Characterization of stress response in human retinal epithelial cells

The pathogenesis of age-related macular degeneration (AMD) involves demise of the retinal pigment epithelium and death of photoreceptors. In this article, we investigated the response of human adult retinal pigmented epithelial (ARPE-19) cells to 5-(N,N-hexamethylene)amiloride (HMA), an inhibitor of Na+/H+ exchangers. We observed that ARPE-19 cells treated with HMA are unable to activate ‘classical’ apoptosis but they succeed to activate autophagy. In the first 2 hrs of HMA exposure, autophagy is efficient in protecting cells from death. Thereafter, autophagy is impaired, as indicated by p62 accumulation, and this protective mechanism becomes the executioner of cell death. This switch in autophagy property as a function of time for a single stimulus is here shown for the first time. The activation of autophagy was observed, at a lesser extent, with etoposide, suggesting that this event might be a general response of ARPE cells to stress and the most important pathway involved in cell resistance to adverse conditions and toxic stimuli.


Introduction
Age-related macular degeneration (AMD) is the major cause of blindness in the Western countries. Moreover, the number of patients is expected to increase from 1.75 million to 3 million in the next 10 years [1]. The pathogenesis of AMD involves demise of the retinal pigment epithelium (RPE) followed by death of photoreceptors [2]. The RPE is a specialized epithelium at the interface between the neural retina and the choriocapillaris where it forms the outer blood-retinal barrier. RPE daily phagocytes important amounts of lipid-rich material (photoreceptor outer segment) in the most oxygenated part of the body. So that, these cells possess high levels of the enzymes required to detoxify reactive oxygen species (ROS) [3]. Because elevated ROS formation is involved in the pathogenesis of many diseases, including cancer, diabetes and neurodegenerative diseases, as well as in ageing, the discovery of biochemical mechanisms that either protect cells or promote cellular recovery from damage is extremely important.
In the intact eye, the transition from light to dark alters pH, [Ca 2+ ] and [K + ], and the oxygen consumption increases inducing the release of CO 2 and H 2 O [4]. The RPE maintains pH and volume homeostasis of the subretinal space by transporting these products to the choroidal compartment. This manoeuvre decreases intracellular pH. In altered metabolic states like hypoxia and hyperglycaemia, H + accumulates because of the elevated glycolysis and failure of retinal circulation, thus the retina is readily acidified, affecting thus the intracellular pH of the surrounding cells. It is worth noting that a decrease in intracellular pH is also a common effect of many apoptotic stimuli such as cytokine deprivation or ROS generation [5][6][7].
To study the biochemistry and molecular biology of oxidative stress in RPE cells, many investigators have used in vitro cultures of ARPE-19 cells. These cells have spontaneously arise from a human RPE cell line derived from the normal eyes of a 19-year-old male who died from head trauma in a motor vehicle accident [8]. ARPE-19 cells presenting a normal karyotype are able to form polarized epithelial monolayers and express RPE-specific markers such as CRALBP and RPE-65.
In this article, we investigated the response of human adult retinal pigmented epithelial (ARPE-19) cells to 5-(N,N-hexamethylene)amiloride (HMA), an inhibitor of Na + /H + exchangers, i.e. a class of proteins acting as intracellular pH regulators [7], and we evaluated the pathways activated by these cells to face this metabolic stress. We focused on the analysis of apoptosis (both caspase-dependent and -independent), parthanatos and autophagy.
We observed that ARPE-19 cells treated with HMA are unable to drive 'classical' apoptosis and succeed in activating autophagy, which protects them under mild stress conditions, and ultimately leads to cell death.

