Interferon-γ Induces Apoptosis and Expression of Inflammation-Related Proteins in Chang Conjunctival Cells

Abstract

purpose. The purpose of this study was to investigate the effect of interferon (IFN)γ on cell viability, cell growth, and apoptosis and on expression of apoptotic and inflammation-related proteins in epithelial conjunctival cells in vitro. Some aspects of transduction pathways of IFNγ-induced alterations were also investigated, especially the role of protein kinase C (PKC) and IFNγ transcriptional factor STAT1.

methods. A human conjunctival cell line was treated with different concentrations (30 and 300 U/ml) of human recombinant IFNγ. After 24, 48, and 72 hours of treatment, cell viability and relative cell number were studied with 3-(4,5-dimethylthiazol-2yl)-2,5-diphenyl tetrazolium bromide (MTT) and crystal violet colorimetric assays. The apoptotic process was sought by phase-contrast microscopy, 4′,6′-diamidino-2-phenylindole dihydrochloride (DAPI) staining, and transmission electron microscopy and was confirmed by DNA electrophoresis and immunoblotting of poly(ADP-ribose) polymerase (PARP). The cell cycle and expression of apoptotic proteins Fas, bax, and p53; of inflammation-related proteins HLA-DR and intercellular adhesion molecule (ICAM)-1; and of IFNγ signal-transducing factor STAT1 were evaluated by flow cytometry and/or western blot analysis. To investigate PKC-related transduction pathways, two PKC modulators, 12-O-tetradecanoyl-phorbol-13-acetate (TPA) and staurosporine, were applied for 3 hours, followed by IFNγ treatment for 72 hours. Moreover, the effects of PKC depletion were studied after a 24-hour application of TPA, also followed by IFNγ treatment for 72 hours. Then, Fas, ICAM-1, and HLA-DR expressions were studied by flow cytometry.

results. IFNγ at 30 U/ml induced no change in cell cycle and in cell viability. Cell viability significantly decreased after 48 hours of treatment with 300 U/ml IFNγ, associated with cell cycle alterations (decrease in number of cells in the S–M phase), apoptotic chromatin condensation and fragmentation, ladder pattern on DNA electrophoresis assay, and cleavage of PARP. Moreover, IFNγ-treated cells overexpressed plasma membrane Fas, HLA-DR, and ICAM-1 in a dose- and time-dependent manner, and STAT1 in both nuclear and cytosolic cell fractions. Only 300 U/ml IFNγ-treated cells overexpressed bax, whereas Bcl-2 and p53 proteins were not modified. HLA-DR and Fas were upregulated after addition of staurosporine or after PKC-depleting treatment and repressed with TPA. Staurosporine, PKC depletion, and TPA all enhanced ICAM-1 expression.

conclusions. In our model, IFNγ induced expression of inflammatory molecules and apoptotic mediators, cell growth arrest, and apoptosis of Chang conjunctival cells. Moreover, our results suggest that activation of PKC is not involved in some IFNγ cellular effects that possibly imply the upregulation and nuclear translocation of STAT1. IFNγ-induced apoptosis could explain in part the recently reported coexistence of inflammation and programmed cell death in ocular surface inflammatory disorders such as Sjögren’s syndrome.

Apoptosis is one of the forms of cell death which, in opposition to necrosis, can be induced by specific stimuli (such as an interaction between death receptors and their ligands) and engages well-defined signal transduction pathways and effector mechanisms. ^{1 2} It is a genetically programmed process, marked by cytoplasm shrinkage, membrane blebbing, chromatin condensation, and DNA fragmentation. Three families of genes and their products play a crucial role in induction and control of apoptosis. ^{3 4} These are the death receptor family (Fas receptor and its ligand and tumor necrosis factor (TNF)-α receptors rTNFα and TNFα), the Bcl-2 family, and the family of caspases. The Bcl-2 family comprises two groups of proteins that either protect (Bcl-2, bcl-Xl) or sensibilize (bax, bak, bad) the cell to undergo apoptosis. Caspases are cysteine proteases that act in cascade in initiating and executing programmed cell death. 

IFNγ, a 25-kDa glycoprotein, plays a crucial role in vivo in many different types of immune responses, such as delayed-type hypersensitivity, inflammation, or graft rejection ^{5 6} and is involved in pathogenesis of inflammatory diseases, for instance of Sjögren’s syndrome. ^{7 8 9 10} This cytokine is secreted exclusively by T cells (cytotoxic and Th1) and natural killer cells. It induces antiviral ⁶ and antiproliferative activities, ¹¹ stimulates macrophages, ¹¹ and controls the expression of several adhesion molecules and surface cell receptors (e.g., intercellular adhesion molecule [ICAM]-1, Fas antigen), ^{12 13 14} of several cytokines and of major histocompatibility complex (MHC) class I and II molecules. ^{15 16} Class II MHC antigens such as HLA-DR play a crucial role in the initiation of immune responses. Their expression by epithelial cells may enable them to act as antigen-presenting cells and to interact with helper T lymphocytes in immune processes. ICAM-1 is one of the accessory molecules essential for communication between lymphocytes and other cells (e.g., epithelial cells) and for control of leukocyte migration and adhesion to different target tissues in inflammatory process. 

