (Bio-Rad Laboratories, Richmond, CA, USA) according to the manufacturer's instructions.
Immunoblotting: Equal concentrations of total protein samples were loaded into the wells of 4-12% Bolt mini gels (Life Technologies) followed by SDS-PAGE electrophoresis. The gels were then transferred onto PVDF membranes. Following transfer, the membranes were blocked in 5% milk for 1 h, then incubated overnight at 4 o C in primary antibodies (Supplementary Table S3). Blots were then washed with 1 × TBST (Tris Buffered Saline-Tween20) and incubated with the respective secondary antibodies for 1 h at room temperature. All primary and secondary antibodies were diluted in 1 × TBST. Following secondary antibody incubation, the blots were washed with 1 × TBST. Protein bands were detected using Clarity Western ECL Blotting Substrate (Cat. #1705060, Bio-Rad). β-actin antibody was used as a housekeeper protein control. Protein bands were visualized using Versadoc imager (Bio-Rad), and quantified using ImageJ software (NIH Image).
Flow cytometry. Normal and AMD cybrids (n = 4-5) were stained with YOPRO-1/propidium iodide (PI) (Cat. # V13243, Invitrogen), and incubated on ice for 20-30 min. Following incubation, stained cells were analyzed by flow cytometry, using 488 nm excitation with green fluorescence emission for YOPRO-1 (i.e., 530/30 bandpass) and red fluorescence emission for PI (i.e., 610/20 bandpass), gating on cells to exclude debris. Standard compensation analysis using single-color stained cells was performed. The YO-PRO-1 dye (green fluorescent) can enter apoptotic cells, whereas PI (red fluorescent dye) cannot. Therefore, YO-PRO-1 dye and PI together provide a sensitive indicator for apoptosis. Live cells show a low level of green fluorescence, apoptotic cells show an incrementally higher level of green fluorescence, and dead cells show both red and green fluorescence.
Statistical analysis. Results between groups were analyzed for differences using the unpaired Student's t-test or one-way ANOVA followed by Tukey-Kramer post-hoc test (GraphPad Prism 5.0; GraphPad Software, CA, USA). Statistical significance was determined at P-values o0.05.