Cell treatments
ARPE-19 cells were treated with 5-(N,N-hexamethylene)amiloride (HMA; stock solution: 80 mM in DMSO, A9561; Sigma-Aldrich, Saint-Quentin Fallavier, France) at increasing concentrations (from 20 to 120 lM) for up to 72 hrs. As a positive control of apoptosis, cell cultures were treated with the DNA topoisomerase II inhibitor, etoposide, which is used as an internal standard for cell proliferation and viability assays. The 50 mM stock solution of etoposide (E1383; Sigma-Aldrich) was prepared in DMSO; ARPE-19 cells were treated with 250 lM etoposide for up to 72 hrs. In some experiments, HeLa cells treated with 100 lM etoposide for 3 hrs, followed by 24 hrs of recovery in drug-free medium were used as a positive example of apoptosis. As a positive control of autophagy, cell cultures were treated with 1 lM rapamycin (R0395; Sigma-Aldrich), an inhibitor of mTOR complex, for 2 hrs. Stock solution was prepared at the concentration of 1.367 mM in DMSO. To inhibit autophagy, 3-methyladenine (3-MA) (M9281; Sigma-Aldrich) and bafilomycin (B1793; Sigma-Aldrich) were used; the first one (stock solution 1 M) was administered at 5 mM for 4 hrs before the treatment with HMA and the second one at 50 nM (stock solution 8 lM in DMSO) for 2 hrs.

Morphological analysis
For the staining with the fluorescent dye 4′,6-diamidino-2-phenylindole (DAPI), cells were seeded in 1 ml of complete medium at a density of 5 9 10 4 cells/well in multiwells containing a coverslip (16-mm diameter) and treated 48 hrs later with HMA at the indicated concentrations. Multiwells with cells still attached to coverslips were put on a bed of ice, washed with cold PBS and fixed with 4% paraformaldehyde for 15 min. After the fixation, cells were permeabilized with 0.3% Triton in PBS for 30 min. at r.t., washed three times with PBS and stained with DAPI (0.2 lg/ml in PBS) for 10 min. in the dark. After five washings of 5 min. with PBS, slides were finally mounted with antifade mounting medium (90% glycerol, 20 mM Tris-HCl, pH 7.5, 2.33% DABCO). Cells were observed using a fluorescence microscope Olympus BX51, equipped with a 40x objective. The images were acquired through a digital camera Camedia C4040 (Olympus, Tokyo, Japan), Adobe Photoshop (Redwood City, CA, USA) was used as elaborating software.

Transmission electron microscopy
ARPE-19 cells were seeded on Lab Tek slides, treated 48 hrs later with HMA at the indicated concentration, for 24 hrs. At the end of the treatment, cells were washed with PBS and fixed with 2.5% glutaraldehyde in sodium cacodylate for 30 min. and then washed in the same buffer twice for 15 min. Cells were post-fixed with 1% osmium tetroxide for 15 min. at 4°C in the dark, washed twice in cacodylate buffer and dehydrated in ethanol at increasing concentrations (50%, 70%, 80%, 90% and 100%). Slides were then incubated with Hydroxypropyl Methacrylate (64180; Sigma-Aldrich) twice for 5 min. each and embedded in epoxy resin. Ultrathin sections (80 nm) were contrasted by using uranyl acetate and observed using an electron microscope (model 100 CX II; JEOL, Tokyo, Japan).

Cytotoxicity assay
Cytotoxicity assay was performed as already described [9].
MTT assay ARPE-19 cells were seeded in 96-multiwell plates at a density of 2.5 9 10 3 in 100 ll of complete medium. Cells were treated 48 hrs later with different drug concentrations for 24-72 hrs. At the end of the treatment, 20 ll of CellTiter 96 non-radioactive cell proliferation assay (MTT, G4000; Promega, Milano, Italy) were added to each well. Plates were then incubated for 4 hrs at 37°C in the dark and analysed using a microplate reader (Gio De Vita, Rome, Italy) at 570/630 nm. Experiments were performed in quadruplicate and repeated three times.

Indirect immunofluorescence experiments
ARPE-19 cells were seeded at the density of 5 9 10 4 /ml in 1 ml of complete medium in multiwells containing 16-mm diameter coverslips.

104
ª 2012 The Authors. Published by Foundation for Cellular and Molecular Medicine/Blackwell Publishing Ltd.
Forty-eight hours later, cells were treated with HMA, etoposide, rapamycin or bafilomycin at the indicated concentrations. The following proteins: AIF, LAMP 2, LC3, LEI/L-DNase II, mtHSP70, p62/SQSTM1 and poly (ADP-ribose) were tested (see in Table 1 the list of the antibodies used).

