Differential Mechanisms of Conjunctival Cell Death Induction by Ultraviolet Irradiation and Benzalkonium Chloride

Abstract

purpose. To determine the molecular mechanisms of conjunctival cell death on exposure to the quaternary ammonium preservative benzalkonium chloride (BAC) and ultraviolet (UV) irradiation.

methods. Chang conjunctival cells, either wild-type or stably transfected with various constructs encoding antiapoptotic molecules or transiently transfected with siRNA targeting the beclin-1 gene, were exposed to BAC or UV radiation Cell death was analyzed morphologically with fluorescence and electron microscopy, and molecular mechanisms of death were studied by using immunofluorescence, cell fractionation, caspase substrates, and immunoblot analysis, with or without immunoprecipitation. The main results were controlled in IOBA-NHC cells.

results. Both agents induced cytochrome c release from the mitochondria, caspase activation, and nuclear chromatin condensation, suggesting caspase-dependent apoptosis. These events are prevented by stable expression of Bcl-2 protein. Both agents also induced a redistribution of Fas in plasma membrane rafts and the Fas-ligand–independent formation of a death-inducing complex leading to caspase-8 activation. Stable expression of either a dominant negative construct of Fas-associated death domain (FADD) or the long or short isoform of FADD-like interleukin-1-β–converting enzyme inhibitory protein (FLIP) inhibited caspase-8 activation in response to both UV radiation and BAC. However, these proteins, as well as permeant peptides and baculovirus p35 caspase-inhibitors, delayed more efficiently the UV irradiation-induced than the BAC-induced nuclear chromatin condensation. BAC specifically activated a caspase-independent pathway by inducing the mitochondrial release of apoptosis-inducing factor. BAC-treated cells contain autophagosomes/autolysosomes, a characteristic feature of autophagy, and siRNA-mediated downregulation of the beclin-1 gene, whose product is crucial for autophagy, increases BAC toxicity.

conclusions. UV irradiation induces typical, caspase-dependent cell death, whereas death induced by BAC associates features of caspase-dependent and –independent apoptosis counteracted by an autophagic process.

Conjunctival and corneal epithelia can undergo cellular damage when exposed to preservative-containing topical drugs. ^{1 2 3} Benzalkonium chloride (BAC), the most commonly used preservative in ophthalmic solutions, is a quaternary ammonium molecule that causes morphologic disruption of the corneal epithelium at high concentrations and induces apoptosis of Chang’s conjunctival cells, ⁴ in part through induction of reactive oxygen species. ⁵ Chronic use of preservative is responsible for apoptosis of conjunctival cells and conjunctival inflammation that has demonstrated negative effects, (e.g., on glaucoma surgery efficacy). ^{6 7} Ultraviolet (UV) irradiation can also be toxic to these cells, leading to photoconjunctivitis, photokeratitis, and pterygium. UV irradiation are potentially involved in oxidative stress involved in ocular surface disease and pterygia. ⁸ In mammalian cells, this agent activates a complex signaling network that implies radical oxygen species (ROS), ⁹ DNA damage, activation of transcription factors, ¹⁰ cell surface receptor changes, ¹¹ and activation of a raft-associated acid, sphingomyelinase, ¹² and various kinases. ¹³ Cells either repair the damage or, when unable to repair, activate the death program. ¹⁴ Several studies suggest that UV-induced cell death involves ligand-independent activation of death receptors. ^{14 15 16 17}  

The most studied mode of cell death is caspase-dependent apoptosis, which involves two main pathways. The intrinsic pathway requires a mitochondria-dependent step through outer mitochondrial membrane permeabilization, leading to cytosolic release of proapoptotic molecules. One of these molecules is cytochrome c which triggers caspase-9 activation in the so-called apoptosome. In turn, caspase-9 activates effector enzymes such as caspase-3. The extrinsic pathway starts with engagement of plasma membrane death receptors such as Fas (CD95/Apo-1), tumor necrosis factor (TNF), and TNF-related apoptosis-inducing ligand (TRAIL) receptors that recruit the adaptor molecule Fas-associated death domain (FADD). In turn, FADD recruits and activates caspase-8 in the death-inducing signaling complex (DISC). Caspase-8 either directly activates the caspase cascade or connects the extrinsic to the intrinsic pathway through cleavage of Bid. ¹⁸  

Permeabilization of the outer mitochondrial membrane can also lead to caspase-independent apoptosis through release of apoptosis-inducing factor (AIF) and endonuclease G. ^{19 20 21} On apoptosis induction, AIF translocates to the cytosol and to the nucleus, ²² binds to DNA, and causes chromatin condensation and DNA degradation. ²³  

Another cell death mechanism is autophagy, which is characterized by the appearance of multiple-membrane cytoplasmic vacuoles, the autophagosomes, engulfing bulk cytoplasm and/or cytoplasmic organelles. ²⁴ Then autophagosomes fuse with lysosomes to become autolysosomes, where their content is destroyed by the lysosomal hydrolases. Boundaries between autophagy and apoptosis remain unclear. ²⁵ Proteins such as death-associated protein (DAP) kinases can induce both apoptosis and autophagy, proteins involved in autophagy such as Beclin-1 (the orthologue of yeast Apg6) can interact with antiapoptotic Bcl-2 proteins, and the Bcl-2 proteins regulate both apoptosis and autophagy. ²⁶  

The present study was undertaken to determine the molecular mechanisms of cell death induced by BAC and UV irradiation in conjunctival cells cultured in vitro. ²⁷ Our goal was to gain a better understanding of the pathophysiology of photoconjunctivitis, photokeratitis, and pterygium, sometimes induced by exposure to these toxic agents. Our results showed that UV irradiation and BAC did not trigger conjunctival cell death in the same manner. We discuss the potential for therapeutic manipulation of the activated death pathways to prevent the toxic effects of these agents. 

