VEGF Rescues Cigarette Smoking–Induced Human RPE Cell Death by Increasing Autophagic Flux: Implications of the Role of Autophagy in Advanced Age-Related Macular Degeneration

Abstract

Purpose.: Cigarette smoking (CS) is the most consistent risk factor for advanced age-related macular degeneration (AMD). To verify the molecular basis for CS-induced RPE alterations, RPE cell survival levels after being exposed to CS in relation with VEGF expression and autophagic flux were evaluated.

Methods.: Cigarette smoking extract (CSE) was added to ARPE-19 cells and hydrogen peroxide (HP) was used as a pure oxidant control. Cell survival was measured by flow cytometry with annexin V–fluorescein isothiocyanate. Cell survival analysis was performed after pretreatment with anti-VEGF or recombinant VEGF. The expression of VEGF-A, VEGF-R1/R2, and soluble VEGF-R1 was determined by semiquantitative RT-PCR. LC3B-I (microtubule-associated protein-1 inhibitors), LC3B-II, and phosphorylation of Akt or Erk were measured with Western blot. Autophagic flux was determined by increasing LC3B-II levels with inhibitors of lysosomal proteases.

Results.: Incubation with 5% CSE for 16 hours induced approximately 30% cell death, which was similar to cell death levels when exposed to concentrations of 200 μM HP. Pretreatment with anti-VEGF did not decrease cell survival under CSE, unlike the decrease in cell survival shown with HP. However, supplementation with VEGF rescued CSE-induced RPE cell death. Interestingly, CSE caused an increase in autophagic flux, which was augmented with VEGF pretreatment. Cigarette smoking extract also degraded the total amounts of Akt levels, and VEGF blunted CSE-induced phosphorylation of Erk.

Conclusions.: Cigarette smoking extract, similar to HP, affects cell viability and induces expression of VEGF and its receptors. Increased autophagic flux accelerated by treatment of exogenous VEGF may have a role in rescuing CSE-induced RPE cell death.

Introduction

Although age-related macular degeneration (AMD) is the leading cause of blindness in elderly people in the Western world, its pathogenesis is not clearly understood. ^1,2  

The retinal pigment epithelium (RPE) is thought to be the primary site of pathology in this disease. ³ Pathogenic mechanisms include both genetic and environmental factors, which are related to primary RPE senescence, alterations in the complement pathway (inflammation), and oxidative stress. ¹  

Smoking is one of the most consistent risk factors in advanced AMD. ^2,4,5 Although smoking is known to be related to exudative and nonexudative types of AMD, there are conflicting reports on which type is more prone to occur. ⁵ Smoking is also associated with increased incidence of choroidal neovascularization (CNV) in presumed ocular histoplasmosis syndrome. ⁶  

Cigarette smoke contains more than 4800 compounds, yet the most important active compound in AMD has not been identified. ^1,5–7 Studies show that smoking is the most established environmental factor for developing AMD, and the identification of the molecular basis for smoking-induced alterations in RPE homeostasis has created potential strategies for treating AMD. ^4,8,9  

Autophagy is a program regulated genetically and was first identified as a cell survival protection mechanism when a cell is in state of nutrient deprivation. Environmental toxins initiate autophagy by endoplasmic reticulum stress—unfolded protein response or increased reactive oxygen species formation. ^10,11 Direct harmful environmental contaminants, such as cigarette smoke, induce protein damage causing autophagy resulting in necrotic cell death. ^9,12 However, it is yet to be determined if changes in autophagic flux are a cause or consequence of a disease or whether autophagic changes reflect alterations in the formation or discharge of autophagosomes. ⁹ As of now, the role of autophagy in AMD has mainly been confined to explain drusen accumulations. A recent study by Wang et al. ¹³ suggests that drusen, an early feature of AMD, may reflect increases in both mitochondrial damage and autophagy. They speculated that an increase in autophagy and release of intracellular proteins via exosomes by aged RPEs contribute to the formation of such drusen. 