LC3
At the end of the treatment, cells were washed three times with PBS for 5 min., fixed with 4% paraformaldehyde for 15 min. on ice, washed three times in PBS, permeabilized with cold 100% acetone for 5 min., washed three times with PBS for 5 min. and then saturated with 4% BSA (bovine serum albumin) in PBS for 10 min. Slides were then incubated with the polyclonal antibody against LC3 (2775; Cell Signaling, Danvers, MA, USA) diluted 1:100 in PBS, for 1 hr at 37°C in a humidified chamber, washed five times with PBS, incubated with the secondary antibody (111-225-003, Cy2conjugated anti-rabbit; Jackson, Suffolk, UK) diluted 1:50 in PBS, for 1 hr at 37°C in a humidified chamber. Coverslips were then washed three times with PBS in the dark and finally mounted with a drop of PBS.

Poly(ADP-ribose)
After fixation performed as described above, cells were permeabilized with 0.3% Triton in PBS for 30 min. at r.t., washed three times with PBS and then saturated with 5% skim milk in PBS for 30 min. Coverslips were then incubated with the monoclonal antibody 10H against poly(ADP-ribose, ALX-804-220-R100; Enzo Lab Sciences, Lausen, Switzerland) diluted 1:100 in PBS, for 1 hr at 37°C in a humidified chamber, washed five times for 5 min. with PBS, incubated with the secondary antibody TRITC-conjugated anti-mouse (115-025-146; Jackson Immu-noResearch) diluted 1:50 in PBS, for 1 hr at 37°C in a humidified chamber. Coverslips were then washed three times with PBS in the dark and finally mounted with a drop of PBS. Cells were observed using a fluorescence microscope Olympus BX51, equipped with a 40x objective. The images were acquired with a digital camera Camedia C4040 (Olympus), Adobe Photoshop was used as elaborating software.

Analysis of survivin protein expression and phosphorylation
Cells were lysed with a specific cell extraction buffer (0.01% NP40, 10 mM Tris-HCl pH 7.5, 50 mM KCl, 5 mM MgCl 2 , 2 mM DTT, 20% glycerol, 1 mM PMSF plus protease and phosphatase inhibitors). For the assessment of the Thr 34 -phosphorylated form of survivin, precleared, detergent-solubilized cell extracts (100 ll) were immunoprecipitated with the rabbit polyclonal anti-human survivin antibody (ab8228; Abcam) for 16 hrs at 4°C, with precipitation of the immune complex by addition of 50 ll of a 50:50 protein A slurry (17-0963-03; GE Healthcare, Limonest, France). Samples were separated on precast 12% NuPAGE TM Bis-Tris gel (NP0321PK2; Invitrogen Life Technology, Merisiers, Saint Aubin, France) and transferred onto Hybond ECL nitrocellulose membranes (RPN203D; GE Healthcare) using the Nu-PAGE transfer buffer (NP0006; Invitrogen Life Technology). Nitrocellulose membranes were blocked in PBS-Tween 20 with 5% skim milk, first incubated with the primary antibodies specific for survivin and Thr 34 -phosphorylated survivin (ab10720; Abcam.) and then with the secondary peroxidase-conjugated antibodies. Bound antibodies were detected using the SuperSignals West PICO chemiluminescent substrate (34077; Pierce Biotechnology, Courtaboeuf, France) and autoradiography. b-Actin monoclonal antibody (ab8229; Abcam) was used to confirm equal protein loading on the gel. Results were quantified by densitometric analysis using the Image-Quant software (Molecular Dynamics, Sunnyvale, CA, USA).

Internucleosomal DNA degradation
Internucleosomal DNA degradation assay was performed as already described [9].

Clonogenic Test
ARPE-19 cells were seeded in 1 ml of complete medium at the density of 2.3 9 10 4 cells/well in 24-well plate and treated 48 hrs later with 80 lM HMA alone or in combination with 5 mM 3-MA or 1 lM rapamycin, as before. At the end of the treatment, cells were washed with PBS and trypsinized. For each condition, 10 3 cells were seeded into 6-well plate and grown with complete medium. One week later, medium was removed and cells were carefully washed with PBS, fixed and stained with 2 ml of 6% glutaraldehyde and 0.5% cresyl violet. After 30 min., the staining solution was removed; plates were washed with tap water, and dried in normal air at room temperature. The surface of the plate occupied by the colonies was calculated as follows: pictures of each plate were taken in constant conditions of distance between the camera and the plate, source of light, ISO, zoom, exposition time and aperture of diaphragm. The area of analysis for the wells on the plate was selected (the radius was the same for all the wells). Using image analysis software, every image was binarized and the rate between the area occupied by colonies and the total area of the well was calculated.