Materials and Methods

Cell Culture and Transfections

Chang cells (Wong-Kilbourne clone 1 to 5c-4) were obtained from the American Type Culture Collection (Manassas, VA) and cultured in standard conditions (5% CO₂, 95% humidified air, 37°C) in Dulbecco’s minimum essential medium (DMEM) supplemented with HEPES and Glutamax (Invitrogen, Cergy Pontoise, France), 10% (vol/vol) fetal bovine serum (BioWhittaker Cambrex, Fontenay sous Bois, France), and 1% penicillin, streptomycin, and amphotericin B (Fungizone mix; BioWhittaker Cambrex). IOBA-NHC cells were kindly provided by Yolanda Diebold (IOBA-University of Valladolid, Spain) and were cultured in DMEM/F12 culture medium (Invitrogen) supplemented with 10% (vol/vol) fetal bovine serum (BioWhittaker Cambrex), human EGF (2 ng/mL; Roche Applied Science, Meylan, France), bovine insulin (1 μg/mL; Sigma-Aldrich, St. Quentin Fallavier, France), hydrocortisone (0.5 μg/mL; Sigma-Aldrich) and 1% penicillin, streptomycin, and amphotericin B (BioWhittaker Cambrex). Cells were seeded at 80% confluence for 24 hours before treatment with 4 μg/mL BAC (Thea, Clermont-Ferrand, France) or irradiation with a 254 nm UV lamp at 30 J/m² (Merck, VWR International, Fontenay-sous-Bois, France). The pTarget vector (Promega, Charbonnière, France) containing the baculovirus p35 cDNA (kindly provided by Jean-Claude Ameisen, INSERM 552, Paris, France), psFF containing full-length human Bcl-2 cDNA (kindly provided by Jacqueline Bréard, INSERM 461, Chatenay-Malabry, France) and pcDNA3 (Invitrogen) containing AIF antisense cDNA (kindly provided by Guido Kroemer, CNRS 8125, Villejuif, France) were transfected in a mixture of 4 μL transfection reagent (FuGENE 6; Roche Applied Science, Meylan, France) and plasmid (1 μg) followed by selection of geneticin-resistant cells. Retrovirus production and cell transduction by control and FADD-DN, FLICE inhibitory protein (FLIP)_S and FLIP_L-containing viruses were performed as previously described. ²⁸  

Antibodies and Chemical Reagents

We used rabbit polyclonal antibodies (Abs) recognizing caspase-3 active fragments (Cell Signaling, Ozyme, Montigny le Bretonneux, France); poly(ADP-ribose)-polymerase (PARP1; Roche Molecular Biosystems); procaspase-3 and Fas (Santa Cruz, Tebu-Bio, Le Perray en Yvelines, France); procaspase-9 (BD-Pharmingen, Heidelberg, Germany); AIF (kindly provided by G. Kroemer); mouse monoclonal Abs targeting Bcl-2 (Dako, Trappes, France); caspase-8 (MBL, Cliniscience, Montrouge, France); FADD (BD-Transduction Laboratories, Heidelberg, Germany); the long and short isoforms of FLIP (Alexis Biochemicals, Illkirch, France); and DR5 for immunoblot experiments (Chemicon International, Souffelweyersheim, France); Hsc70 (Santa Cruz); cytochrome c and flotillin-1 (BD Pharmingen); mitochondrial Hsp70 (Alexis Biochemicals); Fas-L (clone NOK1; Sigma-Aldrich); Fas (clone ZB4 blocking Ab; Immunotech, Marseille, France, and clone DX2, BD Pharmingen); TRAIL, DR4, and DR5 (Alexis Biochemicals); TNF-R1 (R&D Systems, Lille, France); and caveolin 2 (clone 65, BD Transduction Laboratories). Soluble Fas-L, soluble TRAIL, TRAIL-R2-Fc, and Fas-L-Fc were collected from the supernatant of transfected cells, as described ^{28 29} (1 arbitrary unit: 1 μL of a 100-fold concentrated supernatant). We also used caspase peptide inhibitors that include z-VAD-fmk (Bachem, Voisins-le-Bretonneux, France); z-VDVAD-fmk, z-DEVD-fmk, z-IETD-fmk, z-LEHD-fmk, and z-AEVD-fmk (R&D Systems); and fluorogenic caspase peptide substrates that include Ac-DEVD-AMC, Ac-LEHD-AFC, and z-IETD-AMC (Biomol, Plymouth Meeting, PA). AMC (7-amino-4-methylcoumarin) and AFC (7-amino-4-trifluoro-methylcoumarin) released from the substrate were excited at 380 and 400 nm to measure emission at 460 and 505 nm, respectively. Protein concentration in cell lysates was determined by using a protein assay kit (DC kit; Bio-Rad, Ivry-sur-Seine, France). 