In vitro, IFNγ is also among the earliest polypeptides found to inhibit growth and proliferation in cultured cells. ^{17 18} Recently, its proapoptotic action has been demonstrated in some tumoral cell lines. ^{18 19 20}  

Both its apoptotic and immune effects depend on IFNγ-induced cellular transcriptional changes. Molecular mechanisms of IFNγ signal transduction imply the activation, by phosphorylation on tyrosine, of specific latent transcriptional factors or STATs (signal transducers and activators of transcription), which induce activation or repression of numerous genes. ^{21 22} The protein tyrosine kinase–STAT pathway has been shown to be critical for IFNγ-induced expression of ICAM-1, ²³ for growth arrest, and for apoptosis. ^{17 24 25} Moreover, IFNγ was reported to stimulate the PKC-related pathway in many cellular systems, ^{26 27 28 29} and this action was reported to be mandatory in the cytokine-mediated upregulation of ICAM-1. ^{30 31 32} PKC activation was also shown to participate in IFNγ-induced expression of HLA-DR and Fas antigen in some cellular systems. ^{29 33 34}  

In the eye, IFNγ levels are increased in some inflammatory ocular surface disorders, such as corneal allograft rejection ^{35 36} or Sjögren’s syndrome. ^{37 38 39} Moreover, in Sjögren’s syndrome, this cytokine is presumed to be one of the principal molecules responsible for stimulation of salivary, conjunctival, and lacrimal expression of inflammation-related molecules including HLA-DR and ICAM-1, which are constantly upregulated in this disorder and are even used as markers of the pathologic state in clinical practice. ^{38 40 41 42 43 44} In addition to inflammation, apoptotic changes were shown in glandular and ocular surface tissues of patients with Sjögren’s syndrome and in Sjögren-like animal models, ^{42 45 46} and their dependence on lymphocytic cytokines such as IFNγ seems most likely. Apoptosis and inflammation are both resolved with appropriate causal treatment, such as cyclosporin A, ⁴⁶ a fact that could argue for the common origin of both processes. 

We studied the inflammatory and apoptotic potentials of IFNγ in a human continuous conjunctival cell line. The concentrations of IFNγ that we applied were 10 times higher than those detected in human serum or in different tissues in normal or pathologic states, ^{47 48 49 50} but they were comparable to the concentrations of IFNγ secreted by normal human peripheral blood mononuclear cells in in vitro conditions. ⁵¹ We used the inflammatory markers HLA-DR and ICAM-1 as proof of efficiency of IFNγ inflammatory stimulation. We thus showed that the proapoptotic action of IFNγ accompanies expression of inflammatory molecules and thus could help to explain the coexistence of inflammation and apoptosis in ocular surface disorders in human tissues in Sjögren’s syndrome. 

Materials and Methods

Reagents

Eagle’s minimum essential medium, fetal calf serum, and trypsin-EDTA were purchased from Gibco (Paisley, Scotland); human recombinant IFNγ from Pepro Tech (Rocky Hill, NJ); and 12-O-tetradecanoyl-phorbol-13-acetate (TPA), staurosporine, crystal violet, 4′,6′-diamidino-2-phenylindole dihydrochloride (DAPI), and 3-(4,5-dimethylthiazol-2yl)-2,5-diphenyl tetrazolium bromide (MTT) from Sigma (St. Louis, MO). Antibodies specific for the following human antigens were used: anti-Fas (UB2, fluorescein isothiocyanate [FITC] -conjugated; Immunotech, Marseille, France), anti-p53 (DO7, purified and FITC-conjugated; Pharmingen, San Diego, CA), anti-Bcl-2 (N-19, purified; Pharmingen), anti-bax (4F11, purified; Immunotech), anti-poly(ADP-ribose) polymerase (PARP; C2-10, purified; Pharmingen), anti-STAT1 (purified; Transduction Laboratories, Lexington, KY), anti-HLA-DR (Immu-357, FITC-conjugated; Immunotech), and anti-ICAM-1 (6.5B5, purified; Dako, Trappes, France). Control antibodies (mouse FITC-conjugated IgG1 and IgG2κ, mouse phycoerythrin (PE)-conjugated IgG1) were purchased from Immunotech. Staining solutions for cell cycle (DNA Prep Stain) were from Coulter (Miami, FL). 

Conjunctival Cell Line Culture

A human conjunctival cell line (Wong–Kilbourne derivative of Chang conjunctiva, clone 1-5c-4, ATCC CCL-20.2) 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, UK) for MTT and crystal violet assays. Cells were plated in 75-cm² flasks (Falcon) for flow cytometry, western blot analysis, DNA electrophoresis, and transmission electron microscopy and on 20-mm² permanox chamber slide systems (Laboratory-Tek; Nalge Nunc, Naperville, IL), 25,000 cells/chamber, for other morphologic studies. Cells were treated with IFNγ at least 24 hours after the passage (1:4 split ratio at confluence). 

IFNγ Treatment

IFNγ was dissolved in culture medium at concentrations of 30 and 300 U/ml. 