ARPE-19 cell response to HMA treatment
To test ARPE-19 cell response to HMA, we first measured cell viability by the MTT (3-(4,5-Dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide) assay. Treatment of ARPE-19 cells with different HMA concentrations (20-120 lM) for increasing times (up to 72 hrs) revealed that HMA has a dose/time-dependent effect on ARPE-19 cell viability (Fig. 1A), with an IC50~60 lM. This value is higher than the IC 50 for other cells (~20 lM for BHK, HeLa and colon carcinoma cells, not shown). Etoposide (a well-known apoptosis inducer) did not affect ARPE-19 cell viability, possibly causing an arrest in cell proliferation, as suggested by the absorbance values at different time-points (Fig. 1A, Eto).
To evaluate the impact of HMA on cell proliferation, we measured the DNA released from total extracts after alkaline lysis; this parameter is proportional to the cell number [9]. This assay further confirmed the dose/time-dependent decrease in ARPE-19 cell number after HMA treatment (Fig. 1B). On the whole, these results indicated that HMA interferes with the proliferation and viability of ARPE-19 cells, triggering also cell death.
To verify if HMA cytotoxic effect was correlated to the apoptotic process, we evaluated initiator caspase status after drug treatment, and found no evidence of caspase-8, and -9 activation; moreover, effector caspase-3 showed a modest conversion into the active form only in cells treated with 120 lM HMA ( Fig. 2A). Similar results were obtained with etoposide (Eto), thus confirming the intrinsic resistance of ARPE-19 cells to activate caspase-dependent apoptosis, as further supported by the absence of DNA ladder in HMA-treated cells (Fig. 2B). The analysis of PARP-1 proteolysis, which is a classical marker of caspase-dependent cell death [9], further confirmed that HMA did not induce classical apoptosis (Fig. 2C). To give some insights into the resistance of ARPE-19 to activate caspases, we investigated the expression of survivin, a major survival factor that blocks apoptosis through the inhibition of caspases. To our surprise, survivin, which is normally expressed in embryonic and cancer cells is highly expressed in ARPE-19 cells. There was an increase in surviving expression, as compared with the expression of actin (Fig. 2D). The ability of HMA to inhibit survivin activation by interfering with protein phosphorylation on Thr 34 residue was also investigated. Results from immunoblotting experiments using a phosphospecific antibody indicated that the levels of the active, Thr 34 -phosphorylated form of survivin were reduced following treatment of ARPE-19 cells with HMA at both time-points (Fig. 2D), suggesting that this factor, even if present, is not protecting ARPE-19 cells from death.
As mentioned above, no cleavage of PARP-1 was observed but, unexpectedly, HMA-treated cells showed increased levels of PARP-1 expression (Fig. 2C). To investigate if this event was correlated to increased PARP-1 activity, we performed immunofluorescence experiments using an anti-poly(ADP-ribose) (PAR) antibody (Fig. 3A). PAR synthesis, which is a stress response signal, appeared to be increased after HMA exposure. In fact, HMA stimulated PAR accumulation, visible as the appearance of speckled nuclei after 40 lM HMA treatment for 24 hrs. The treatment of cells with higher HMA concentrations (80 lM and 120 lM) or 250 lM etoposide revealed the appearance of brilliant nuclei (Fig. 3A).
Given that PAR synthesis is involved in the caspase-independent apoptotic pathway called parthanatos, driven by AIF (Apoptosis Inducing Factor) translocation from mitochondria to the nucleus [10], we investigated intracellular AIF localization. In untreated ARPE-19 cells, AIF co-localized with the mitochondrial marker mtHSP70 Results are expressed as the mean ± SD of three independent experiments. All measurements were significantly different as calculated using a one way ANOVA test P < 0.05 (B) To monitor cell proliferation, cells were lysed in alkaline buffer and the released DNA was measured. Results are expressed as the mean ± SD of three independent experiments All measurements were significantly different as calculated using a one way ANOVA test P < 0.05.  (Fig. 3B). The treatment with 40 lM and 80 lM HMA for 24 hrs did not promote a significant AIF translocation from mitochondria to the nucleus, which was visible in cells treated with 120 lM HMA, where cells became rounded and AIF appeared to move to the nucleus. These data suggest that the activation of this pathway occurs only at high HMA concentrations. We investigated also a paradigm of apoptosis that involves leucocyte elastase inhibitor (LEI) conversion into its DNase active form, called L-DNase II [11] and that implies the activation of PARP-1. [12,13] Immunofluorescence experiments revealed that in untreated cells, the labelling for LEI/L-DNase II proteins showed a low and diffuse pattern, while a treatment of 72 hrs with 250 lM etoposide induced the translocation of LEI inside the nucleus of ARPE-19 cells (Fig. 4A). In HMA-treated cells (40 lM), the nuclear fluorescence of LEI/L-DNase II supported its translocation inside the nucleus. Using higher drug concentration (80 lM), L-DNase-II aggregated into foci inside the nucleus; however, this phenomenon was no longer visible in cells treated with 120 lM HMA (Fig. 4A). The conversion, and therefore activation, of LEI into its L-DNase II form was further confirmed using Western blot analysis, which revealed the appearance of an additional immunoreactive band (corresponding to L-DNase II) after HMA treatment (Fig. 4B). This body of evidence indicates that the activation of L-DNase II, an endonuclease known to be active in apoptotic-like cell death, was involved in ARPE-19 cell demise. As apoptosis displays very characteristic changes in nuclear morphology, we evaluate HMA effect on this parameter. To do this, we stained ARPE-19 cells with DAPI after 40, 80 and 120 lM HMA treatment for 24 hrs. ARPE-19 nuclei showed no signs of canonical apoptosis (i.e. nuclear shrinkage and apoptotic bodies); interestingly, the cytoplasm of treated cells displayed various fluorescent and brilliant dots (Fig. 5A).
To get inside the nature of this fluorescence, we analysed HMA spectral behaviours, providing the evidence that HMA emits blue light when irradiated with UV light [14], and that the compound is accumulated in these vesicles. Electron microscopy observation revealed that 40 lM HMA treatment for 24 hrs triggered the formation of several vacuole-like structures different in shape and content, being either filled with electron-dense material, likely of mitochondrial origin, or empty and surrounded by double membranes, suggestive of autophagosomes (Fig. 5B, upper panel). Moreover, HMA treatment induced nuclear as well as endoplasmic reticulum membrane dilatation and phase I chromatin condensation, indicating that HMA affected both organelles and nuclear integrity (Fig. 5B, lower panel).

