Fas- and Interferon γ-Induced Apoptosis in Chang Conjunctival Cells: Further Investigations

Abstract

purpose. Previously interferon (IFN)γ-induced apoptosis and expression of inflammation-related proteins in a human conjunctival cell line were demonstrated. The aim of this study was to further investigate the mechanisms of IFNγ-, Fas-, and cycloheximide (CHX)-induced programmed cell death, with special attention to the role of transcriptional factors NF-κB and STAT1.

methods. In a human conjunctival cell line (Chang conjunctival cells) apoptosis was induced with 500 ng/ml anti-Fas antibody (anti-Fas ab) alone (24 or 48 hours) or, as previously reported, with 300 U/ml of human recombinant IFNγ alone (48 hours). To study the role of IFNγ on Fas-induced apoptosis, cells were treated first with IFNγ at 30 U/ml during 24 hours (nontoxic dose), and then anti-Fas ab was applied for 24 hours. Moreover, to study the influence of CHX on Fas- and IFNγ-induced apoptosis, cells were treated for 24 hours with 300 U/ml IFNγ together with a nontoxic concentration (1 μg/ml) of CHX, or with 500 ng/ml anti-Fas ab together with 1 μg/ml CHX (24 hours). After treatment, cell viability (neutral red assay), mitochondrial membrane potential (rhodamine 123 assay), chromatin condensation (Hoechst 33342 assay), and the index Hoechst/neutral red were studied by cold light microplate cytometry. The apoptotic process was sought for by contrast phase microscopy and DAPI staining and was confirmed by immunoblotting of PARP. Activation of caspase-3 (CPP32) and caspase-8 were investigated by Western blot analysis. NF-κB and STAT DNA-binding activities were studied by electrophoretic mobility shift assays (EMSA).

results. After 24 and 48 hours of treatment with anti-Fas ab alone, 15% to 20% and 30%, respectively, of apoptotic cells were observed. When anti-Fas sera were applied after IFNγ pretreatment or together with CHX, 50% to 80% of cells demonstrated morphologic characteristics of programmed cell death. Apoptosis was confirmed by a cleavage of PARP and CPP32, by caspase-8 activation, and by an index Hoechst/neutral red greater than one. All these modifications were preceded by a decrease in mitochondrial membrane potential. EMSA revealed that NF-κB was activated after IFNγ and anti-Fas ab treatments and inhibited after CHX treatment. STAT1 was strongly activated after IFNγ treatment and only in a minor degree after anti-Fas ab treatment. STAT1-binding activity persisted after CHX treatment.

conclusions. The relative resistance of Chang cells toward Fas-induced apoptosis could be related to the activation of NF-κB. IFNγ-induced programmed cell death preferentially involves the activation of STAT1 that counterbalances NF-κB antiapoptotic effects. In fact, Fas-induced apoptosis was potentiated by IFNγ or CHX treatments. These results suggest that NF-κB activation could maintain cell viability as well as participate in IFNγ-induced inflammatory modifications, whereas STAT1 activation could provide, in this model, a proapoptotic signal.

Programmed cell death, or apoptosis, is an essential process for the normal development, homeostasis, and maintenance of a multicellular organism, for its defense and for removal of individual cells without damage to anatomic and functional structures. ^{1 2 3 4} Various stimuli can induce programmed cell death. One of them is interferon (IFN)γ, an inflammatory cytokine and a lymphocyte effector molecule implicated in many different types of immune responses (inflammation or graft rejection), ^{5 6 7} and involved in the pathogenesis of ocular surface inflammatory diseases, such as Sjögren’s syndrome. ^{8 9 10 11 12 13} IFNγ can kill cells by apoptosis as was demonstrated in several in vitro models. ^{14 15 16} However, the biological role of IFNγ-induced programmed cell death is still not well defined, and the precise correlation between IFNγ-mediated inflammatory changes and apoptosis has not been well established, ^{17 18} probably because of the pleiotropism of IFNγ-induced effects and the great number of signaling and effector proteins downstream of IFNγ membrane receptor. ^{19 20} The sensitivity of different cell types to IFNγ-induced apoptosis is extremely variable, thus signing the complexity of intracellular signal transduction pathways. In most cell studies so far, IFNγ was shown to principally activate signal transducer and activator of transcription (STAT) family, especially one of its members, namely STAT1. ^{20 21} The dimer of activated STAT1 could bind to the short stretches of DNA, called gamma interferon activation site (GAS), and there could then be a rapid transcriptional induction of several genes and their products such as ICAM-1, HLA DR, or apoptotic proteases (caspases). ^{22 23} A requirement of STAT1 activation in IFNγ-induced programmed cell death is now clearly established. ^{23 24 25 26 27} Other IFNγ-induced signal transduction pathways are less known, for instance that implying NF-κB activation. ²⁸  