TPA and Staurosporine Treatments

TPA, similar to other tumor-promoting phorbol esters, at first activates, but then depletes cells of PKC during prolonged treatment. Cells were treated for 3 hours with 10 ng/ml TPA (PKC-stimulating action) and 50 μM staurosporine (a PKC inhibitor), followed by a 300-U/ml IFNγ treatment for 72 hours. The effects of PKC depletion were studied after a 24-hour application of 10 ng/ml TPA, followed by a 300-U/ml IFNγ treatment for 72 hours. After these periods, expressions of Fas, HLA-DR, and ICAM-1 were studied by flow cytometry. 

Cell Viability and Cell Number Assays

Assays were conducted using 96-well microtiter plates. At 24, 48, and 72 hours of IFNγ treatment, cell viability was assessed with MTT assay, as described previously. ⁵² MTT is bioreduced in metabolically active cells into a colored formazan product insoluble in tissue culture medium. At times indicated previously, 5 mg/ml MTT solution was added to the culture medium (10 μl per 100 μl of medium), and plates were incubated at 37°C for 4 hours. After this period, the liquid was carefully discarded. Acid-isopropanol (0.04 N HCl in isopropanol) was added (100 μl/well) and mixed thoroughly to dissolve all formazan crystals. Plates were then rapidly read on an enzyme-linked immunosorbent assay (ELISA) multiwell plate reader (iEMS Reader; Labsystems, Helsinki, Finland) at 570nm. 

To determine relative cell number, cells were stained with crystal violet at 24, 48, and 72 hours of IFNγ treatment, as described previously. ⁵³ Briefly, the cells were rinsed twice with sterile phosphate-buffered saline (PBS; pH 7.4) and then fixed in 70% cold ethanol for 10 minutes at room temperature; 100 μl/well of 0.5% crystal violet solution was added. The relative cell number was determined by eluting the dye from stained cells with 33% acetic acid, and absorbance was measured at 540 nm on an ELISA multiwell reader. 

In both experiments, absorbance was expressed as a percentage of control values. The background absorbance was determined on wells without cells, but containing the dye solution. At each time point, cell viability or relative cell number values were the mean of three to six determinations. 

Nuclear DNA Isolation and Electrophoresis

After 72 hours of treatment with 300 U/ml IFNγ, DNA was isolated from adherent cells cultured in 75-cm² flasks by a proteinase K–phenol method, as previously described. ⁵⁴ DNA samples were treated with 50 μg/ml DNase-free RNase, extracted twice with phenol/chloroform, precipitated with ethanol, and dissolved in 10 mM Tris-HCl (pH 7.6) and 1 mM EDTA. DNA samples (10 μg) were fractionated by electrophoresis on 1% agarose gels and visualized by staining with ethidium bromide (0.5 μg/ml). 

Morphologic Procedures

Phase-Contrast Microscopy.

Treated cells were observed after 24, 48, and 72 hours of treatment with 30 and 300 U/ml IFNγ. 

Nuclear Staining.

Cells were processed for DAPI staining after 48 and 72 hours of 30- and 300-U/ml IFNγ treatment. Cells cultured on chamber slides and supernatants were rinsed twice with PBS, fixed, and permeabilized for 10 minutes in ice-cold 70% ethanol, 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 washed extensively and mounted (Quantafluor Mounting Medium; Kallestad, Chaska, MN) before examination. A Leica DML light microscope (Leica, Heildelberg, Germany) was used for visualization. Morphologic analysis was performed in a masked manner by the same investigator during the whole experimental procedure. 

Transmission Electron Microscopy.

After 72 hours of 300-U/ml IFNγ treatment, cells cultured in 75-cm² flasks were harvested in PBS by gentle scraping and pelleted by centrifugation. The cells were fixed in 2.5% buffered glutaraldehyde for 1 hour at 4°C, rinsed in PBS, postfixed with 1% osmium tetroxide for 2 hours at room temperature, and then dehydrated in a graded ethanol series, followed by embedding of the cell pellets in Epon 812. Ultrathin sections were stained with uranyl acetate and lead citrate and examined with an electron microscope (JEOL, Tokyo, Japan) operating at 80 kV. 

Flow Cytometry

All measurements were performed on a flow cytometer(FACScan; Becton Dickinson, Mountain View, CA) equipped with an argon laser emitting at 488 nm, using software (Lysis II, Becton, Dickinson) for data analysis. Forward-scatter and side-scatter, FITC fluorescence (FL1, 525 nm band pass), and propidium iodide fluorescence (FL3, 630 nm band pass) were measured. At least 10,000 events were collected per sample. The flow cytometry data were reported as mean fluorescence intensities. 

Expression of Inflammation and Apoptosis-Related Proteins.

Besides two principal tested IFNγ concentrations (30 and 300 U/ml), we introduced an intermediate concentration at 150 U/ml to study the dose-dependence of protein expressions. 