Viability Assays

Nuclear chromatin condensation is a morphologic characteristic of apoptosis and was identified by staining trypsinized cells with 10 μg/mL Hoechst 33342 (Sigma-Aldrich) for 10 minutes at 37°C. The percentage of cells with condensed chromatin was determined by analyzing 100 cells in triplicate. To measure cell viability, we also used a methylene blue assay in which washed cells were incubated for 5 minutes in ethanol and dried 15 minutes at room temperature, then incubated for 15 minutes in methylene blue dye (100 mM boric acid, 25 mM di-sodium tetraborate, 120 mM NaCl, and 0.5 mg/mL methylene blue) and washed. HCl (0.1 M) was added before the cells were completely dry, and absorbance was read at 630 nm. Cell viability was calculated as the ratio of absorbance in the treated sample to absorbance in the control sample × 100 and expressed as a percentage. 

Caspase Activity Measurement

The cells were washed in DPBS and incubated 30 minutes in lysis buffer (150 mM NaCl, 50 mM Tris-HCl [pH 8.0], 0.1% SDS, 1% Nonidet P-40, 0.5% deoxycholate, and 1 mM phenylmethylsulfonyl fluoride) at 4°C. After centrifugation (10,000g, 20 minutes, 4°C), supernatant was collected and 50 μg of proteins were incubated in a buffer assay (100 mM HEPES [pH 7.0], 1 mM EDTA, 0.1% CHAPS [3-[3-cholamidopropyl]dimethylammonio-2-hydroxy-1-propanesulfonate], 10% glycerol, and 20 mM dithiothreitol) in the presence of 100 μM fluorogenic caspase peptide substrate. Fluorescence was monitored continuously at 37°C for 30 minutes in dual-luminescence fluorometer (MicroTek OS; Bio-Tek Kontron Instruments, Winooski, VT). Enzyme activities were determined as initial velocities expressed as relative intensity per minute per milligram. 

Flow Cytometry Analyses

Analysis of death receptors and their ligand at the cell surface was performed by incubating the cells for 1 hour at 4°C with either a specific Ab or its isotype-matched control (Dako). Abs were diluted in DPBS containing 01% NaN3 and 1% bovine serum albumin (BSA). After two washes, cells were incubated with a 488-Alexa goat anti-mouse Ab (Molecular Probes Europe BV, Leiden, The Netherlands). Bcl-2 expression was measured by incubating saponin-permeabilized cells for 1 hour at room temperature with appropriate Abs. Analysis was performed by using flow cytometry (FacScan; BD Biosciences). 

Immunofluorescence Staining

The cells were seeded on glass coverslips for 24 hours, treated, and fixed by incubation for 15 minutes at room temperature in 2% paraformaldehyde. After three washes, the cells were preincubated in DPBS with 5% BSA for 15 minutes at room temperature and incubated with the primary Ab diluted in DPBS with 2% BSA, with or without 0.1% saponin, for 90 minutes at room temperature. After washing, the cells were incubated for 40 minutes at room temperature with 488-Alexa goat anti-mouse or anti-rabbit Abs (Invitrogen-Molecular Probes, Eugene, OR). The nuclei were labeled with Hoechst 33342, and analysis was performed with a fluorescence microscope (Nikon, Champigny, France). 

Electron Microscopy

Briefly, the cells were fixed for 30 minutes at 20°C in 4% paraformaldehyde and 1.5% glutaraldehyde in 0.1 M Störensen buffer (pH 7.3) before they were washed in Störensen buffer, stained with 1% toluidine blue, and embedded in 2% agar. The cells were fixed for 1 hour in 1% OsO₄ in the dark and embedded in Epon before analysis of ultrathin sections by transmission electron microscope (H-7500; Hitachi, Velizy, France), at 80 kV. 

Autophagy Vacuole Staining

The cells were incubated for 20 minutes at 37°C in complete medium containing 2.5 μg/mL acridine orange or 0.05 mM monodansylcadaverine (Sigma-Aldrich), then washed and incubated for 10 minutes at 37°C in complete medium. Analysis was immediately performed either by fluorescence microscopy or flow cytometry. 

Immunoblot Analysis

Cells were lysed for 15 minutes at 4°C in boiling buffer (10 mM Tris-HCl [pH 7.4] 1% SDS, 1 mM sodium vanadate, and 1:50 Complete protease inhibitor mixture; Roche Applied Science). The viscosity of the sample was reduced by sonication. Mitochondrial and cytosolic fractions were prepared as described. ³⁰ Briefly, cells were washed in cold DPBS and resuspended in hypotonic buffer A (sucrose 250 mM, HEPES [pH 7.4] 20 mM, KCl 10 mM, MgCl₂ 1.5 mM, EDTA 1 mM, EGTA 1 mM, dithiothreitol 1 mM, and 1:50 protease inhibitor). Homogenization was performed until 50% of cells were trypan blue positive. Nuclei were pelleted by a 10-minute 750g spin at 4°C. The supernatant was centrifuged at 10,000g at 4°C for 25 minutes, and pellet was suspended in buffer A (mitochondrial fraction). The supernatant was spun at 100,000g at 4°C for 1 hour and supernatant was conserved (cytosolic fraction). Fifty micrograms of proteins were boiled for 5 minutes in loading buffer (125 mM Tris-HCl [pH 6.8] 10% β-mercaptoethanol, 4.6% SDS, 20% glycerol and bromphenol blue), separated on SDS-polyacrylamide gel and transferred onto nitrocellulose membrane (Schleicher & Schuell, Dassel, Germany). Membranes were incubated overnight at 4°C with the primary Ab in TPBS (DPBS with 0.1% Tween) and washed three times before incubation with an anti-mouse or anti-rabbit Ab coupled with horseradish peroxidase (Jackson ImmunoResearch Laboratories, West Grove, PA). Protein detection was performed by using an enhanced chemiluminescence detection kit (Santa Cruz). 