NF-κB is a transcriptional regulatory protein complex participating in the regulation of gene expression of many modulators of inflammatory, proliferative, and immune reactions. It is activated in response to the great number of stimuli, most of which represent pathogenic stress. Among the many target genes of NF-κB, some are involved in apoptosis, such as p53, c-myc, Fas ligand or interleukin-1–converting enzyme (ICE or caspase-1). ^{29 30 31} On the contrary, in other cellular systems activation of NF-κB was found to block the action of some proapoptotic proteins such as caspase-8, TRAF1 (tumor necrosis factor [TNF] receptor–associated factor 1) and to suppress TNF-α–mediated programmed cell death. ^{32 33 34 35 36 37} It is also known that Fas receptor–Fas ligand complex could activate NF-κB complex, but the significance of this pathway is still unknown. ^{38 39 40 41 42}  

Nevertheless, either pro- or antiapoptotic, STAT1 or NF-κB participate in the process of programmed cell death in an indirect manner, as signal transducers and gene activators, preparing or inhibiting programmed cell death by influencing expression of a receptor, a ligand, or other genes involved in cell death or survival. It is worth noting that there are other apoptotic pathways that classically do not require de novo protein synthesis, such as that mediated by Fas/Fas ligand interaction. ^{43 44} Fas receptor (CD95) is one of the members of the TNF receptor family. Binding of its specific ligand (Fas ligand) or of agonistic anti-Fas ab to this receptor induces a process of programmed cell death. Classically recognized signal transduction pathways of Fas involve the activation of a cascade of apoptotic proteases called caspases that, in turn, induce the cleavage of intracellular substrates and a decrease in mitochondrial membrane potential, all of which occur without activation of transcriptional factors. ^{45 46 47}  

The two processes apoptosis and inflammation frequently coexist in some ocular surface diseases, such as Sjögren’s syndrome or drug-induced pathologies. ^{48 49} We previously demonstrated that in a human conjunctival cell line (Chang cells), IFNγ at the concentration of 300 U/ml induced programmed cell death with a concomitant dose-dependent upregulation of Fas and STAT1. ⁵⁰ The aim of this study, which is a continuation of our former work, was therefore to investigate Fas-induced programmed cell death and to better characterize the connections between IFNγ- and Fas-induced apoptosis in our conjunctival cell line. We also tried to approach some molecular basis of Fas- and IFNγ-mediated processes, more particularly the role of transcriptional factors STAT1 and NF-κB. Moreover, we examined the role in the apoptotic process of an inhibitor of protein synthesis, cycloheximide (CHX), which classically sensitizes cells to Fas-induced programmed cell death, and the effect of IFNγ on Fas proapoptotic action. 

Materials and Methods

Reagents

Eagle’s minimum essential medium, fetal calf serum, and trypsin-EDTA were purchased from Gibco BRL (Paisley, Scotland). Human recombinant interferon gamma was from Pepro Tech (Rocky Hill, NJ). 4′,6′-diamidino-2-phenylindole dihydrochloride (DAPI) and CHX were from Sigma (St. Louis, MO). Antibodies specific for the following human antigens were used: anti-Fas (CH-11, purified; Immunotech, Marseilles, France), anti-poly (ADP-ribose) polymerase (PARP; C2-10, purified; Pharmingen, San Diego, CA), anti–caspase-3 (CPP32) recogzhynizing native and cleaved forms (32- and 17-kDa fragments, purified; Transduction Laboratories, Lexington, KY), anti–caspase-8 recognizing only 55-kDa native form (B9-2, purified; Pharmingen). Fluorescent probes: neutral red, Hoechst 33342, and rhodamine 123 were from Molecular Probes (Leiden, The Netherlands). All reagents were used as recommended by suppliers. 

Conjunctival Cell Line Culture

A human conjunctival cell line (Wong-Kilbourne derivative of Chang conjunctiva, clone 1-5c-4; ATCC CCL-20.2; Manassas, VA) was cultured under standard conditions (5% CO₂, 95% humidified air, 37°C) in Eagle’s minimal essential medium supplemented with 10% fetal calf serum, 2 mM l-glutamine, 50 mg/ml streptomycin, and 50 IU/ml penicillin. Cells were plated at a density of 10,000 cells/well in 96-well plates (Falcon; Becton Dickinson Labware, Plymouth, England) for cold light microplate cytofluorometric assays. Cells were plated in 75-cm² flasks (Falcon) for Western blot analysis and for electrophoretic mobility shift assays and on 20-mm² permanox chamber slide systems (Laboratory-Tek; Nalge Nunc International, Naperville, IL), 25,000 cells per chamber, for morphologic studies (phase contrast microscopy, DAPI staining). Cells were treated with anti-Fas ab, IFNγ, or CHX at least 24 hours after the passage (1:4 split ratio at confluence). 

Anti-Fas Antibody, IFNγ, and CHX Treatments

Anti-Fas monoclonal ab was dissolved in culture medium at the concentration of 500 ng/ml recommended by suppliers, and the cells were then treated for 24 or 48 hours. 

IFNγ was dissolved in culture medium at concentrations of 30 and 300 U/ml as previously reported. ⁵⁰ To induce apoptosis, cells were treated for 48 hours with 300 U/ml IFNγ as it was previously described. ⁵⁰ To study the influence of IFNγ on Fas-induced apoptosis, cells were treated first with 30 U/ml IFNγ for 24 hours (noncytotoxic dose) ⁵⁰ ; then the IFNγ containing medium was discarded, and cells were rinsed twice and treated for 24 hours with 500 ng/ml anti-Fas solution. 