For flow cytometric analysis of Fas, HLA-DR, and ICAM-1 expression, cells were harvested with trypsin-EDTA, pelleted, washed twice in PBS, and incubated for 30 minutes with FITC-conjugated anti-Fas, FITC-conjugated anti-HLA-DR, or purified anti-ICAM-1 antibodies, with FITC-conjugated and purified mouse IgG1 as negative controls. After incubation with anti-ICAM-1 antibody and purified mouse IgG1, cells were washed twice and pelleted, and a secondary antibody, FITC-conjugated goat anti-mouse immunoglobulin, was applied for 30 minutes. For p53 and bax labeling, cells were fixed and permeabilized for 5 minutes with 1% paraformaldehyde in PBS, followed by 100% cold methanol (10 minutes at −20°C). ⁵⁵ Then labeling was performed with FITC-conjugated anti-p53 antibody and FITC-conjugated mouse IgG2κ as a negative control. Bax labeling was performed with purified anti-Bax antibody and FITC-conjugated goat anti-mouse antibody. Purified mouse IgG2κ was used as a negative control. The results are presented in flow cytometric tracings, graphs, or bar charts. 

DNA Content Analysis.

After 72 hours of 300-U/ml IFNγ treatment, cells were trypsinized, washed with cold PBS, and fixed with 70% ethanol in PBS at −20°C. After 12 hours, samples were washed with cold PBS and stained (DNA Prep Stain), containing propidium iodide and RNase III-A, for 30 minutes at room temperature, according to the manufacturer’s instructions, then stored in the dark before analysis (within 24 hours) with the flow cytometer. 

Gel Electrophoresis and Western Blot Analysis

Cytosol and nuclei-containing cell extracts 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, 1 mM phenylmethylsulfonyl fluoride, 1 μg/ml aprotinin, 1 μg/ml pepstatin, and 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 tubes (Ependorf, Fremont, CA). The cytosol- and nuclei-containing samples were fractionated (30 μg protein per lane) by sodium dodecyl sulfate–polyacrylamide gel electrophoresis and transferred to nitrocellulose (Protran BA 83; Schleicher & Schuell, Dassel, Germany) by a semidry transfer procedure (Trans-Blot SD, Bio-Rad, Ivry sur Seine, France). 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-STAT1, anti-PARP, anti-bax, anti-Bcl-2, and anti-p53), used as recommended by the suppliers. Blots were developed using enhanced chemiluminescence reagent (ECL; Amersham, Arlington Heights, IL). 

Statistical Analysis

Flow cytometric results were calculated as arithmetic means ± SD, and significance values were calculated by means of the unpaired Student’s t-test with P < 0.05 regarded as significant. Results of colorimetric assays were calculated as arithmetic means ± SEM, and significance values were calculated by means of the two-way analysis of variance with P < 0.05 regarded as significant. All experiments in this study were performed at least in duplicate. 

Results

Cell Viability and Relative Cell Number Assays

Cell viability significantly decreased after 48 hours of treatment with 300 U/ml IFNγ (P < 0.01), but it was not modified by 30 U/ml IFNγ at any tested time point (Fig. 1 A). A significant decrease (P < 0.01) in relative cell number was observed after 72 hours of treatment with 300 U/ml IFNγ (Fig. 1B) . 

Morphologic Analysis

Cells were analyzed first with a phase-contrast microscope. There were a considerable number of dead cells in samples treated with 300 U/ml IFNγ after 48 hours of treatment, and this number increased with time of treatment. Density of adherent cells was decreased when compared with control (Fig. 2) . After DAPI staining, the supernatant cells and most of the adherent cells showed chromatin condensation and fragmentation (Fig. 3) . Characteristic apoptotic morphology of 300 U/ml IFNγ-treated cells with nuclear and cytoplasmic alterations was confirmed by electron microscopy (Fig. 4) . 

DNA Electrophoresis Assay

A DNA electrophoresis assay showed a characteristic ladder pattern in the sample treated for 72 hours with 300 U/ml IFNγ, confirming the presence of an apoptotic process (Fig. 5) . 

Expression of Apoptosis- and Inflammation-Related Proteins

Figure 6 shows the results of western blot analysis of STAT1 expression. STAT1 was detected at the basal level in untreated cells and increased in cytosolic and nuclear cell extracts from 24 hours of treatment with 30, 150, and 300 U/ml IFNγ, but the level of expression did not vary with time or with concentrations of IFNγ. Immunoblotting of PARP showed the increase of quantity of the native form of 116 kDa in nuclear cell extracts after 24 hours of treatment with 300 U/ml IFNγ and its proteolytic cleavage and the presence of an 85-kDa fragment after 48 hours, but not after 24 hours of treatment with 300 U/ml IFNγ (Fig. 7) . This cleavage signified the presence of an apoptotic process. As shown by western blot and flow cytometry, bax was slightly expressed in untreated cells, was upregulated after 48 hours of treatment with 300 U/ml IFNγ, and was increased further after 72 hours of treatment (Figs. 8 A, 8B). Western blot analysis showed no modification of Bcl-2 expression after IFNγ treatment at all tested concentrations (Fig. 9 A). As shown by western blot and flow cytometry, the expression of p53 was not modified by treatment with IFNγ at 30, 150, and 300 U/ml (Figs. 9B 9C) . 

Flow cytometric analysis of ICAM-1 showed that the protein was expressed in the cell line at the basal level, and this expression was significantly upregulated with all tested IFNγ concentrations after 24 hours of treatment. (Fig. 10 A, 10B). HLA-DR was negative at the basal level and became positive after 48 hours of treatment (Figs. 11 A, 11B). Fas was expressed at a low level in nontreated cells. The upregulation of expression was observed with all tested concentrations of IFNγ after 24 hours of treatment (Fig. 12) . The intensity of expression of these three proteins increased with time of treatment and concentrations of IFNγ. 