HMA induces autophagy in ARPE-19 cells
To investigate the nature of vesicle-like structures, we monitored the status of LC3 protein, a specific marker for autophagy ( Fig. 6A and B). In untreated ARPE-19 cells, LC3 labelling was low and cytoplasmic; in rapamycin-treated samples, which are representative of autophagy, several LC3-labelled dots corresponding to autophagosomes were detectable (Fig. 6A, left panel). These structures were visible in cells treated with 40 and 80 lM HMA for 24 hrs, while 120 lM treated cells showed a more diffused labelling. In fact, the punctuated labelling of LC3 is apparent from 2 hrs of treatment with HMA (data not shown). DMSO alone, the vehicle of HMA produced a labelling identical to the control. Interestingly, also 250 lM etoposide caused the appearance of few autophagosomes in ARPE-19 cells (Fig. 6A, right panel).
The analysis of LC3 status using Western blot further confirmed that HMA induces autophagy, as demonstrated by the appearance of LC3-II protein, the lipidated form of LC3 (Fig. 6B, left panel). In etoposide-treated cells, a modest amount of LC3-II was also present, suggesting that autophagy activation is a general feature of ARPE-19 cells. We further investigated the expression of different proteins involved in the autophagic process. Western blot analysis revealed that Beclin-1, one of the most important autophagy regulators, was up-regulated by 40 and 80 lM HMA; the same effect was recorded for ATG5/12 (Fig. 6B, left panel). By contrast, no significant changes in ATG7 expression were detected. Remarkably, HMAinduced autophagy was found to be regulated by the MAPK pathway, as supported by the phosphorylation of ERK 1/2 proteins, which activation was sustained at least up to 24 hrs (Fig. 6B,   central panel). On the contrary, JNK, which is known to be involved in several paradigms of autophagy, was poorly activated by HMA (Fig 6B, right panel).
Finally, up-regulation of Beclin-1 and ATG5 protein levels was confirmed by RT real-time PCR. As shown in Figure 6C, 40 lM HMA induced an up-regulation of BECN-1 and ATG5 gene expression. In fact, although an 8-hr treatment had no effect on BECN-1 and ATG5 gene transcription (1.03 and 1.02 A.U. respectively), a 24 hrs treatment up-regulated BECN-1 (1.53) and ATG5 (1.55) gene transcription. were treated as before and analysed by Western blot using anti-beclin 1, ATG-7, AGT5-12 and LC3 antibodies. c tubulin was used as a loading control (left panel). On middle and right panels, cells were treated as before and analysed using anti-ERK and phospho-ERK (*) antibodies (middle panel) or with anti JNK1 or phosphorylated JNK1 (*) antibodies (right panel). Under the western images the quantification of the bands is reported showing a significant increase of Beclin 1 (Means are different from each other as calculated from a one-way ANOVA test (P < 0.0001). ATG7 did not change (P = 0.165, one-way ANOVA). ATG 12 is increased in treated samples (Means are different from each other as calculated from a-one way ANOVA test (P < 0.05) The lipidated form of LC3 was also significantly increased as compared with the control (P < 0.0001) for all treatments. The same holds true for the phosphorylated form of ERK (P < 0.0001, except for etoposide, ANOVA test). JNK was significantly increased at 120 lM HMA and in the presence of rapamycin (ANOVA test P < 0.0001). (C) RT real-time PCR experiments were performed on untreated and HMA or rapamycin-treated ARPE-19 cells. HMA was used at 40 lM concentration for the indicated times, whereas rapamycin was used at 1 lM concentration for 4 hrs. GAPDH expression level was used as internal standard. Results are expressed as the mean ± SD of three independent experiments. Relative expression levels for each gene were calculated using the DDCt method. Only the 40 lM HMA results are significantly different from the others P < 0.05.