CHX was dissolved in culture medium at a concentration of 1μ g/ml. ⁵¹ To study CHX cytotoxic action, cells were treated for 24 hours with 1 μg/ml CHX alone. To study the influence of CHX on Fas- and IFNγ-induced apoptosis, cells were treated with 1μ g/ml CHX together with 500 ng/ml anti-Fas ab or with 1 μg/ml CHX together with 300 U/ml IFNγ during 24 hours. 

Control cells were treated with the unmodified culture medium. 

Morphologic Procedures

Phase Contrast Microscopy.

Treated cells were observed after (a) 24 or 48 hours of treatment with 500 ng/ml anti-Fas ab, (b) 24 hours of 1 μg/ml CHX treatment, and (c) 24 hours of pretreatment with 30 U/ml IFNγ and then 24 hours of treatment with anti-Fas ab. Moreover, the culture aspect was analyzed after 24 hours of combined treatments with (d) 1 μg/ml CHX and 500 ng/ml anti-Fas, or (e) 1 μg/ml CHX and 300 U/ml IFNγ. Small, adherent, round shape or bubbling and shrunken cells were considered as dying apoptotic cells. These cells were counted in the microscopic field and reported as a percentage of total number of cells. The distinction was made with mitosis. Detached cells were excluded from the count. Morphologic analysis was performed in masked manner by the same investigator during the whole experimental procedure. 

Nuclear Staining.

Cells were processed for DAPI staining after the treatments indicated above. Cells cultured on chamber slides were rinsed twice with PBS, fixed, and permeabilized for 10 minutes in ice-cold 70% ethanol and then washed in PBS and stained with DAPI at a concentration of 0.5 mg/ml for 5 minutes at room temperature. After staining, the slides were extensively washed and mounted in Quantafluor Mounting Medium (Kallestad, Chaska, MN) before examination. A Leica DML microscope (Leica, Heidelberg, Germany) was used for visualization. Cells with chromatin condensation and nuclear fragmentation (apoptotic cells) were counted in the microscopic field and reported as a percentage of total number of cells. Morphologic analysis was performed in masked manner by the same investigator during the whole experimental procedure. 

Cold Light Cytometry

Microplate cold light cytometry is a recently described method that allows realization of toxicological tests in 96-well microtiter plates with excellent reproducibility and sensitivity. The very large spectrum of detected fluorescence (280–870 nm) permits utilization of a considerable number of cellular probes and testing of many different cell functions and characteristics. ^{52 53}  

Cytotoxicity tests were carried out on a microplate cold light fluocytometer (Fluorolite 1000; ThermoBioAnalysis; Dynex, Issy-Les-Moulineaux, France), according to ECVAM (European Center for the Validation of Alternative Methods) recommendations. Probes were used according to manufacturer’s instructions. Assays were conducted using 96-well microtiter plates. 

Neutral red is a viability probe retained in lysosomes of cells with undamaged cell membrane. ^{54 55} Thus, it determines viable or early apoptotic cells. For neutral red test, cells were incubated with 0.005% neutral red solution in culture medium for 3 hours under standard culture conditions. ^{54 55} After this period, the medium was carefully discarded, and cells were rinsed twice in PBS. Cell viability was determined by eluting the dye from stained cells with a solution of 1% acetic acid/50% ethanol (100 μl/well). After thorough mixing to dissolve all neutral red crystals, the plates were rapidly read with specific filters (excitation wave length λ [exc.]= 535 nm, emission wave length λ [em.] = 600 nm). 

Hoechst 33342 is an intercalating dye that allows determination of total chromatin quantity variations and the degree of chromatin condensation. ^{56 57} For Hoechst 33342 test, culture medium was discarded, and plates were incubated in the dark with 10 μg/ml Hoechst 33342 in PBS for 30 minutes. Then, the plates were directly read at λ exc., 360 nm; λ em., 450 nm. We have not presented the results of the Hoechst test separately, but only integrated in the index Hoechst 33342/neutral red as explained below. 

The index of fluorescence Hoechst 33342/neutral red (Ho/NR) is an empiric method for better discrimination of apoptotic and necrotic cell populations. In fact, utilization of this index is especially recommended when stimuli could alter cell proliferation (and thus the total DNA quantity detected by Hoechst 33342) and falsely induce the reduction in Hoechst 33342 fluorescence suggestive of necrosis. This is, for instance, the case of IFNγ. ^{58 59} The index Ho/NR allows evaluation of the importance of chromatin condensation in comparison with cell viability reduction. The schema reported below was established in an empiric manner, using a dilution set of proapoptotic, pronecrotic, and neutral agents in the Laboratory of Toxicology of The School of Pharmaceutics and Biological Science in Paris: Ho/NR greater than 1 is highly suggestive of apoptosis, Ho/NR equal to 1 indicates the proliferating cell system, and Ho/NR less than 1 is most likely related to the necrotic process. 