A 3-hour treatment with 50 μM staurosporine increased the expression of Fas (Fig. 13 A) and ICAM-1 (Fig. 14 A), whereas HLA-DR was not modified (data not shown) and remained negative in cells examined after 72 hours of recovery in normal cell medium. PKC depletion, obtained after 24 hours of application of 10 ng/ml TPA, increased the cellular expression of Fas (Fig. 13A) , with no effects on HLA-DR, still negative (data not shown) after 72 hours of recovery period in culture medium without TPA. When followed by 72 hours of 300-U/ml IFNγ treatment, staurosporine and PKC depletion potentiated an IFNγ-induced increase in Fas (Fig. 13B) , ICAM-1 (Fig. 14B) , and HLA-DR (Fig. 15) expressions. 

A 3-hour application of 10 ng/ml TPA alone induced a significant decrease in Fas expression (Fig. 13A) , whereas ICAM-1 was upregulated (Fig. 14A) . HLA-DR expression remained negative (data not shown). When followed by 72 hours of 300-U/ml IFNγ treatment, TPA induced reduction in expressions of Fas (Fig. 13B) and HLA-DR (Fig. 15) , when compared with cells treated with 300 U/ml IFNγ alone, but potentiated IFNγ-induced upregulation of ICAM-1 (Fig. 14B) . 

Therefore, all results considered, the PKC stimulator TPA had a negative effect on HLA-DR and Fas expression. The PKC inhibitor staurosporine and PKC-depleting treatment enhanced the expression of Fas, when applied alone, and potentiated IFNγ-induced stimulation of HLA-DR and Fas. PKC inhibitors and stimulator, all enhanced the basal expression of ICAM-1 and IFNγ-induced upregulation of this protein. 

Cell Cycle Alterations

DNA content analysis showed alterations of cell cycle after 48 hours of treatment with 300 U/ml IFNγ. A significant reduction in number of cells in the S–M phase was observed, and this reduction amplified with time of treatment, as shown in Figure 16 . There was no modification in cell cycle in 30 U/ml IFNγ-treated cells (data not shown). 

Discussion

Our data show that in a human conjunctival cell line IFNγ at 300 U/ml induced both apoptosis and expression of inflammatory markers. Furthermore, we observed cell cycle alterations involving a reduction in number of cells in the S–M proliferative phase and a consequent cell growth arrest. Cell cycle alterations and IFNγ inhibition of the G1 transit into the S phase has been reported in some in vitro and in vivo models. ^{56 57} Besides the determinant action of STAT1, ^{17 58} growth inhibition can involve other parallel intracellular steps. Some of them are reported in literature, such as downregulation of c-myc ⁵⁹ ori cyclin A, ⁶⁰ hypophosphorylation of the retinoblastoma gene product, ⁶⁰ and inhibition of cyclin-dependent kinase 2 (CDK2). ⁶¹  

IFNγ-induced apoptosis has been shown recently in some tumor cell lines. ^{18 19 20} Similar to cell growth arrest, the role of STAT1 also seems to be crucial in this IFNγ-mediated process and in apoptosis induced by other proapoptotic stimuli such as actinomycin D or TNFα. ⁶² In fact, recent data show that TNFα-treated U3A cells deficient in STAT1 cannot undergo apoptosis because of the absence of constitutive expression of caspases. ⁶² Therefore, STAT1 seems to play the role of a proapoptotic mediator, per se. ^{63 64} Additionally, some STAT1-transactivated gene products such as Bcl-2 family members (proapoptotic bax, bak) or the Fas receptor and STAT-induced growth arrest can also participate in a cell’s increased susceptibility to programmed cell death after IFNγ treatment. ^{19 65} In our model, upregulation of Fas; alteration of the balance between anti- and proapoptotic factors Bcl-2 and bax (Bcl-2 was not modified, whereas bax expression was stimulated by IFNγ treatment); activation of caspases, evidenced indirectly by cleavage of their substrate PARP; and cell growth arrest could explain, at least in part, the proapoptotic action of IFNγ. That we observed upregulation of STAT1 with all tested concentrations of IFNγ does not contradict the hypothesis of its important role in the induction of apoptosis (observed only after 300-U/ml IFNγ treatment). In fact, western blot analysis is only a semiquantitative method, and small differences of protein levels cannot be detected. We did not test the activity and DNA binding capacity of STAT1, however, which could be decisive in induced effects. Moreover, it cannot be excluded that in our system, IFNγ-induced growth arrest and apoptosis result from the combination of activation of STAT1 and modification of other transduction factor activity, such as that of IFN-regulating factor 1 or NF-κB, which can be modulated by IFNγ treatment. Moreover, as in the case of other intracellular molecules, the apoptotic process can lead to the cleavage of STAT1 by caspases, ⁶⁶ which could explain the absence of increase in apoptosis after 300-U/ml IFNγ treatment. 