Immunoprecipitation

The cells (100 × 10⁶) were incubated for 15 minutes at 4°C in lysis buffer (150 mM NaCl, 50 mM Tris-HCl [pH 7.4], 0.1% SDS 1% Nonidet P-40, 0.5% sodium deoxycholate) containing protease inhibitor mixture tablets (Complete; Roche Applied Science). After centrifugation for 10 minutes at 10,000g at 4°C, Fas-L-Fc (3 AU/mL) was added for 1 hour at 4°C before the immune complexes were precipitated by 30 μL of mixed Sepharose 6B (Sigma-Aldrich) and G-Sepharose (GE Healthcare, Piscataway, NJ) for 4 hours at 4°C. The precipitate was washed 4 times with lysis buffer and boiled 5 minutes in loading buffer before immunoblot analysis. 

Isolation of Rafts

Briefly, 100 × 10⁶ cells were incubated for 30 minutes at 4°C in 1 mL MES buffer (25 mM MES, 150 mM NaCl, and protease inhibitor mixture) with 1% Triton X-100. The lysates were diluted in 1 mL of MES buffer and 2 mL of MES buffer containing 80% sucrose, placed at the bottom of a linear sucrose gradient (5% and 40%), and centrifuged at 250,000g for 20 hours at 4°C. One-milliliter fractions were collected from the top to the bottom of the gradient. Fractions 1 to 8 (450 μL) were precipitated by adding trichloroacetic acid (TCA, 5%) and centrifuged 10 minutes (8000g, 4°C) and the precipitate was resuspended in 60 μL loading buffer (62.5 mM Tris-HCl [pH 6.8], 2% SDS, 10% glycerol, 5 mM EDTA, 100 mM dithiothreitol, 2M urea, and 0.02% bromphenol blue). Thirty microliters of these fractions and 10 μL of the others were subjected to SDS-polyacrylamide gel electrophoresis and immunoblot. 

RNA Extraction and RT-PCR

RNA was prepared (NucleoSpin RNA II kit; Macherey-Nagel, Hoerd, France), and cDNA was synthesized from 1 μg total RNA by the use of Moloney murine leukemia virus reverse transcriptase (Invitrogen). PCR was performed with polymerase (AmpliTaq Gold; Roche Applied Science) on a thermocycler (iCycler; Bio-Rad). Specific primers were synthesized by Invitrogen and ProLigo Primers and Probes (Boulder, CO), respectively: beclin-1 (forward: TCTGGGACAACAAGTTTGAC; reverse: CCACTTAA GATTCGTCAGCA), β2-microglobulin (forward: ACCCCCACTGAAAAAGATGA; reverse: ATCTTCAAACCTCCATGATG) and baculovirus p35 (forward: ATGGATTATAAAGATG ATGATGATAAATGTGTAATTTTT; reverse: TTATTTAATTGTGTTTAATATTACATTT TTGTTGAGTGC). The PCR products were separated on 8.8% SDS-polyacrylamide gel and stained with ethidium bromide before fluorescence analysis by video camera imaging (Photomat software; Microvision Instruments, Evry, France). 

RNA Interference

Two small interfering (si)RNAs specific for the beclin-1 gene were used, starting at positions 189 (CUCAGGAGAGGAGCCAUUU) and 1206 (GATTGAAGACACAGGAGGC) from the ATG codon (Qiagen, Courtaboeuf, France). As the control, we used a scrambled siRNA from B189 (AGCAGCUGACCGGAUUAGU). Exponentially growing cells were treated with a mixture of oligofectamine reagent (Invitrogen) and siRNA for 4 hours before culture for 24 to 48 hours. 

Results

Effect of UV Irradiation and BAC on Nuclear Chromatin Condensation and Caspase Activation in Chang Cells

Exposure of Chang cells to UV radiation and BAC induced the condensation of their nuclear chromatin (Fig. 1A) . The percentage of cells with condensed chromatin increased with time (Fig. 1B)and dose (not shown) of each agent. This event was associated with the activation of several caspases. Western blot analyses of UV- and BAC-treated cells showed a time-dependent decrease in procaspase-3, -8, and -9 expression, which correlated with the accumulation of their active fragments (Fig. 1C) . Caspase activation, which was further confirmed by time- and dose-dependent cleavage of several specific peptide substrates (not shown), was associated with the release of cytochrome c from the mitochondria into the cytosol (Figs. 1D 1E) . The caspase inhibitor z-VAD-fmk failed to prevent cytochrome c release from the mitochondria. These events were associated to the proteolytic cleavage of poly(ADP-ribose) polymerase 1 (PARP1), a well-known target of effector caspases (Fig. 1F) . Overexpression of the anti-apoptotic Bcl-2 protein prevented the release of cytochrome c and the condensation of the nuclear chromatin (Fig. 1G) . Altogether, these data suggested that both UV radiation and BAC could induce caspase-dependent apoptosis of Chang cells. 