Early disruption of the mitochondrial membrane potential (ΔΨ_m) was shown to precede nuclear signs of apoptosis in a variety of different systems. ^{60 61} BecauseΔΨ _m results from the unequal distribution of protons on the inner site of the mitochondrial membrane, theΔΨ _m cytofluorometric quantification is based on the use of cationic lipophilic dyes that are sequestered in the mitochondrial matrix according to Nernst equation. We used rhodamine 123 as the fluorescent probe. ⁶² For the rhodamine 123 test, culture medium was discarded, and plates were incubated in the dark with 10 μg/ml rhodamine 123 in PBS for 20 minutes. After this period, the liquid was carefully discarded, and the cells were rinsed twice in PBS and incubated for 1 hour in normal culture medium in standard culture conditions to eliminate nonretained intracellular probe and to equilibrate the quantity of intramitochondrial probe. Then cells were rinsed twice with PBS. ΔΨ_m was determined by eluting the dye from stained cells with a solution of 1% acetic acid/50% ethanol (100 μl/well). After thorough mixing, the plates were rapidly read with specific filters (λ exc., 490 nm; λ em., 530 nm). 

In all experiments, the background fluorescence was determined on wells without cells, but containing the dye solution. At each time point, reported values were the mean of 12 determinations. In all experiments, fluorescence was expressed as the percentage of control values. 

Gel Electrophoresis and Western Blot Analysis

Cytosol- and nuclei-containing cell extracts were separated to make more sensitive the detection of cytosolic and nuclear proteins, respectively. They were prepared by lysing cells at 4°C in hypotonic buffer (10 mM HEPES-KOH, pH 7.9, 1.5 mM MgCl₂, 10 mM KCl, 1 mM spermidine, 1 mM dithiothreitol (DTT), 1 mM PMSF, 1μ g/ml aprotinin, 1 μg/ml pepstatin, 1 μg/ml leupeptin) for 10 minutes on ice. Lysates were centrifuged for 30 seconds at 500g, and the cytosol-containing supernatants were separated from the nuclei-containing pellets. The pellets were resuspended in high-salt buffer (hypotonic buffer with 20% glycerol and 400 mM NaCl) for 30 minutes on ice and then centrifuged for 2 minutes at 18,000g. The nuclei-containing supernatant was transferred into Ependoff tubes. Protein concentration was determined by method of Bradford using a Bio-Rad Protein Determination Kit (10-minute incubation with Coomassie Brilliant Blue dye and subsequent spectrophotometric measurement at wave length = 595 nm) from Bio-Rad SA (Ivry sur Seine, France). The cytosol- and nuclei-containing samples were fractionated (30 μg of protein per lane) by SDS-polyacrylamide gel electrophoresis, transferred to nitrocellulose (Protran BA 83; Schleicher & Schuell, Dassel, Germany) by semidry transfer procedure (Trans-Blot SD; Bio-Rad SA). The nitrocellulose membranes were then incubated for 1 hour in blocking buffer (PBS, containing 0.1% Tween 20, and 5% nonfat milk powder), rinsed and incubated for 1 hour with specific antibodies (anti-PARP, anti–CPP32, anti–caspase-8) used as recommended by suppliers. Blots were developed using enhanced chemiluminescence reagent ECL (Amersham, Arlington Heights, IL). We previously published the results of immunoblotting of PARP after 24 and 48 hours of treatment with 300 U/ml IFNγ (Fig. 6 , lanes 300 U IFNγ/24 h and 300 U IFNγ/48 h). ⁵⁰ In the present study, we present these data to review the effects of IFNγ when applied alone and to better compare them with other treatments. 

Electrophoretic Mobility Shift Assay

The activation of DNA-binding sites specific to transcriptional factors NF-κB and STAT1 ⁶³ can be determined with this procedure. 

Cells were treated with (a) 300 U/ml IFNγ for 30 minutes, 2, 6, and 24 hours; (b) 1 μg/ml CHX for 24 hours; (c) 500 ng/ml anti-Fas ab for 24 hours; (d) 300 U/ml IFNγ and 1 μg/ml CHX for 24 hours; or (e) 500 ng/ml anti-Fas ab and 1 μg/ml CHX for 24 hours. Nuclear extracts were prepared as described above. Nuclear protein/DNA-binding reactions were performed in a 20-μl volume containing 10 μg nuclear extract protein, 10 mM HEPES-KOH, pH 7.9, 100 mM NaCl, 1 mM EDTA, 1 mM DTT, 10% glycerol, and 2 μg poly(dI-dC) as a nonspecific competitor. After the preincubation (20 minutes at 25°C), 2 μl ³²P-labeled double-stranded target oligonucleotide was added in each reaction, in the absence or presence of a 100-fold molar excess of nonradioactive competitor oligonucleotide (for analysis of the specificity of induced DNA-binding complexes), and incubated for 20 minutes at 25°C. The sequence of the nucleotides that corresponds to the NF-κB consensus DNA-binding site was as follows: forward, 5′-AGT TGA GGG GAC TTT CCC AGG C-3′; reverse, 5′-GCC TGG GAA AGT CCC CTC AAC T-3′. The sequence of the nucleotides that corresponds to STAT consensus DNA-binding site (GAS/ISRE) was as follows: forward, 5′-AAG TAC TTT CAG TTT CAT ATT ACT CTA-3′; reverse, 5′-TAG AGT AAT ATG AAA CTG AAA GTA CTT-3′. As interferon gamma transduction pathway preferentially implies STAT1 factor, this sequence in our model could therefore be considered as specific to STAT1. 