Using human ARPE-19 cell lines, we investigated and compared the mechanism of cell survival in response to CSE exposure and compared it to cell survival under pure oxidant hydrogen peroxide (HP). We attempted to investigate the pathophysiologic role of cigarette smoke in both atrophic (RPE cell death) and neovascular (VEGF-related) AMD. ^2,5 In doing so, we first reconfirmed the role of autocrine VEGF signaling in CSE-induced cell death, as has been proposed to occur under oxidative stressful conditions. ³ Also, cigarette smoke, being an environmental toxic substance and not just a pure oxidant, we investigated if there was a change in autophagic flux and searched for its association with VEGF signaling. ^14,15  

Methods

Chemical Reagents and Cell Culture Medium

Dulbecco's modified Eagle's medium (DMEM) along with fetal bovine serum, F-12 nutrient mixture, HEPES buffer, gentamicin, and amphotericin B were all purchased from Hyclone Laboratories, Inc. (Logan, UT). Recombinant human VEGF165 (rhVEGF) was purchased from R&D Systems, Inc. (Minneapolis, MN); anti-VEGF neutralizing antibodies (PC315), from Calbiochem (San Diego, CA); and horseradish peroxidase (HRP)–conjugated secondary antibody, from Dako (Glostrup, Denmark). 

Cell Culture

The ARPE-19 cell line was purchased from ATCC (Manassas, VA) and preserved in DMEM, Ham's F-12 nutrient medium mixture (DMEM F-12; Invitrogen-Gibco, Carlsbad, CA), and 10% fetal bovine serum. The ARPE-19 cells were used within 10 passages, plated in 6-well plates at 1.5 × 10⁵ cells per well and incubated at 37°C under 5% (vol/vol) CO₂ and exposed to CSE or HP at 70% confluence. ³ The cells were washed with PBS and incubated in serum-free DMEM 1 hour before treatment with H₂O₂ or CSE. The cells were then harvested for cell death analysis after 16 hours. 

Preparation of CSE

The cigarette smoke extract (CSE) was made by bubbling smoke from 2 commercially available, filtered cigarettes (Marlboro 20 class A cigarettes containing 8.0 mg of tar and 0.7 mg of nicotine; Philip Morris Korea, Inc., Yangsan, Gyeongsangnam-do, Korea) at a rate of 1 cigarette per 2 minutes through 20 mL of prewarmed serum-free DMEM/F12 (1:1), as described in previous reports. ¹⁶ The CSE pH was adjusted to 7.4 and was sterile filtered through a 0.2 M filter (Sartorius Stedim Biotech, Goettingen, Germany). Standardization of the CSE preparation was done by measuring its absorbance (optical density = 0.65 ± 0.05 at 320 nm). The spectrographic pattern of absorbance at 320 nm was similar between the different preparations of CSE. The CSE was prepared for usage within 1 hour of each experiment after being diluted with a culture medium adjusted to pH 7.4 and sterile filtered as described for 10% CSE. 

Flow Cytometric Analysis of Apoptosis

The cell apoptosis rates were analyzed by flow cytometry (BD, Franklin Lakes, NJ) after FITC-conjugated Annexin V (Annexin V; Sigma-Aldrich, St. Louis, MO)/propidium iodide (PI, Sigma-Aldrich) and Hoechst 33342 (Sigma-Aldrich)/PI staining. ¹⁷ Following the Annexin V/PI method, the RPE cells were harvested by treatment with 0.25% trypsin, washed with PBS, and finally incubated at 37°C for 15 minutes in the dark with FITC-conjugated Annexin V (a phospholipid-binding protein used as a probe for phosphatidylserine on the apoptotic cell's outer membrane) and PI. The specific fluorescence of 10,000 cells was analyzed with FACS Calibur (BD) within 1 hour after the addition of FITC-conjugated Annexin V. The data were analyzed by FSC express version 3.0 (DeNovo Software, Los Angeles, CA). Anti–VEGF-A neutralizing antibody was added 2 hours before H₂O₂ or CSE treatment. 