Role of autophagy in HMA-treated ARPE-19 cells
To depict the exact role of autophagy in HMA-treated ARPE-19 cells, we used either rapamycin (autophagy inducer) or 3-methyladenine (3-MA), an autophagy inhibitor that blocks autophagosome fusion with lysosome, in combination with HMA or alone. The clonogenic assay performed in ARPE-19 cells treated with HMA for 24 hrs showed low survival and no difference between untreated and treated cells (data not shown). However, a short incubation with HMA (2 hrs) alone or combined with either 3-MA or rapamycin, revealed that inhibition of autophagy by 3-MA increased cell death, while rapamycin had the opposite effect ( Fig. 7A and B). Surprisingly, after further 2 hrs of HMA exposure the situation was inversed, 3-MA bringing protection, rapamycin increasing cell death ( Fig. 7A and B). These results suggest that early activation of autophagy protects cells from HMA insult, but after few hours a prolonged activation of the autophagic process might be deleterious. The effect of HMA treatment in ARPE-19 cells on autophagic flux was evaluated by indirect immunofluorescence focusing on p62 protein localization. p62 is involved in the recognition and delivery to autophagosomes of proteins to be degraded by the autophagic machinery; its expression levels are a specific marker for autophagic flux [15]. In untreated cells, p62 showed a diffuse intracellular pattern, as in the case of rapamycin-induced autophagy (Fig. 7C). HMA triggered the re-localization of p62 giving rise to a punctate cytoplasmic pattern, as in the case of bafilomycin (an inhibitor of lysosomal ATPase, which blocks autophagosome fusion with lysosome) -treated cells, where several puncta appeared inside the cytosol. Interestingly, in HMA-treated cells p62 formed brilliant foci, possibly because of high protein damage caused by HMA in specific cellular districts.
To investigate the fusion between autophagosomes and lysosomes, we analysed specific proteins resident in these two organelles, namely, LC3 in the autophagosomes and LAMP-2 in the lysosomes (Fig. 8). Untreated cells show only a red puncture labelling corresponding to lysosomes (Fig. 8 upper row, merge, inset). In ARPE-19 cells driven to autophagy by aminoacid depletion (W/O aa), a complete overlapping of both proteins was visualized (see merge and inset). This was not the case for bafilomycin-and HMA-treated cells, in which the proteins did not colocalize (see Fig. 8, BAF and HMA rows), supporting the hypothesis that HMA inhibits the fusion of autophagosome with lysosome. It is worth noting that in HMA-treated cells the DAPI staining labels mainly the vesicles. This is because of the autofluorescence of HMA [15].