Electrophoretic mobility shift assay (EMSA) was carried out on 4% 37.5:1 acrylamide-bisacrylamide gels in 45 mmol TBE containing 1 mmol EDTA, pH 8.0, at 4°C for 1 hour. Gels were dried under vacuum and subjected to autoradiography with intensifying screens at −70°C. 

Statistical Analysis

Results of microplate fluorometric assays were calculated as arithmetic means ± SD, and significance values were calculated by means of the two-way analysis of variance (ANOVA), with P < 0.05 regarded as significant. All experiments in this study were at least duplicated. 

Results

Morphologic Analysis

Cells were analyzed first with a phase contrast microscope. When compared to control, in samples treated with 500 ng/ml anti-Fas ab there were 15% to 20% dead cells after 24 hours of treatment and 25% to 30% after 48 hours of treatment, with a concomitant decrease in density of adherent cells (Figs. 1A 1B 1C) . 

Compared with control, 30 U/ml IFNγ or 1 μg/ml CHX did not induce any modification of culture aspect (Figs. 1D 1E) . When cells were treated for 24 hours with 500 ng/ml anti-Fas monoclonal ab after 24 hours pretreatment with 30 U/ml IFNγ or together with 1 μg/ml CHX, there were 50% to 80% dead cells in the treated sample (Figs. 1F 1G) . 

There were 50% dead cells in samples treated with 300 U/ml IFNγ alone only after 48 hours of treatment (Fig. 1H) but not after 24 hours of treatment (data not shown). When 300 U/ml IFNγ was applied together with 1 μg/ml CHX, there were 80% dead cells after 24 hours of treatment (Fig. 1I) . 

After DAPI staining, comparable fractions of cells showed chromatin condensation and nuclear fragmentation suggestive of apoptosis (Figs. 2A 2B 2C 2D 2E 2F 2G) . 

Cold Light Microplate Flow Cytometry

Cell Viability Assay.

Cell viability, evaluated by neutral red assay (Fig. 3) , significantly decreased after 24 hours of treatment with 500 ng/ml anti-Fas ab alone (Fig. 3A) . This reduction was 20% and 30%, respectively, after 24 and 48 hours of treatment. Cell viability significantly decreased after 1 hour of treatment with 500 ng/ml anti-Fas ab applied after 24 hours of pretreatment with 30 U/ml IFNγ (Fig. 3B) . After 24 hours, there were only 42% cells with cell membrane integrity (viable or early apoptotic cells). 

Index Ho/NR.

Index Ho/NR (Fig. 4) was markedly above 1 after 48 hours of treatment with 300 U/ml IFNγ (Fig. 4A) , 24 hours of treatment with 500 ng/ml anti-Fas ab (Fig. 4B) , and 3 hours of treatment with 500 ng/ml anti-Fas solution when a pretreatment with 30 U/ml IFNγ was applied (Fig. 4C) . Ho/NR index superior to 1 is highly suggestive of apoptosis. 

ΔΨ_m.

All results are represented in Figure 5 . A disruption in ΔΨ_m started after 3 hours of treatment with 300 U/ml IFNγ (Fig. 5A) , after 1 hour of treatment with anti-Fas ab alone (Fig. 5B) , or after a pretreatment with 30 U/ml IFNγ (Fig. 5C) , which was suggestive of an apoptotic process. 

Western Blot Analysis

Figure 6 represents the immunoblotting of PARP, which demonstrated the increase in quantity of the native form of 116 kDa, its proteolytic cleavage, and presence of an 85-kDa fragment in nuclear cell extracts after 24 and 48 hours of treatment with 500 ng/ml anti-Fas aab. When anti-Fas monoclonal antibody was applied after 24 hours of treatment with 30 U/ml IFNγ, the cleavage was quantitatively more significant, with a decrease in the amount of the native form of PARP. The same pattern was observed when cells were treated with anti-Fas together with CHX. 

The increase in quantity of the native form of PARP and its slight cleavage was also noted after 24 hours of treatment with 300 U/ml IFNγ (previously published data ⁵⁰ ). This cleavage became important after 48 hours of treatment with 300 U/ml IFNγ alone (previously published data ⁵⁰ ) or after 24 hours of the combined treatment with IFNγ/CHX. 

Immunoblotting of caspase-3 CPP32 (Fig. 7) showed the unusual cleavage of the native form (cleaved fragment more than 17 kDa) after anti-Fas ab treatments (24 or 48 hours). This cleavage became regular (molecular weight of the cleaved fragment of 17 kDa) and quantitatively more significant (decrease in the native form of CPP32) when cells were pretreated with 30 U/ml IFNγ or when anti-Fas ab was applied together with 1 μg/ml CHX. A slight irregular cleavage was also observed after 24 hours of treatment with 300 U/ml IFNγ and became potentiated and regular after a combined treatment with IFNγ/CHX (24 hours). 