Semiquantitative RT-PCR

The expression of VEGF, VEGF-R1, sVEGF-R1, membrane-bound VEGF-R1, VEGF-R2, and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) were evaluated by using RT-PCR. The total RNAs in the RPE cells were isolated by TRIzol reagent (Invitrogen-Gibco), and the cDNA was synthesized by using 1g total RNAs according to the instructions of the manufacturer. The PCR primer sequences were as follows: human VEGF, 5′-ATG GCA GAA GGA GGG CAG CAT-3′ and 5′-TTG GTG AGG TTT GAT CCG CAT CAT-3′; VEGF-R1, 5′-GTAGCTGGCAAGCGCTCTTACCGGCTC-3′ and 5′-GGATTTGTCTGCTGCCCAGTGGGTAGAGA-3′; sVEGF-R1, 5′-CCA GGA ATC ACA CAG G-3′ and 5′-CAA CAA ACA CAG AGA AGG-3′; mbVEGF-R1, 5′-CCA CCT TGG TTG CTG AC-3′ and 5′-TGG AAT TCG TGC TGC TTC CTG GTC C-3′; VEGF-R2, 5′-TCT GGT CTT TTG GTG TTT TG-3′ and 5′-TGG GAT TAC TTT TAC TTC TG-3′; GAPDH, 5′-GCC AAG GTC ATC CAT GAC AAC-3′ and 5′-GTC CAC CAC CCT GTT GCT GTA-3′. All the PCR products were separated by using 1% agarose gel and observed by Gel-Doc system (Bio-Rad, Hercules, CA). 

Enzyme-Linked Immunosorbent Assay

The RPE cells were treated with differing concentrations of H₂O₂ at baseline (0 hour) and at 16 hours. The supernatants were collected, centrifuged, and stored at 70°C before ELISA (R&D Systems) was performed according to the manufacturer's instructions. VEGF-A levels were adjusted to reflect total protein concentrations. The level of VEGF-A proteins was measured in cell-free supernatant by using a human VEGF-A ELISA kit (Quantikine; R&D Systems). 

Immunoblotting

Western blot was done with a standard method. ³ Cells were rinsed twice with cold Tris-buffered saline (TBS) and lysed in a lysis buffer (50 mM Tris-HCl, pH 7.5, 100 mM NaCl, 0.1% SDS, 1 mM EDTA, 1% NP-40, 0.5% sodium deoxycholate, 10 mM sodium fluoride, 1 mM sodium orthovanadate, and 100 mg/mL PMSF). After incubation with ice for 30 minutes, the cell lysates were clarified by centrifugation at 4°C at 16,000g for 10 minutes and the supernatant was separated for Western blot and protein analysis. The total protein concentration was measured by using BCA kit (Sigma-Aldrich). Equal amounts of the proteins were fractionated by using 15% SDS-PAGE, then transferred to a nitrocellulose membrane, and incubated at 4°C overnight with primary antibodies against LC3 and GAPDH (SantaCruz, Santa Cruz, CA). Anti–phosphor-Akt antibody, anti-Akt antibody, anti–phosphor-Erk1/2 antibody, and anti-Erk1/2 antibody were all obtained from Cell Signaling Technology (Beverly, MA). 

The membranes were washed twice with TBS Tween 20 (Sigma-Aldrich) and then probed at room temperature for 1 hour with the corresponding secondary antibodies conjugated with HRP. Detection was carried out by an enhanced chemiluminescence detection kit (Pierce, Rockford, IL) and by autoradiography. The relative band intensities were determined densitometrically by the Quantity One software (Quantity One, Hercules, CA). All the data from 3 independent experiments were noted as the ratio to optical density values of the corresponding controls for statistical analyses. 

Statistical Analyses

Statistical analyses were performed by using SPSS version 12.0 for Windows (SPSS, Chicago, IL). The results were presented as the mean SD, with P value of 0.05 as statistically significant. 

Results

Effects of CSE and Hydrogen Peroxide on Cell Survival in RPE

Cigarette smoking extract from 1% to 10% significantly reduced cell survival in a concentration-dependent manner. After 16 hours of incubation, 1% CSE induced approximately 10% cell death, 5% CSE induced approximately 30%, and 10% CSE induced nearly 85% RPE cell death (Figs. 1A, 1B). 