Discussion
The aim of the present study was to exploit an in vitro model system representative of the in vivo dangerous milieu of retinal cells, i.e. where cells have to face a lot of exogenous insults, causing often adverse growth conditions. Among the deleterious situations recorded in vivo, changes in the pH occur, which can be reproduced in vitro by the use of HMA, a NHE inhibitor belonging to the amiloride family, which is a potent inhibitor of Na + /H + exchangers through the plasma membrane, able to induce apoptosis in most cells [7].
Adult retinal pigmented epithelial-19 cells proved to be extremely resistant to HMA treatment, and a similar cell response was recorded also for etoposide, forcing the concept, supported by other's results, that these cells are very resistant to stress [16]. It is worth noting that etoposide activates classical (caspase-dependent) apoptosis in most cells, while HMA triggers caspase-independent, apoptotic-like cell death [7].
We show here that ARPE-19 cells overcome cell death by activating autophagy. In the first 2 hrs of HMA exposure, autophagy is efficient in protecting cells from death. Thereafter, autophagy is impaired, as indicated by a p62 accumulation, and this protective mechanism becomes the executioner of cell death. This switch in autophagy role as a function of time for a single paradigm is shown here for the first time.
Our results show that, depending on concentration, HMA has two different effects. Up to 80 lM, an autophagic response is observed, but at higher concentrations (120 lM) a modest activation of caspase-3, AIF translocation and LEI conversion into L-DNase II have been detected, suggesting that multiple cell death pathways are simultaneously triggered as result of this extremely high dose.
Given that the apoptotic response (caspase activation; Fig. 2A) occurs at very high HMA concentration, we focused our attention on the events occurring at lower concentrations, possibly involved in cell resistance. The main response of these cells to HMA concentration ranging from 20 to 80 lM is the activation of the autophagic process that, likely, protects ARPE-19 cells (Fig. 6). Actually, the clonogenic test showed that this is true in the first 2 hrs of HMA Fig. 8 Lysosome, autophagosome fusion in HMA treated cells. ARPE cells were cultured in normal medium (untreated), in a complete medium without aminoacids (to induce autophagy), in the presence of 50 nM bafilomycin (to inhibit the lysosomal proton pump) or 40 lM HMA. Immunofluorescence was performed using DAPI staining (blue), anti-LC3 (green) and anti-LAMP 2 (red) antibody. Insets in the 'merge' panel represent a zoomed region of the merged picture. In 'untreated cells' only lysosomes (red) show a punctuate staining. In the absence of amino acids (W/O aa) red and green stained (lysosomes and autophagosomes respectively) merged. These stains are not overlapped in bafilomycin-treated cells neither in HMA-treated cells. To note: given that HMA is fluorescent in UV conditions autophagosomes appear white in the merged pictures, lysosomes are red. The nucleus stained with DAPI is less fluorescent than the vesicles and appears artificially unlabelled. Scale bar: 25 lM.  (Fig. 7A). In this condition, as expected, the activation of autophagy by rapamycin increases cell viability, while inhibition by 3-MA increases cell death. This protection is probably a result of the sequestration of damaged organelles/proteins into autophagosomes (Fig. 5). It is worth noting that in ARPE-19 cells we detected also HMA in autophagosomes, indicating that the compound itself is sequestered in the vesicles [14] (Fig. 5A). Western blot analysis of molecular actors of autophagy supports the activation of this process; in fact, LC3 is transformed in LC3-II, this event being a marker of autophagy; furthermore, in situ experiments confirm the formation of autophagosomes after HMA treatment (Fig. 6). Moreover, overexpression of beclin-1 suggests that this HMA-triggered pathway of autophagy is beclin-dependent. Stress kinases ERK 1/2 and JNK 1 are also activated (even if JNK1 activation is very low if related to rapamycin-treated cells; Fig. 6B). In this regard, it has been shown that a transient activation of ERK may induce autophagy, although its sustained activation might be deleterious for a correct autophagy execution [17] by inhibiting autophagosome maturation, giving rise to giant autophagosomes unable to fuse with lysosomes, thus blocking the final step of autophagy. It is possible that, in our biological system, ERK protein is at least in part involved in the formation of large autophagosomes as well as in the possible inhibition of their maturation, as indicated by p62 accumulation within the cytosol (Fig. 7B) and the lack of fusion between phagosomes and lysosomes (Fig. 8).