Caspase-8 (Fig. 8) was upregulated in the cytosolic extracts of cells treated with anti-Fas sera (24 and 48 hours). The decrease in the quantity of caspase-8 native form suggestive of its quantitatively more significant cleavage was observed when cells were pretreated with 30 U/ml IFNγ before anti-Fas ab treatment or after 24 hours of combined treatment with anti-Fas/CHX. This cleavage could not be objectified by the antibody used in these experiments, which only detected the native, uncleaved form. Upregulation of caspase-8 also was observed after 24 and 48 hours of treatment with 300 U/ml IFNγ. 

Electrophoretic Mobility Shift Assays

As shown in Figure 9 , cell treatment with 300 U/ml IFNγ was associated with induction of a nuclear NF-κB–binding activity at the 30th minute of treatment. After 2 hours of treatment, the DNA-binding activity decreased and reappeared only after 24 hours of IFNγ treatment, thus realizing a biphasic pattern. CHX treatment completely suppressed IFNγ-induced DNA-binding complex. Mobility shift retardation was also detected after 24 hours of treatment with anti-Fas ab and was suppressed when the combined treatment with anti-Fas/CHX was applied. 

Figure 10 represents STAT-binding activity. The GAS/ISRE sequence is a consensus binding site for IFNγ-activation factor STAT. The STAT-binding activity was revealed after 2 hours of treatment with 300 U/ml IFNγ and persisted during 24 hours of treatment without any modification of intensity. The retardation complex was still detected with diminished intensity after IFNγ/CHX combined treatment. A slight binding activity was also detected after 24 hours of anti-Fas ab treatment and was suppressed by CHX. 

For all samples, after the competition analysis, the formation of slowly migrating complexes suffered competition from a 100-fold excess of unlabeled oligonucleotide. 

Discussion

We have previously shown that in a human conjunctival cell line, IFNγ at a concentration of 300 U/ml induced programmed cell death accompanied by overexpression of the inflammation-related proteins HLA DR and ICAM-1, apoptosis-related proteins Fas/CD95 and bax, and of the transcriptional factor STAT1. In contrast, although IFNγ at 30 U/ml was nontoxic, it was also found to upregulate the expression of Fas after 24 hours of treatment. ⁵⁰ Our present data confirm that Chang conjunctival cells constitutively express Fas and that this receptor is functional. In fact, our cell line is, to some extent, sensitive to CD95-mediated apoptosis. Actually, 25% to 30% of cells undergo the process of programmed cell death after treatment with anti-Fas ab alone, a fact that, moreover, suggests the relative resistance of the Chang cell line to Fas-mediated programmed cell death. Similarly, the low susceptibility of other cell systems to Fas-induced apoptosis was previously reported in some in vitro models as well as the possibility of overcoming this resistance by administration of IFNγ. ^{28 64 65 66 67 68} Accordingly, in our model, the application of 30 U/ml IFNγ potentiated Fas proapoptotic effects. The simplest explanation of this phenomenon could be the existence of IFNγ-induced upregulation of CD95. However, in our opinion, a slight upregulation of Fas expression observed after 24 hours of treatment with 30 U/ml IFNγ, as we previously demonstrated, ⁵⁰ is not sufficient to justify 80% dead cells after a subsequent anti-Fas ab treatment. Thus, we were more particularly interested in the mechanisms of Fas- and IFNγ-induced apoptosis and their possible connections and interactions. The activation of STAT1 by IFNγ is now well established as well as the requirement of this transcriptional factor in IFNγ proapoptotic potential. ^{25 58 69 70} In agreement with these data, in our model, IFNγ not only upregulated STAT1 in cytosolic and nuclear extracts ⁵⁰ but also activated STAT consensus binding site (GAS/ISRE) in a lasting and strong manner which could, partly, explain IFNγ-mediated induction and facilitation of programmed cell death. 

The stimulation by IFNγ of another transcriptional factor, NF-κB, has been much less explored. It is well known that NF-κB participates in the activation of proinflammatory genes and plays an important role in inflammatory processes. ^{71 72} Several teams have recently reported controversial findings concerning the role of NF-κB in apoptosis. In different cellular systems, NF-κB can either suppress ^{73 74 75} or promote ^{28 76 77 78} the apoptotic process. In our model, NF-κB–binding activity evolved in a biphasic manner after IFNγ treatment. The early activation phase, after 30 minutes of treatment with IFNγ, most likely concerned the preexisting cytosolic NF-κB. The late complex detected at the 24th hour was strongly suggestive of synthetic activity of the cell and was suppressed by the protein synthesis inhibitor, CHX. NF-κB activation was also detected after 24 hours of treatment with anti-Fas ab and was suppressed by CHX. Our hypothesis is that, in this model, NF-κB may be part of a survival mechanism used by the cell to escape death. In fact, the strong IFNγ-induced activation of STAT1 counterbalanced this mechanism, whereas the Fas-mediated STAT1 transduction pathway was not important enough (the STAT-binding activity detected was very slight) to reverse NF-κB antiapoptotic effect. 