In our model, cell exposure to the proapoptotic agent IFNγ induced the early activation of the DNA repair enzyme PARP, evidenced by the increase in PARP in nuclear cell extracts after 24 hours of treatment with IFNγ. The later proteolytic cleavage of PARP occurred only after 48 hours of treatment, signaled by the appearance of an 85-kDa fragment. The role of the activation of PARP during the apoptotic process still remains unclear. PARP inhibition in human T cells or lymphoblasts decreases apoptosis induced by different triggers such as alkylating agents. ^{67 68} In contrast, PARP-deficient primary bone marrow cells are extremely sensitive to apoptosis induced by DNA-damaging stimuli. After transfection of these PARP −/− cells, the expression of the wild-type PARP or of an uncleavable PARP mutant significantly delays cell death. ⁶⁹ Furthermore, a product of PARP cleavage was suggested to bind irreversibly to broken DNA ends, blocking the access of repair enzymes to DNA strand breaks. ⁷⁰ Thus, PARP activation seems to be induced to protect the cell from DNA damage, but the cleavage of the increased amount of PARP accelerates the apoptotic process and leads to its irreversibility. In opposition, the susceptibility to apoptosis of PARP-deficient mouse thymocytes, hepatocytes, or neurons is not modified in comparison with the wild-type cells, and PARP activation in these systems seems not to be crucial during programmed cell death. ⁷¹ Thus, further investigations involving PARP inhibitors are necessary to increase understanding of the role of PARP in our model of Chang conjunctival cells. 

Although cell cycle arrest was observed in our model, IFNγ-induced apoptosis seemed to be p53-independent. In fact, p53 did not vary (in western blot and flow cytometry analysis) after 72 hours of 300-U/ml IFNγ treatment, whereas cells underwent apoptosis. Similarly, IFNγ-induced apoptosis was reported to be p53-independent in other epithelial cell systems. ^{18 65}  

The signal transduction pathways of IFNγ inflammatory effects (upregulation of MHC II and ICAM-1) vary between different cell systems. The role of PKC in generation of these changes seems important but rather controversial and seems also to depend closely on the cell type. ^{16 30 33 72 73} IFNγ was shown to induce activation and membrane translocation of PKC in several in vitro cellular systems such as murine macrophages, human glioma and retinoblastoma cell lines, rat astrocytes, and other models. ^{26 31 74 75} PKC activation has often been reported as an inducer of ICAM-1, ^{23 32 72} whereas its effects on HLA-DR are less constant. ³³ Furthermore, PKC is reported to stimulate Fas expression in several murine and human tumor cell lines (especially in lymphocyte-derived cell lines), most likely by activation of the newly identified Fas-regulatory genes, the IPL gene, the murine TDAG51, or its human homologue TSSC3. ^{34 76 77} Therefore, we investigated whether the PKC pathway may participate in IFNγ-stimulatory effects on ICAM-1, HLA-DR, and Fas expression in the conjunctival cell line. IFNγ alone induced a very rapid increase of ICAM-1 expression in a dose- and time-dependent manner. As expected, the application of 10 ng/ml TPA, a PKC activator, elicited an increase of the steady state ICAM-1 level and significantly enhanced IFNγ stimulation of ICAM-1. However, surprisingly, the PKC inhibitor staurosporine and the PKC-depleting treatment not only had no negative effect on the IFNγ-mediated response but also enhanced IFNγ-induced upregulation of ICAM-1. Thus, our first conclusion was that as far as ICAM-1 was concerned, the IFNγ signal was independent of PKC activation and that the cytokine probably acts by stimulation of other intracellular secondary messengers. Recent data have shown that IFNγ-induced expression of ICAM-1 relies on a tyrosine kinase–dependent mechanism distinct from the PKC pathway activated by TPA. ⁷² Moreover, the ICAM-1 gene promoter possesses several consensus sites that are important for regulating gene expression, such as phorbol ester–responsive element (implied in the interaction with TPA), IFN-responsive element (gamma-activating sequence GAS), and NF-κB motif, which enable concurrent intervention of multiple modulators of ICAM-1 expression. ^{78 79} In fact, IFNγ-dependent transcriptional factor STAT1 and its interaction with the sequence GAS seems to play an important role in ICAM-1 expression in epithelial cell models in vitro, because STAT1-dominant negative mutants or a STAT1-deficient cell line U3A fail to upregulate ICAM-1 after IFNγ stimulation. ²³  

Our subsequent investigations concerned Fas and HLA-DR. In our cell line Fas was expressed at low constitutive levels, and IFNγ-induced upregulation was observed after 24 hours of treatment. HLA-DR was constitutively negative, and IFNγ stimulated its expression in a dose- and time-dependent manner, with kinetics less rapid than that of Fas or ICAM-1. The PKC inhibitor staurosporine and PKC depletion both enhanced IFNγ-induced HLA-DR and Fas upregulation, whereas TPA, a PKC activator, had an antagonist effect and tended to reduce IFNγ-stimulatory effects. Thus, the PKC activation was inhibitory for HLA-DR and Fas induction in our system. IFNγ-positive effects on HLA-DR and Fas expressions in our cell model confirmed that the PKC pathway, even if present, was not predominant after IFNγ stimulation. Consequently, we conclude that transduction pathways other than those that are PKC-related, probably those of tyrosine kinases and STAT1, controlled the effects of IFNγ in our system. This conclusion is in agreement with some recent reports of IFNγ’s effects independent of PKC in other epithelial cell lines such as the bronchial epithelial cell line NCl-H292. ⁷²  