Effect of UV Irradiation and BAC on the Formation of a Fas-Containing DISC

In cells undergoing apoptosis, caspase-8 can be activated either upstream of the mitochondria, in the DISC, or downstream of the mitochondria in the caspase cascade activated by cytochrome c ³¹ The observation that caspase-8 was activated in Chang cells on UV irradiation and BAC exposure led us to explore whether this activation took place at the level of a death receptor. We detected the expression of Fas and the two agonistic receptors of TRAIL, DR4, and DR5 at the surface of untreated cells. These cells did not express Fas-ligand, TRAIL, or TNFR1 (Fig. 2A) . Exposure to UV radiation or BAC induced a decrease in the expression of Fas, DR4, and DR5 at their surface, without inducing the appearance of the other studied receptors (Fig. 2A) , and triggered a redistribution of Fas, FADD, procaspase-8, and, to a lesser extent, DR5, in caveolin-2-rich cell fractions (Fig. 2B) . Immunofluorescence and confocal analyses indicated that Fas partly colocalized with flotillin in UV- and BAC-treated cells (Fig. 2C) . 

Coimmunoprecipitation experiments demonstrated that both UV radiation and BAC induced the formation of a Fas-containing DISC including Fas, FADD, and procaspase-8 in Chang cells (Fig. 3A) . Stable expression of a dominant negative construct of FADD or the long or the short isoform of FLIP (Fig. 3B)prevented the appearance of IETD-AMC cleavage activity, suggesting caspase-8 activation, in lysates of UV- and BAC-treated cells (Fig. 3C) . Of note, these constructs delayed UV-induced nuclear chromatin condensation without affecting that induced by BAC (Fig. 3D) . Neither ZB4 anti-Fas blocking Ab (2 μg/mL), or NOK1 anti-Fas-L Ab (5 μg/mL), or TRAIL inhibition using TRAIL-R2-Fc (50 AU/mL) could prevent the nuclear chromatin condensation induced by the two studied agents (Fig. 3E) . These observations suggested that the extrinsic pathway to caspase-8 activation, that was activated in both situations, played a central role in UV-induced chromatin condensation, whereas an alternative pathway might be activated by BAC to trigger this event in Chang cells. 

Contribution of Apoptosis-Inducing Factor to BAC-Induced Nuclear Chromatin Condensation

To further analyze the role of caspases in these pathways, we introduced the baculoviral p35 caspase inhibitor in Chang cells by stable transfection (Fig. 4A) . Caspase activation induced by exposure to UV radiation or BAC was decreased in p35-expressing Chang cells compared with those transfected with an empty vector (Fig. 4B) . Expression of p35 also prevented the cleavage of PARP1 in cells treated with UV radiation or BAC (Fig. 4C)but did not prevent the condensation of nuclear chromatin as efficiently in BAC-treated as in UV-treated cells (Fig. 4D) . The differential consequences of caspase inhibition on UV- and BAC-induced chromatin condensation were further confirmed by using permeant peptide inhibitors of caspases. These peptides strongly decreased nuclear chromatin condensation induced by exposure to UV irradiation, whereas their ability to prevent that induced by BAC was less strong (Fig. 4E) . Such a difference was still observed with the lowest tested concentrations of z-VAD-fmk caspase inhibitor (Fig. 4F) . The low sensitivity of BAC-induced chromatin condensation to caspase inhibition suggested an alternative pathway activated by the drug. Accordingly, whereas UV irradiation hardly induced the redistribution of AIF from the mitochondria to the nucleus, such a redistribution was observed in several BAC-treated cells (Fig. 4F) . Transient expression of an antisense construct targeting AIF (Fig. 4H , inset) partially prevented BAC-induced without affecting UV-induced nuclear chromatin condensation. z-VAD-fmk and AIF-targeting antisense construct demonstrated an additive effect in protecting Chang cells from BAC-induced nuclear changes, which was not observed after UV irradiation (Fig. 4H) . 

BAC Induction of Autophagy

Using electron microscopy, UV radiation was observed to trigger nuclear chromatin condensation and cell fragmentation in apoptotic bodies without any vacuoles in the cytoplasm (Fig. 5A) . In contrast, the majority of BAC-treated cells contained several multimembrane vacuoles (Fig. 5A) , suggesting an autophagic process. Cell staining with acridine orange and monodansylcadaverine confirmed the formation of these autophagosomes-autolysosomes (Figs. 5B 5C) . A decrease in beclin-1 gene expression, whose product was identified as a molecular actor of autophagy, ³² was obtained by using two distinct siRNAs (Fig. 5D)and increased BAC-induced whereas its effect of UV-induced cell death remained limited (Fig. 5E) . A decrease in beclin-1 gene expression also prevented BAC-induced autophagy (Fig. 5F)and enhanced BAC-induced nuclear chromatin condensation (Fig. 5G) . These data suggested that an autophagic process was activated in BAC-treated cells to counteract the caspase-dependent and -independent pathways to death. 

Activation of Death Pathways in IOBA-NHC Cells

As the reliability of the Wong-Kilbourne derivative of Chang conjunctiva cells to study the pathophysiology of human conjunctiva has been discussed, we checked whether the identified pathways were also activated in IOVA-NIH cells. In these cells, both UV and BAC induced apoptosis in a dose- (Fig. 6A)and time- (Fig. 6B)dependent manner. A DEVD-AMC cleavage activity appeared on both stimuli, and this activity was inhibited by the large-spectrum caspase inhibitor z-VAD-fmk (Figs. 6C 6D) . AIF antisense contructs failed to prevent UV-induced apoptosis whereas it partially decreased the level of BAC-induced apoptosis, and this effect was enforced by addition of z-VAD-fmk (Fig. 6D) . BAC but not UV irradiation induced the formation of acridine orange–labeled large vacuoles (Fig. 6E)and beclin-1-targeting siRNAs slightly enhanced the level of apoptosis induced by BAC without modifying the level of UV-induced cell death (Fig. 6F) . 