Figure 1

 

Cigarette smoking extract and H₂O₂ induction of cell death in RPE cells. RPE cells were incubated with 1%, 5%, and 10% CSE and 100, 200, and 400 μM H₂O₂ for 16 hours and the cells were analyzed by flow cytometric analysis by using annexin V-fluorescein isothiocyanate and PI staining. Incubation in serum-free DMEM for 16 hours was used as a control. (A) Each panel shows a typical flow cytometric scatter diagram of 10,000 cells per sample from a representative experiment. Lower left: Viable and undamaged cells (annexin V, PI). Lower right: Cells undergoing early apoptosis (annexin V, PI). Upper right: Necrotic or late apoptotic cells (annexin V, PI). (B) Each bar shows the mean ± standard deviation (SD) of results in 9 to 12 wells in 3 independent experiments. *P < 0.05 compared with the control.

To investigate the mechanism of CSE cytotoxicity, we used HP as a pure oxidant control for comparison because CSE is also known to be a strong oxidant. ⁸ After 16 hours of incubation, 100 μM HP exposure induced approximately 10% cell death; 200 μM, approximately 30% cell death; and 400 μM, nearly 80% RPE cell death. In our experiment, the cytotoxic effects of 5% CSE on RPE cells were comparable to those of 200 μM HP. 

Role of VEGF Signaling Pathway on RPE Cell Viability Under CSE and Hydrogen Peroxide

Autocrine VEGF signaling is known to participate in RPE cell survival under oxidative stress. To confirm whether autocrine VEGF signaling also plays a role under CSE, we first examined RPE cell expression of VEGF and VEGF receptors (VR) after exposure to CSE. According to RT-PCR results, levels of VEGF, VEGF-R2, VEGF-R1, and soluble VEGF-R1 increased after exposure to 1% and 5% CSE. This effect is similar to increases in response after exposure to 100 μM and 200 μM HP (Fig. 2A). The concentration of VEGF in the cell supernatant also increased after CSE exposure as in HP (Fig. 2B). 

Figure 2

 

Expression of VEGF-A and VEGF receptors by CSE or H₂O₂ in RPE cells. (A) After 1 hour of CSE or H₂O₂ treatment, VEGF-A mRNA and VEGF-R1 and VEGF-R2 mRNA expression in RPE cells was determined. The mRNA levels were measured 1 hour after inoculation of each of the 2 concentrations of CSE or H₂O₂ showing increasing values. (B) VEGF-A excretion into the medium was measured by ELISA. After 16 hours of treatment with 5% CSE or 200 μM H₂O₂, the supernatant was collected and analyzed by ELISA. Data are expressed as the mean ± SD of the results in 3 independent experiments. sVR-1, soluble VEGF-R1; mbVR-1, membrane-bound VEGF-R1; VR-2, VEGF-R2.

Next, to investigate the involvement of “autocrine” VEGF signaling on cell survival, cell survival rates after pretreatment with anti-VEGF were measured. Surprisingly, blocking “autocrine” VEGF signaling with anti-VEGF did not significantly influence RPE cell survival under CSE in concentrations of 1% and 5%. This is a much different response from that of HP where introduction of anti-VEGF caused a dramatic decrease in cell survival (Figs. 3A, 3B). When adding exogenous VEGF, there was an increase in cell survival, suggesting the possibility that although the autocrine VEGF signaling pathway may not be activated for cell survival from CSE, VEGF may rescue RPE cells from further cell death through a different route (Figs. 4A, 4B). 

Figure 3

 

Contribution of autocrine VEGF survival signaling under CSE or H₂O₂. (A) Anti-VEGF antibody was added 2 hours before treatment with 1% or 5% CSE for 16 hours. Cell death was analyzed by flow cytometry of cells tagged with FITC-labeled annexin V and PI. Anti-VEGF treatment did not affect cell survival rates under either 1% or 5% CSE. (B) Cell survival analysis showed that RPE cell survival decreased by pretreatment with anti-VEGF antibody and 100 or 200 μM H₂O₂ for 16 hours. *P < 0.05.

Figure 4

 

Influence of VEGF supplements in CSE-induced RPE cell death. Each scatter diagram shows cell survival analysis for RPE cell death after 5% CSE exposure for 16 hours. Pretreatment of anti-VEGF did not influence cell survival rates. However, CSE-induced RPE cell death was inhibited by concomitant supplementation with VEGF. *P < 0.05.