Thus, in our in vitro model of conjunctival cells, Fas, HLA-DR, and ICAM-1 were all upregulated by the Th1 cytokine IFNγ. This observation coincides with the close correlation of intensities of expressions of these proteins in ocular surface inflammatory disorders. In fact, some recent data concerned concomitant and correlated expressions of Fas, HLA-DR, and ICAM-1 in conjunctival cells in Sjögren’s syndrome and in lacrimal gland acinar cells in Sjögren-like animal models. ^{44 80 81} Fas and HLA-DR were also shown to be simultaneously overexpressed in conjunctival epithelium in lens wear–induced inflammatory disorders and in patients undergoing topical long-term antiglaucoma treatments containing preservatives. ⁴⁴ Because the human continuous conjunctival cell line that we used in this study differs in its characteristics from normal epithelium, further investigations in vivo or in first-passage culture of human conjunctival epithelium could be required to confirm our findings and to allow better extrapolation to these pathologic states. ⁸²  

However, the coexistence of apoptosis and inflammation is one of the principal features of Sjögren’s dry eye syndrome, and both processes are improved by cyclosporin A treatment. ⁴⁶ That apoptosis and inflammation are possibly mediated by the same stimuli and transduction pathways opens up new prospects with on the therapeutics of some ocular diseases. Pharmacologic research on antiapoptotic drugs such as caspase inhibitors, very promising in other diseases such as liver or heart failures, in the future also may resolve ocular surface and lacrimal gland inflammatory disorders. 

 Submitted for publication December 21, 1998; revised March 29, 1999; accepted April 12, 1999.

 Proprietary interest category: N.

 Corresponding author: Christophe Baudouin, Service d’Ophtalmologie, Ambroise Paré Hospital, 9 Avenue Charles de Gaulle, 92104 Boulogne Cedex France. E-mail: [email protected]

 

Figure 1.

 

(A) Viability of conjunctival cells determined by MTT colorimetric assay. When compared with control, cell viability significantly decreased after 48 hours of treatment with 300 U/ml IFNγ (P < 0.01). Cell viability was not modified with 30-U/ml IFNγ treatment. (B) Relative cell number after 300-U/ml IFNγ treatment was determined by crystal violet colorimetric assay. When compared with control, there was a significant decrease in relative cell number after 72 hours of treatment.

Figure 2.

 

Phase-contrast microscopy analysis after IFNγ treatment. (A) Control cells. (B) Cells treated for 72 hours with 30 U/ml IFNγ. No alteration in cell culture is visible. (C) Cells treated for 24 hours with 300 U/ml IFNγ. Cell aspect is comparable to the control. (D) Cells treated for 48 hours with 300 U/ml IFNγ. A few round, detached cells are visible in the microscopic field. (E) Cells treated for 72 hours with 300 U/ml IFNγ. A considerable number of dead cells were present in the supernatant. The density of adherent cells was diminished when compared with the control.

Figure 3.

 

Nuclear DAPI staining after IFNγ treatment. (A) Control cells. (B) Cells treated for 72 hours with 30 U/ml IFNγ. Cell nuclei aspect is comparable to the control. (C) Cells treated for 72 hours with 300 U/ml IFNγ. Numerous nuclei show chromatin condensation and fragmentation characteristic of apoptosis. (D) Supernatant of culture treated for 72 hours with 300 U/ml IFNγ with a considerable number of apoptotic nuclei.

Figure 4.

 

Transmission electron microscopy analysis after IFNγ treatment. Cells treated for 72 hours with 300 U/ml IFNγ showed characteristic apoptotic ultrastructure, with membrane blebbing, chromatin condensation, and fragmentation. Arrowheads: autophagic vacuoles. N, nuclear fragments; b, plasma membrane blebbing; m, mitochondria.

Figure 5.

 

Nuclear DNA samples isolated from cells after 72 hours of treatment with 300 U/ml IFNγ. Lane 1: DNA molecular markers (1 kb DNA). Lane 2: Control cells. Lane 3: Apoptotic ladder pattern in cells treated for 72 hours with 300 U/ml IFNγ.

Figure 6.

 

Western blot analysis of expression of STAT1 in cells treated with IFNγ. STAT1 expression was significantly increased in (A) cytosolic and (B) nuclear extracts after 24 hours of treatment with IFNγ at 30, 150, and 300 U/ml. C, control cells; 30 U/24h, treated for 24 hours with 30 U/ml IFNγ.

Figure 7.

 

Western blot analysis of expression of PARP in nuclear cell extracts after IFNγ treatment. A proteolytic cleavage characteristic of apoptosis was observed after 48 hours of treatment with 300 U/ml IFNγ. C, control cells; 300 U/48h, cells treated for 48 hours with 300 U/ml IFNγ.

Figure 8.