Moreover, we presume that STAT1 activation in our model does not need de novo protein synthesis to constitute a sufficient proapoptotic signal. Indeed, the cytosolic and nuclear protein upregulation was constant, ⁵⁰ and the activated DNA-STAT complex was detected in nuclear cell extracts also in a constant manner, without any variation of intensity between 30 minutes and 24 hours of IFNγ treatment. In addition, CHX did not completely neutralize the detected complex. In contrast, NF-κB activation was biphasic and thus highly suggestive of a synthetic process. Its complete interruption by CHX stopped the antiapoptotic signal in our cells. In the light of this explanation, it is obvious why CHX could sensitize cells to undergo Fas- or IFNγ-mediated apoptosis. 

Therefore, in our model, IFNγ-induced facilitation of Fas-mediated programmed cell death is probably the result of a complex process. IFNγ-induced upregulation of the Fas receptor might be one of the factors facilitating Fas-dependent apoptosis, but most likely not the predominant one. Indeed, we observed this facilitation in terms of cell viability modifications, but also in the patterns of PARP, CPP32, and caspase-8 immunoblotting. Fas antibody alone induced only partial and irregular cleavage of these intracellular substrates. After IFNγ pretreatment, the amount of cleavage increased with appearance of well-defined regular-weight products. We presumed that NF-κB activation induced by Fas could conflict with caspase activation and consequently with their proteolytic action. We suggest that IFNγ-induced modification of the NF-κB/STAT1 imbalance and the resultant shift to the proapoptotic signal are crucial in potentiation of Fas-mediated programmed celldeath. 

Besides the two principal intracellular mediators NF-κB and STAT1, IFNγ and anti-Fas could imply other transcription factors such as Smad7 ^{79 80} or JNK/SAPK (c-Jun N-terminal kinase/stress-activated protein kinase). ^{81 82 83 84} Furthermore, our study concerns only some aspects of modulation of NF-κB and STAT1. Thus, other investigations should be necessary to confirm our hypotheses about significance of NF-κB– and STAT1-related molecular events. 

In several cellular systems, a reduction of mitochondrial membrane potential is a crucial, very early event, independent of the transcription mediators NF-κB and STAT1, in the programmed cell death process. ^{85 86} This phenomenon is due to the activation of a high conductance permeability transition pore in the mitochondrial inner membrane. Its opening abruptly increases the permeability of the mitochondrial inner membrane to solutes of molecular mass up to 1500 Da, among which are some of proapoptotic molecules such as cytochrome c. In our model, both IFNγ and agonistic anti-Fas generated a similar decrease in ΔΨ_m. However, Chang cell line appeared to be relatively resistant to Fas-induced death. This fact suggests that, in this model, the detected decrease ofΔΨ _m might be only an additional factor in apoptotic process. In fact, in some cellular systems, mitochondrial permeability transition could occur without leading to a programmed cell death cascade. ⁸⁷ Conversely, the release of proapoptotic molecules from mitochondrial intermembrane space can be simultaneous to the steady state or even to a rise ofΔΨ _m. ^{88 89 90}  

A technical reserve regarding our conclusions is that we did not perform the supershift assay to determine the specificity of DNA–protein complexes. It is now well established that the IFNγ transduction pathway principally implies STAT1. The fact that after IFNγ treatment STAT1 was upregulated in nuclear and cytosolic extracts, as we previously demonstrated, ⁵⁰ and that IFNγ induced formation of a retardation complex specific to the GAS/ISRE sequence were suggestive enough for activation of this member of STAT family. In contrast, the NF-κB–binding consensus site is specific to the NF-κB factor and does not require the supershift analysis. 

Another reserve concerns our model of conjunctival cell line. These cells present some characteristics of conjunctival epithelium (desmosomes, microvilli, expression of EGF and Fas receptors, absence of expression of HLA DR ^{48 50 91 92} ), which are, however, not sufficient to directly extrapolate our findings to human pathology of the ocular surface. In fact, the Chang epithelium is an immortalized monolayer constituted from only one type of cells in which there is no tear film or well-defined mucus. Therefore, our model remains an experimental approach even if it could serve as a basis to other in vivo or clinical studies. In the light of the role of IFNγ in modulation of Fas-induced programmed cell death in vitro, it could be interesting to investigate further the interactions of these factors in vivo, for instance in a pathology associating inflammation and apoptosis such as Sjögren’s syndrome. In fact, in this disease lymphocytic infiltration, the hallmark of inflammatory process is accompanied by Fas-mediated programmed cell death of acinar cells. ⁴⁹ A better understanding of the mechanisms of how inflammation could influence apoptotic destruction of lacrymal glands could therefore lead to additional therapeutic approaches toward this disease and toward many other ones. 

 Submitted for publication November 18, 1999; revised March 2, 2000; accepted March 22, 2000.

 Commercial relationships policy: N.

 

Figure 1.