CSE But Not Hydrogen Peroxide Increases Autophagic Flux in RPE Cells

Cigarette smoking extract is a complex chemical compound and such environmental toxic materials usually cause increased autophagic flux. After exposing RPE to CSE or HP, changes in autophagic flux were determined by analyzing levels of LC3-II. Autophagic flux was determined by changes in the amount of LC3-II in the presence and absence of a lysosomal protease inhibitor. 

Treatment with 5% CSE and 200 μM HP for 1 hour increased levels of LC3-II, implying the presence of a large number of autophagosomes. With the addition of a protease inhibitor, the relative levels of LC3-II further increased in response to CSE, which implies an increase in autophagic flux. However, under 200 μM HP, LC3-II levels did not increase, suggesting a reduction in autophagic flux (Fig. 5). 

Figure 5

 

Cigarette smoking extract but not H₂O₂ increases autophagic flux. Autophagic flux was assayed by immunodetection of the LC3 protein of RPE cells exposed to CSE or H₂O₂ for 1 hour. The ratio of LC3 protein suggests enhanced autophagy with 5% CSE but not with 200 μM H₂O₂. GAPDH served as the loading control and was used as the housekeeping gene. AF, autophagic flux (change in LC3-II level between samples treated with inhibitors or not).

Applying (Exogenous) VEGF Increases Autophagic Flux Under CSE

To determine whether VEGF can influence the recovery or enhance autophagic flux for protecting cell death, the changes in amounts of autophagy were determined by analyzing LC3-II levels in VEGF-treated and VEGF-untreated cells. 

When the cells were exposed to 5% CSE, autophagic flux increased at both times (1 and 8 hours), and 1 hour of pretreatment with rhVEGF further increased the levels of autophagic flux at each time points (Fig. 6). 

Figure 6

 

Pretreatment with VEGF increases autophagic flux under CSE. Western blot analysis was done for LC3 in samples taken at the indicated times after 5% CSE exposure in RPE cells. Levels of LC3-II, the active form, began to increase at 1 hour and were substantially higher at 8 hours. With the addition of a protease inhibitor, the relative levels of LC3-II further increased, implying an increase in autophagic flux. The autophagic flux was further increased with pretreatment of VEGF.

Influence of CSE and VEGF on Cell Survival Signal; Akt and Erk Signaling Pathways

Among the growth factor–autophagy signal pathways, the growth factor–Akt–autophagy pathway is the most well-established pathway. The Akt survival pathway was also the major pathway in our previously reported studies on autocrine VEGF survival signaling. However, in this study, cell death by CSE does not involve the use of this circuit because blocking VEGF did not have any apparent detrimental effects on cell survival. Therefore, we investigated if there were any changes in Akt signals after exposure to CSE. 

With CSE treatment, total Akt levels continued to decrease over the period starting as early as 15 minutes after CSE exposure, despite p-Akt levels increasing as early as 15 minutes but soon returning to previous levels (Fig. 7A). Total Akt levels still continued to decrease during the period from 15 minutes to 180 minutes even with pretreatment of VEGF (Fig. 7B). Such degradation of total Akt protein may be responsible for the minimal role of autocrine VEGF signaling (VEGF–Akt-survival) under CSE. On the other hand, CSE induced phosphorylation of another stress-related cell signal, Erk, as early as 15 minutes but was blunted with pretreatment of VEGF. Also, total Erk levels showed no definite decrement over the time course irrespective of VEGF treatment (Figs. 8A, 8B). 

Figure 7

 

Cigarette smoking extract mediated Akt phosphorylation in RPE cells. Total protein was extracted from serum-starved RPE cells after being incubated with 5% CSE for various times. (A) The phosphorylation of Akt started as early as 15 minutes and returned to the pre-level with time. Pretreatment of VEGF increased the p-Akt level initially, which increased slightly more at 15 minutes. However, total Akt levels showed a continuous decrement over the period irrespective of VEGF pretreatment. (B) Densitometry results expressed as a ratio of loading control to β-actin. Data are expressed as the mean ± SD of the results in three independent experiments.