Discussion

The present study indicates that UV-induced apoptosis of conjunctival cells is mainly dependent on caspase activation whereas BAC-induced conjunctival cell death is a more complex process that associates caspase-dependent and -independent nuclear chromatin changes with autophagic features. 

Several studies have shown that ligand-independent activation of death receptors, including TNFR1 and Fas, was responsible for UV-induced cell death. ^{14 15 16} Here we show that, in Chang cells exposed to UV irradiation, the Fas-containing DISC forms, leading to caspase-8 and downstream caspase activation and the appearance of apoptotic features. Ligand independence was suggested by the lack of preventive effect of either ZB4 or NOK1 Abs that efficiently prevented Fas-L-mediated apoptosis in these cells. Thus, the mechanism of UV-induced cell death in Chang cells may not significantly differ from that identified in malignant epithelial cell lines. ^{14 15 16 17}  

BAC-induced apoptosis of conjunctival cells also demonstrated caspase-dependent apoptotic features. However, caspase inhibition only partially prevented nuclear chromatin condensation, suggesting that additional pathways participate in the BAC-induced death phenotype. One of these pathways involves the flavoprotein AIF ^{19 22 23 33 34} that, on exposure to BAC, translocates from the mitochondria to the nucleus. So far, HSP70 remains the only identified intracellular protein that negatively interferes with caspase-independent, AIF-mediated chromatinolysis. ^{35 36} Strategies aiming to induce HSP70 expression may partially prevent the toxic effect of BAC on conjunctival cells whereas therapeutic approaches aiming to sensitize tumor cells to cytotoxic drug-induced apoptosis by preventing AIF/HSP70 interaction ³⁷ could potentially increase the toxic effects of BAC containing topical drugs. 

Another aspect of BAC-induced death of conjunctival cells is autophagic vesicle formation. Autophagy was initially described as an adaptive response to nutrient starvation and a mechanism of cytoplasmic homeostasis. ³⁸ Autophagy is involved in various other cellular functions, ³⁹ whereas its link with apoptosis remains controversial. Autophagy could partially prevent apoptosis by removing damaged mitochondria that may activate death pathways. ⁴⁰ Accordingly, the ability of beclin-1 gene downregulation to increase BAC-induced cell death suggested that autophagy played a protective role in BAC-treated conjunctival cells. Beclin-1 is part of a class III PI-3 kinase complex that is crucial for autophagy, suggesting that pharmacological manipulation of the PI-3K pathway could partially prevent BAC-mediated death of conjunctival cells. 

Apoptosis plays a role in the pathophysiology of acute and chronic diseases of the conjunctiva such as inflammatory ocular disorders, ^{41 42} infection by Chlamydiae psittaci ⁴³ and viruses, ⁴⁴ and exposure to toxic agents. ^{4 5} Studies have suggested that the molecular pathways leading to cell death could depend on the death inducer (e.g., Chlamydia activates a Bcl-2-regulated, caspase-independent pathway ⁴³ whereas viruses activate a Fas-L/Fas-mediated caspase-dependent pathway) ⁴⁴ The present study demonstrates that two other common toxic insults activate distinct pathways of conjunctival cell death. Identification of these molecular pathways could suggest strategies to prevent conjunctival cell damage induced by these toxic insults by inhibiting apoptosis. Current therapeutic approaches used to alleviate ocular surface disease (OSD) include the use of artificial tears, ⁴⁵ anti-inflammatory agents, ⁴⁶ and, in the setting of dry eye disease, cyclosporine. ^{47 48} Some of these molecules have demonstrated antiapoptotic effects, together with their anti-inflammatory activity. Manipulation of HSP70 and the PI-3K pathway could prevent BAC-mediated death of conjunctival cells and facilitate the chronic use of this otherwise efficient preservative. In contrast, inhibition of apoptosis could be deleterious in UV-exposed cells by preventing elimination of those in which DNA has been severely damaged. Thus, in vivo studies are now needed to determine whether manipulation of these pathways could be therapeutically useful. 

 Supported by INSERM Ligue Nationale Contre le Cancer.

 Submitted for publication November 15, 2005; revised April 20 and May 24, 2006; accepted July 28, 2006.

 Disclosure: N. Buron, None; O. Micheau, None; S. Cathelin, None; P.-O. Lafontaine, None; C. Creuzot-Garcher, None; E. Solary, None

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Figure 1.

 