 

Phase contrast microscopy analysis after anti-Fas ab, IFNγ and cycloheximide (CHX) treatments. (A) Control cells. (B) Cells treated for 24 hours with 500 ng/ml anti-Fas ab. Fifteen percent to 20% of dead cells are visible. (C) Cells treated for 48 hours with 500 ng/ml anti-Fas ab. Twenty-five percent to 30% of dead cells are present in the microscopic field. (D) Cells treated for 24 hours with 30 U/ml IFNγ. Cell aspect is comparable to the control. (E) Cells treated for 24 hours with 1 μg/ml CHX. No alteration in cell culture is visible. (F) Cells treated for 24 hours with 500 ng/ml anti-Fas ab after a 24-hour pretreatment with 30 U/ml IFNγ. A considerable number of dead cells are present in the microscopic field. (G) Cells treated for 24 hours with 500 ng/ml anti-Fas ab and 1 μg/ml CHX (combined treatment). The density of adherent cells is diminished when compared to the control and almost all cells are detached. (H) Cells treated for 48 hours with 300 U/ml IFNγ. There are 50% of round detached apoptotic cells. (I) Cells treated for 24 hours with 300 U/ml IFNγ combined to 1 μg/ml CHX treatment. The majority of cells are detached.

Figure 2.

 

Nuclear DAPI staining. (A) Control cells. (B) Cells treated for 24 hours with 30 U/ml IFNγ. Cell nuclei aspect is comparable to the control. (C) Cells treated for 24 hours with 1 μg/ml cycloheximide (CHX). There is no modification of nuclei compared with the control. (D) Cells treated for 24 hours with 500 ng/ml anti-Fas ab. Few apoptotic cells are visible (15%–20%). (E) Cells treated for 48 hours with 500 ng/ml anti-Fas ab. There are approximately 30% of apoptotic cells. (F) Cells treated for 24 hours with 500 ng/ml anti-Fas ab after 24 hours pretreatment with 30 U/ml IFNγ. There are a considerable number of apoptotic nuclei in the visual field. (G) Supernatant of culture treated for 24 hours with 500 ng/ml anti-Fas ab together with 1 μg/ml CHX. The majority of nuclei are apoptotic.

Figure 3.

 

Microplate cold light cytometry. Relative cell number as determined by neutral red assay. Dashed arrows parallel to x-axis indicate a change in the time scale. (A) There is a significant decrease in cell viability respectively 25% to 33% after 24 and 48 hours of treatment with 500 ng/ml anti-Fas ab alone. (B) Anti-Fas antibody treatment preceded by 24 hours of pretreatment with 30 U/ml IFNγ. There is a significant decrease in cell viability after 1 hour of treatment with 500 ng/ml anti-Fas ab. Only 42% of cells maintain cell membrane integrity (viable and early apoptotic cells) after 24 hours of treatment.

Figure 4.

 

Index Hoechst 33342/neutral red (Ho/NR) as determined by microplate cold light cytometry. Dashed arrows parallel to x-axis indicate a change in the time scale. Index Ho/NR superior to 1 is suggestive of apoptosis. (A) Index Ho/NR is significantly superior to 1 after 24 hours of treatment with 300 U/ml IFNγ. (B) Index Ho/NR is significantly superior to 1 after 24 hours of treatment with 500 ng/ml anti-Fas ab alone. (C) Index Ho/NR is significantly superior to 1 after 3 hours of treatment with 500 ng/ml anti-Fas ab when 24 hours of pretreatment with IFNγ was applied.

Figure 5.

 

Mitochondrial membrane potential (ΔΨ_m) as determined by microplate cold light cytometry. Dashed arrows parallel to x-axis indicate a change in the time scale. (A) There is a significant decrease inΔΨ _m after 3 hours of treatment with 300 U/ml IFNγ. (B) There is a significant decrease inΔΨ _m after 1 hour of treatment with 500 ng/ml anti-Fas ab alone. (C) There is a significant decrease inΔΨ _m after 1 hour of treatment with 500 ng/ml anti-Fas ab when a pretreatment with 30 U/ml IFNγ was applied.

Figure 6.

 

Western blot analysis of expression of poly (ADP-ribose) polymerase (PARP) in nuclear cell extracts. The increase in the quantity of native form of PARP and its proteolytic cleavage are characteristic of apoptosis.

Figure 7.

 

Western blot analysis of expression of caspase-3 (CPP32) in cytosolic cell extracts. The increase in the quantity of native form and a proteolytic cleavage of CPP32 are characteristic of apoptosis.

Figure 8.

 

Western blot analysis of expression of native form of caspase-8 in cytosolic cell extracts.

Figure 9.

 

Electrophoretic mobility shift assay with ³²P-labeled double-stranded oligonucleotides corresponding to the consensus NF-κB–binding site. The retarded DNA-protein complex indicated by the arrowhead was detected after 30 minutes and 24 hours of IFNγ treatment and after 24 hours of treatment with 500 ng/ml anti-Fas ab. CC, competition with a 100-fold excess of unlabeled oligonucleotide; C, control cells.

Figure 10.

 

Electrophoretic mobility shift assay with ³²P-labeled oligonucleotides corresponding to the Gamma Activating Sequence (GAS). The retarded GAS-STAT complex indicated by the arrowhead was detected after 30 minutes, 2, 6, and 24 hours of treatment with IFNγ alone, after 24 hours of the combined treatment IFNγ/CHX, and after 24 hours of treatment with anti-Fas ab alone. CC, competition with a 100-fold excess of unlabeled oligonucleotide; C, control cells.

The authors thank Annie-France Bringuier and Alain Moreau for their excellent technical assistance.