No signs of caspase-dependent cell death have been detected after HMA treatment, even if different features of cellular damage such as nuclear and endoplasmic reticulum membrane dilatation and chromatin condensation have been observed using electron microscope analysis (Fig. 5B), confirming the notion that the ARPE-19 cell line is extremely refractory to caspase-dependent cell death induction.
It is interesting to note that PARP-1, a 'guardian of the genome' with a crucial role in maintaining DNA integrity under stress condition [18] is overexpressed (Fig. 2C) and hyperactivated (Fig. 3A) after HMA treatment. At the same time, we found the activation of L-DNase II (Fig. 4), an endonuclease [12,13] which is in charge of DNA degradation to promote the dismantling of the nucleus of the dying cell and could mediate PARP-1 activation, as we already reported [12]. In our study, the mechanism of activation of PARP-1 seems quite different from the mechanism described by Muñoz-Gamez et al. [19], who showed that PARP-1 is activated by the damage induced in DNA and autophagy is abolished in PARP-1 À/À cells, suggesting a role of this enzyme in the regulation of the autophagic process [20]. Here, PARP-1 activation might be just a consequence of L-DNase II activity, which is not very high because no DNA degradation is visible in agarose gels.
PARP-1 activation and consequent PAR synthesis have been related to a caspase-independent cell death pathway called parthanatos, in which PAR stimulates AIF translocation from mitochondria to the nucleus, where it induces DNA condensation and cell death (Fig. 3B) [10]. This could be the case for the cell death observed after 120-lM HMA treatment, in agreement with previous data obtained in our laboratory showing that caspase-independent pathways are preferentially activated in ARPE-19 cells [21].
Taken together, our data allow the demonstration that under cellular stress ARPE-19 respond by triggering autophagy. This response is not completely efficient in protecting cells because of the pleiotropic effects of HMA, possibly unrelated, being able to induce autophagy, on the one hand, and to block it, on the other hand [22]. In this scenario, sustained autophagy activation has a deleterious effect on cell viability and its forced and prolonged activation could switch its role from cell rescue to PCD type II. Beyond the time-dependent role of autophagy in manipulating cell death, our data support the idea that autophagy can be a general response after drug treatment. Actually, this behaviour is also observed at a lesser extent with etoposide, suggesting that induction of autophagy might be a general response of RPE to stress and the most important pathway involved in cell resistance to adverse conditions and toxic stimuli. Experiments in progress are devoted to extend our analysis to in vivo pathological conditions.
It is worth noting that lysosomes play a major role in RPE physiology because of their involvement in the management of photoreceptor waste. Actually, it has been shown that N-retynildiene-Nretynilethanolamine, a lipofuscin pigment causes inhibition of the proton pump and lysosomal disfunction [23].This is also the case for the lysomotropic agent chloroquine [24]. In addition, mice deficient in Nrf2, an antioxidant gene develop deregulated autophagy and cardinal features of age-related macular degeneration, including drusen formation [25]. Thus, the investigation of the regulation of the autophagic process in RPE cells could be a clue point for the understanding of the physiopathology of age-related macular degeneration and the development of new therapeutic approaches.