UV irradiation and BAC induced nuclear chromatin condensation and caspase activation in Chang cells. (A, B) BAC and UV irradiation induced nuclear chromatin condensation in Chang cells. These cells were either left untreated or treated with BAC (4 μg/mL, □) for the indicated times or exposed to UV irradiation (30 J/m²) and the percentage of cells with condensed nuclear chromatin, suggesting apoptosis, was analyzed at the indicated times (♦) after labeling with Hoechst 33342. (A) Magnification, ×40. (B) Mean ± SD of three independent experiments. (C) Immunoblot analysis of indicated proteins in cells treated as in (A) and (B). The appearance of caspase cleavage fragments suggests activation. Hsc70 was used as a loading control. (D) Immunoblot analysis of cytochrome c (Cyt. c) in the mitochondrial (m) and cytosolic (c) fractions of Chang cells left untreated, treated with 4 μg/mL BAC for 8 hours, or exposed to UV irradiation (30 J/m²) and analyzed 4 hours later. Mitochondrial Hsp70 (mHsp70) was used as a control of cell fractions. The appearance of cytochrome c in the cytosol of conjunctival cells suggest permeabilization of the mitochondria. (E) Immunofluorescence analysis of cytochrome c in cells treated as in (D), in the absence or presence of 10 μM z-VAD-fmk Magnification, ×60. Cytochrome c is in the mitochondrial and living cells and is released in the cytosol of cells undergoing apoptosis under UV irradiation and BAC exposure. (F) Immunoblot analysis of PARP1 in cells treated as in (B), reveals proteolytic cleavage of this well-characterized caspase target. (G) Percentage of cells with condensed chromatin, as identified by Hoechst 33342 staining in control- and bcl-2-transfected Chang cells, either left untreated or treated with 4 μg/mL BAC for 16 hours or exposed to UV irradiation (30 J/m²) and analyzed 8 hours later (mean ± SD of three experiments). The mitochondria-associated Bcl-2 protein protected the cells from apoptosis, further suggesting the involvement of the mitochondria in the death processes. Inset: Bcl-2 protein expression analyzed by flow cytometry (F.I., fluorescence intensity) in control ( ) and bcl-2-transfected (□) Chang cells. Molecular mass is indicated in kDa.

Figure 2.

 

UV irradiation and BAC induced redistribution of some death receptors in raft domains of the plasma membrane. (A) Flow cytometry was used to analyze the expression of indicated proteins in Chang cells either left untreated (Co) or treated with 4 μg/mL BAC for 3 hours or exposed to UV irradiation (30 J/m²) and analyzed 3 hours later ( , isotype-matched control antibody; □, specific antibody). (B) Immunoblot analysis of indicated proteins in fractions obtained from Chang cell lysates after separation on a sucrose gradient. Cells were either untreated (Co) or exposed to UV irradiation and analyzed 1 hour later or treated with BAC for 3 hours Molecular mass is indicated in kDa. The protein caveolin was used to identify raft fractions. Fas, FADD, caspase-8, and DR5 partially colocalized with caveolin on UV irradiation and BAC treatment, but not in the control cells. (C) Immunofluorescence analysis of Fas and flotillin (another protein associated with raft domains of the plasma membrane) at the surface of Chang cells treated as in (B). Magnification × 60. Again, the colocalization of Fas and flotilin in treated cells, not in untreated cells, suggests that Fas migrates to the rafts on treatment.

Figure 3.

 

UV- and BAC-induced chromatin condensation differentially depended on the death receptor pathway. (A) Cells were collected 1 hour after UV irradiation exposure or after a 3-hour exposure to BAC at doses indicated in Figure 1Afor immunoblot analysis of indicated proteins, before (extracts) or after (IP Fas) immunoprecipitation with a Fas-L-Fc chimeric protein. Molecular mass is indicated in kDa. The results indicated that Fas, FADD, and caspase-8 associated in cells exposed to either UV irradiation or BAC, to form the DISC. (B) Immunoblot analysis of the indicated proteins in Chang cells transfected with an empty vector (Co) or a vector expressing either a dominant negative construct of FADD (FADD-DN) or the short (FLIP_S) or the long (FLIP_L) isoform of FLIP. (C) IETD-AMC cleavage activity, used to measure caspase-8 activity, was measured in the various cell populations described in (B), either left untreated (□) or exposed to UV irradiation and examined 6 hours later (▪) or treated for 16 hours with BAC ( ) at doses described in Figure 1A . (D) The percentage of cells with nuclear chromatin condensation, suggesting apoptosis, was measured after Hoechst staining at the indicated times after exposure of cells described in (B) (♦, control; ▪, FADD-DN; ⋄, FLIP_S; □, FLIP_L) to UV irradiation or treatment with BAC as in (B). (E) Chang cells were incubated for 1 hour with either nothing (□) or ZB4 anti-Fas antagonistic Ab (2 μg/mL; ) or NOK1 anti-Fas-L Ab (5 μg/mL; ▪) or TRAIL-R2-Fc chimeric protein (50 AU/mL; ) before exposure to UV radiation (analyzed 6 hours later) or BAC for 16 hours, before analyzing nuclear chromatin condensation after Hoechst staining. As the control, cells were exposed either to Fas-L (5 AU/mL) or TRAIL (25 μg/mL). Results indicate that interaction of death receptors with their respective ligand is not requested for Fas involvement in the studied death processes. In (C), (D), and (E), results are the mean ± SD of three independent experiments.

Figure 4.

 