 

(A) Western blot analysis of expression of bax in cytosolic cell extracts after IFNγ treatment. Control cells (C) constitutively expressed bax. Bax expression was upregulated after 48 hours of treatment with 300 U/ml IFNγ. (B) Flow cytometric analysis of bax expression. The open gray graph represents an isotypic negative control. The open black graph represents bax expression in untreated cells. The filled gray graph represents bax expression in cells treated for 72 hours with 300 U/ml IFNγ. Fifty-two percent of these cells showed a significant increase in bax expression after IFNγ treatment. 300 U/48h, 48-hour treatment with 300 U/ml IFNγ.

Figure 9.

 

(A) Western blot analysis of Bcl-2 expression in cytosolic cell extracts after IFNγ treatment. There was no modification of Bcl-2 expression after IFNγ treatment. (B) Western blot analysis of expression of p53 in nuclear cell extracts after IFNγ treatment. There was no modification of p53 expression after IFNγ treatment. (C) Flow cytometric analysis of p53 expression. The open dashed graph represents an isotypic negative control. The open black graph represents p53 expression in untreated cells. The filled gray graph represents p53 expression in cells treated for 72 hours with 300 U/ml IFNγ. There was no modification of p53 expression after IFNγ treatment. C, control cells; 300 U/48h, cells treated for 48 hours with 300 U/ml IFNγ.

Figure 10.

 

(A) Flow cytometric analysis of ICAM-1 expression after IFNγ treatment. The open black graphs represent an isotypic negative control. The percentage of cells positive for ICAM-1 is indicated above the marker lines. The relative mean intensity of ICAM-1 expression is indicated as “mean.” The control cells constitutively expressed ICAM-1. IFNγ treatment induced progressive dose- and time-dependent increase of ICAM-1 expression. (B) Graph representing a flow cytometric analysis of expression of ICAM-1 after IFNγ treatment.

Figure 11.

 

(A) Flow cytometric analysis of HLA-DR expression after IFNγ treatment. The open black graphs represent an isotypic negative control. The percentage of cells positive for HLA-DR is indicated below the marker lines. The relative mean intensity of HLA-DR expression is indicated as “mean.” The control cells were constitutively negative for HLA-DR. There was a dose- and time-dependent increase in HLA-DR expression induced by IFNγ after 48 hours of treatment. (B) Graph representing a flow cytometric analysis of the expression of HLA-DR after IFNγ treatment.

Figure 12.

 

Graph representing a flow cytometric analysis of expression of Fas antigen after IFNγ treatment. Fas was expressed at a low level in untreated cells (88% of positive cells, mean fluorescence 12). A gradual time- and dose-dependent increase in Fas expression was observed after IFNγ treatment.

Figure 13.

 

Influence of PKC-modulating treatment on Fas expression. (A) Three hours of treatment with 10 ng/ml TPA or 50 μM staurosporine or 24 hours of treatment with 10 ng/ml TPA (PKC-depleting treatment) were followed by 72 hours of recovery period in normal cell culture conditions. PKC-inhibitory treatments (staurosporine, PKC depletion) induced Fas upregulation. TPA reduced intensity of Fas expression. (B) Three hours of treatment with 10 ng/ml TPA or 50 μM staurosporine or 24 hours of treatment with 10 ng/ml TPA (PKC-depleting treatment) were followed by 300-U/ml IFNγ treatment for 72 hours. PKC-inhibitory treatments potentialized IFNγ-induced increase in Fas expression. TPA reduced IFNγ-induced upregulation of Fas.

Figure 14.

 

(A) Three hours of treatment with 10 ng/ml TPA or 50 μM staurosporine or 24 hours of treatment with 10 ng/ml TPA (PKC-depleting treatment) were followed by 72 hours of recovery period in normal cell culture conditions. PKC-inhibitory treatments (staurosporine, PKC depletion) and PKC-stimulating treatment (TPA) all induced upregulation of ICAM-1 expression. (B) Three hours of treatment with 10 ng/ml TPA or 50 μM staurosporine or 24 hours of treatment with 10 ng/ml TPA (PKC-depleting treatment) were followed by 300-U/ml IFNγ treatment for 72 hours. PKC-inhibitory (staurosporine, PKC depletion) and PKC-stimulating treatments (TPA) all potentialized IFNγ-induced upregulation of ICAM-1 expression.

Figure 15.

 

Three hours of treatment with 10 ng/ml TPA or 50 μM staurosporine or 24 hours of treatment with 10 ng/ml TPA (PKC-depleting treatment) were followed by 300-U/ml IFNγ treatment for 72 hours. PKC-inhibitory treatments (staurosporine, PKC depletion) induced a small but significant increase in IFNγ-induced expression of HLA-DR. TPA reduced IFNγ stimulatory effect on HLA-DR expression.

Figure 16.

 

Flow cytometric analysis of DNA content after IFNγ treatment. Only adherent cells were analyzed after exclusion of debris gate. (A) Control cells. Cell cycle G1 phase represented 48% of cells; 41% of cells were proliferating (S–M phase of cell cycle). (B) Cells treated for 48 hours with 300 U/ml IFNγ. Cell cycle G1 phase represents 61% of cells; 33% were in the S–M proliferating phase. (C) Cells treated for 72 hours with 300 U/ml IFNγ. The number of proliferating cells (S–M phase) is reduced to 27% of the population.

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