Figure 8

 

Cigarette smoking extract mediated Erk phosphorylation in RPE cells. (A) Cigarette smoking extract induced Erk phosphorylation in RPE cells as early as 15 minutes. CSE-induced Erk phosphorylation was blunted by pretreatment of VEGF. Total Erk protein expression showed no definite decrement. (B) Densitometry results expressed as a ratio of loading control to β-actin. Data are expressed as the mean ± SD of the results in 3 independent experiments.

Comparing the Effects of VEGF and b-FGF on Autophagic Flux

The changes in amount of autophagy were determined by analyzing LC3-II levels in VEGF- and b-FGF–treated RPE cells. Both VEGF and b-FGF were treated 1 hour before treatment with 5% CSE. The levels of autophagic flux increased at each time point (before and 1 hour after CSE treatment) with both VEGF and b-FGF (Fig. 9). 

Figure 9

 

Comparing the effects of VEGF and b-FGF on autophagic flux. Autophagic flux was determined by analyzing LC3-II levels in VEGF- and b-FGF–treated cells. VEGF and b-FGF were treated 1 hour before treatment with 5% CSE. The levels of LC-II increased before and 1 hour after CSE treatment with both VEGF and b-FGF. When adding a protease inhibitor, the relative levels of LC3-II further increased, implying an increase in autophagic flux. We found that b-FGF also increased autophagic flux in CSE-treated RPE cells similarly to VEGF.

Discussion

Autophagy is a catabolic mechanism that degrades cellular constituents. This process of autophagy also plays a significant role in cell survival, aging, and death. ^1,18 When the cell is stressed and activates autophagy, the proceeding cascade of pathways can result in either cell survival or cell death. ^9,10,15 In previous reports, oxidative stress in the form of HP has been reported to cause reduction of autophagic flux in RPE cells. ¹⁹ However, we found that contrary to HP, CSE induced an increase in autophagic flux in the RPE cells, which may be a protective response mechanism to toxic substances. The autocrine VEGF survival signaling pathway, which played a substantial part under HP, did not play a major role in RPE cell death induced by CSE. ^3,20 This may be related to the decrease of the Akt signaling protein by CSE. ¹⁸  

Our previous report has shown that the autocrine VEGF signaling pathway involves the VEGF–VEGFR-2 and Akt signaling pathway. ³ The growth factor–Akt–mTOR pathway usually involves reduction of the autophagic flux to reserve cell organelles before cell division. ^10,21,22 However, it was interesting that the treatment of the growth factor (VEGF) enhanced the autophagic flux under CSE-induced cell death in our study. Autophagic flux usually increases under such conditions as nutrition deprivation, growth factor withdrawal, or exposure to biological, chemical, or physical hazards causing hypoxia. ¹¹  

Cell responses to growth factors may be different for cells according to their cell cycles. Growth stimulation leads to cellular senescence when the cell cycle is blocked. ²³ In dividing cells, cell organelles are reserved before cell division. The removal of damaged organelles or proteins is not as urgent as in nondividing cells because in dividing cells, the burden of such toxic material or wastes can be halved between the divided cells. Therefore, we can hypothesize that the effects of growth factors on autophagic flux may be different according to the state of the cells, resulting in differing cell fates. ^10,11  

We found that CSE can induce increased expression of VEGF and receptors as it did with HP. ³ In epidemiologic studies, smoking is associated not only with atrophic changes but also with neovascularization. ^1,3,6 Moreover, the association with CNV is greater for current smokers than past smokers, which implies a potential relationship between CSE and VEGF. ^4,7  

In previous reports, we and others have established that autocrine VEGF-A signaling affects the Akt signaling pathway, which under conditions of oxidative stress (HP), may be used by RPE cells for survival. ^3,20 We found in this study that the autocrine survival signaling of VEGF was not the major pathway for rescue from cell death induced by CSE. We suspect that CSE-induced degradation of (total) Akt protein may be responsible for such findings. ¹⁸ Cigarette smoke is also known to induce reduction in VEGF-R2 phosphorylation and to decrease PI3K and p-Akt levels. ^14,24  