AIF specifically contributed to BAC-induced nuclear chromatin condensation. (A) Chang cells were stably transfected with the empty (Co) or the baculovirus p35-containing pTarget vector (p35) before RT-PCR analysis of p35 gene expression. The β2-microglobulin (β2) gene was used as the control. (B–D) Chang cells described in (A) (control, □; p35, ▪) were collected 6 hours after exposure to UV irradiation or after 16 hours of BAC treatment at doses described in Figure 1Ato study the cleavage of the fluorogenic substrates Ac-DEVD-AMC, Ac-IETD-AMC, and Ac-LEHD-AFC, suggesting activation of caspase-3, -8, and -9, respectively (B), the cleavage of PARP1 as in Figure 1F(C), and the percentage of cells with nuclear chromatin condensation as in Figure 1B(D). (E, F) Chang cells were either left untreated (□) or exposed to UV irradiation and examined 6 hours later ( ) or treated for 16 hours with BAC ( ). Cells were either preincubated for 1 hour with 50 μM of indicated permeant caspase inhibitor. *P < 0.05, Mann-Whitney test (E) or cotreated with indicated concentrations of the large spectrum caspase inhibitor z-VAD-fmk (F) before measuring the percentage of cells with nuclear chromatin condensation as in Figure 1B . (G) Chang cells were either left untreated or exposed to UV irradiation and examined 4 hours later or treated for 4 hours with BAC before studying AIF redistribution in the nucleus by immunofluorescence (see the percentage of cells with nuclear AIF in one representative experiment; magnification × 60). (H) To determine the role of AIF in apoptosis induced by UV irradiation and BA, the cells were transiently transfected with pCDNA3 vector (Co) or AIF antisense-containing vector (AS), then treated as in (E) and (F) before measuring the percentage of cells with nuclear chromatin condensation. Inset: immunoblot analysis of AIF expression in cells transiently transfected with indicated vectors. When indicated, cells were cotreated with 10 μM z-VAF-fmk. Mean ± SD of at least three independent experiments are shown.

Figure 5.

 

Autophagy specifically contributes to BAC-induced cell death. (A) Chang cells were either left untreated or examined 8 hours after UV irradiation or after a 16 hour treatment with BAC as described in Figure 1Abefore morphologic examination of cells by electron microscopy. One representative cell is shown. Nuclear chromatin condensation and plasma membrane blebbing are typical features of apoptosis. BAC-treated cells also demonstrate the presence of typical autophagic vacuoles (right). (B) Chang cells were either left untreated or examined 4 hours after UV irradiation or after a 4-hour treatment with BAC, then labeled with either acridine orange (AO) or monodencyl cadaverin (MDV) and observed by fluorescence microscopy. Magnification, ×60. The increase in the size of labeled vacuoles in BAC-treated cells suggests autophagic vesicle formation. (C) Acridine orange staining was quantified by flow cytometry in Chang cells left untreated ( ) or examined 8 hours after UV irradiation or after an 8-hour treatment with BAC (□). This experiment completes the one shown in (B) by showing the increased staining. (D–G) Cells were left untreated or transiently transfected with scrambled- (Sc and ○) or beclin-1-specific (Sp1 and ▪, 189; Sp2 and ▴, 1206) siRNAs. (D) Beclin-1 gene expression was measured by RT-PCR at indicated times. β2-Microglobulin gene expression was used as the control. (E) Cells treated with siRNAs for 48 hours were exposed to UV radiation or BAC as in Figure 1A , and their viability was measured at indicated times by using a methylene blue assay (mean ± SD of three independent experiments). (F) Cells treated with siRNAs for 48 hours were either left untreated ( ) or treated with BAC for 8 hours before staining with acridine orange (□) and flow cytometry analysis to study the effect of beclin-1 expression decrease on autophagic vesicle formation. (G) Cells treated with siRNAs for 48 hours were subsequently left untreated (□) or examined 6 hours after UV irradiation ( ) or 16 hours after BAC treatment ( ) before measuring the percentage of cells with condensed nuclear chromatin (mean ± SD of three independent experiments). Altogether, these experiments confirm the autophagic process activated by exposure of Chang cells to BAC, but not to UV radiation.

Figure 6.

 

UV irradiation and BAC activated a similar death pathway in IOBA-NHC conjunctiva cells. (A) IOBA-NHC cells were either left untreated or exposed to UV radiation and analyzed 24 hours later ( ) or treated with BAC ( ) for 16 hours before the percentage of cells with condensed nuclear chromatin was measured, suggesting apoptosis. (B) Cells were either left untreated or exposed to UV irradiation (100 J/m²; ▪) or BAC (4 μg/mL; □) for the indicated times before measuring the percentage of cells with condensed nuclear chromatin. (C) DEVD-AMC cleavage activity, suggesting caspase-3 activation, was measured in untreated cells, those exposed to 100 J/m² UV irradiation and examined 8 hours later, and those treated for 16 hours with 4 μg/mL BAC ( ) at doses described in (B). Treatment was performed in the absence (□) or presence ( ) of 10 μM z-VAD-fmk. (D) Cells were transiently transfected with pCDNA3 vector (Co) or AIF antisense encoding vector (AS), then treated as in (B) (□, untreated; , UV irradiation; , BAC) before percentage of cells with condensed chromatin was measured When indicated, cells were cotreated with 10 μM z-VAD-fmk. (E) Cells were either left untreated or examined 4 hours after exposure to UV radiation or treatment with BAC, then labeled with acridine orange (AO, red) and observed by fluorescence microscopy to study autophagic vesicle formation. Nuclei were stained with Hoechst 33342 (green). One representative experiment. Magnification, ×60. AO-labeled vacuoles were enlarged in BAC-treated IOVA-NHC cells, suggesting autophagia. (F) Cells transiently transfected with scrambled (Sc) or beclin-1-specific (S1: 189; S2: 1206) siRNAs for 36 hours were subsequently left untreated (□) or examined 8 hours after UV radiation exposure ( ) or 16 hours after BAC treatment ( ) at the doses described in (B) before the percentage of cells with condensed nuclear chromatin was measured. Inset: beclin-1 gene expression measured by RT-PCR 36 hours after transfection. In all but (E), data are the mean ± SD of three independent experiments.

The authors thank Yolanda Diebold for providing IOBA-NHC cells, Frank Menetrier for his help in electron microscopy studies, and Alain Bron and Christophe Baudouin for helpful advice.