Cigarette smoke is not a simple oxidant compound but contains numerous heavy metal toxins such as cadmium. ^8,9,14 Oxidative stress from cigarette smoking is caused by the process of combustion and amplified by heavy metals. ¹³ In vivo studies with young cigarette smokers has shown increased levels of serum cadmium in comparison with nonsmokers. ²⁵ Cadmium is reported to induce reactive oxygen species–dependent activation of poly ADP ribose polymerase, resulting in a depletion of ATP and stimulating LKB1-AMPK signaling. ²⁶ Activation of this pathway initiates autophagy through the inhibition of mTOR. ¹⁵  

Interestingly, our studies showed that applying additional amounts of VEGF can rescue cell death from CSE. We speculate that VEGF may prevent cell death from CSE via opposing effects on MAPK/Erk activation. ^27,28 Oxidative stress is a potent stimulator of MAP kinase activities and CSE induces phosphorylation of Erk, as early as 15 minutes in our study. ^27,29,30 However, pretreatment of VEGF blunted this transient phosphorylation of Erk. Constitutive activation of active Erk (by active Raf or cadmium) has been reported to induce a form of cell death with massive vacuolization, implying autophagic cell death. ^29,31,32 VEGF is known to prevent apoptosis of human microvascular endothelial cells via opposing effects on MAPK/ERK signaling. ³³  

Autophagic reaction is itself a critical pathway determining degrees of health and morbidity at different points in time. ⁹ Under certain circumstances, autophagy activation serves as a protective mechanism through recycling cellular constituents and disposing of damaged organelles and potentially cytotoxic aggregates. Thus, we believe that VEGF affects cell survival favorably through blunting Erk phosphorylation and the induction of autophagic flux in critical situations. In the retina, especially in a pathologic retina, the RPE is not the only source of VEGF, which may be derived from various immune cells or other types of cells. ¹ Under toxic stress, increased environmental VEGF may help these stressed RPE cells to survive through the induction of autophagic flux. 

Currently, rapamycin (mTOR inhibitor–autophagy inducer) is under evaluation for clinical usage in the treatment of AMD. Rapamycin is known to prevent mitogen-induced hypoxia-inducible factor-1 and hypoxia-inducible factor-1–dependent transcription and the secretion of VEGF. ^21,22 Thus, rapamycin reduces VEGF expression in RPE with resultant decrease in RPE-induced angiogenesis through mTOR inhibition. Recently, other possible mechanisms of rapamycin for AMD treatment have been proposed. In AMD, like in choroidal neovascularization animal models, rapamycin has prevented significant abnormalities in RPE cells, in the retinal barrier, and in neurodegeneration of photoreceptors without the reduction of VEGF levels. ³⁴ As our studies have shown, we think that VEGF-induced autophagic flux increase may have some relationship with these findings. 

Dysfunction and atrophy of the RPE subsequently leads to CNV or geographic atrophy in late-stage AMD. Therefore, when RPE damage is not too substantial, inducing autophagy may be a feasible treatment option for early macular degeneration. On the other hand, if RPE damage progresses beyond a critical point, as in late-stage AMD, autophagy may lead to cell death and disease exacerbation. Regarding the association of autocrine survival signaling of VEGF and VEGF-induced autophagic flux, the abundance of VEGF may be a favorable environment for RPE cells, rescuing them from toxic stimulations. 

In our study, we found that CS may have a close relation with both RPE cell death (atrophic AMD) and VEGF expression (neovascular AMD). Cigarette smoke, as an environmental toxic substance, causes cellular stress and leads to the activation of autophagy in RPE cells. However, plenary circumferential VEGF may rescue toxicant-induced RPE cell death by augmenting autophagic flux. This may also hold true for other growth factors such as b-FGF (Fig. 9), necessitating further investigations regarding the role of each growth factor along different time points. Our results may provide a clue to understanding the pathophysiologic role of CS in both atrophic and neovascular AMD, with possible incorporation of autophagic modulators into future treatment strategies. 

Acknowledgments

Supported by the Converging Research Center Program through the Ministry of Science, ICT and Future Planning, Korea (2013K000365), and a faculty research grant of Yonsei University College of Medicine for 2009 (6-2009-0078). The authors alone are responsible for the content and writing of the paper. 

Disclosure: Y.K. Chu, None; S.C. Lee, None; S.H. Byeon, None