Modulation of TNF-α–Induced Apoptosis in Corneal Fibroblasts by Transcription Factor NF-κB

Abstract

purpose. Previous studies have suggested no role for tumor necrosis factor (TNF)-α in the modulation of apoptosis in corneal fibroblasts. However, recent investigations have demonstrated that nuclear factor (NF)-κB activation by TNF-α mediates negative apoptotic effects that must be blocked to unmask the apoptotic effects of TNF-α in vitro. The purpose of this study was to investigate the role of transcription factor NF-κB in the suppression of TNF-α–induced apoptosis of corneal fibroblasts.

methods. mRNA was detected by reverse transcription–polymerase chain reaction (RT-PCR) and RNase protection assay. Proteins were detected by immunocytochemistry and immunoprecipitation with Western blot analysis. Cell death was evaluated by trypan blue exclusion assay in corneal fibroblasts treated with TNF-α in presence or absence of the specific inhibitor of NF-κB activation, SN50, actinomycin D, or actinomycin D with dexamethasone, ketorolac tromethamine, or diclofenac sodium. Apoptosis was monitored by trypan blue exclusion, colorimetric cell assay, CPP32 activation assay, DNA fragmentation assay, and transmission electron microscopy. NF-κB activation was monitored using electrophoretic gel shift assay.

results. TNF-α, TNF receptor (R)I, and TNFRII mRNAs were detected in all three cultured corneal cell types and in ex vivo corneal epithelium using RT-PCR. TNF-α mRNA was also detected in ex vivo corneal epithelium, corneal epithelial cells, and stromal fibroblasts with the RNase protection assay. TNF-α, TNFRI, and TNFRII proteins were detected by immunocytochemistry in all three major corneal cell types in human corneal tissue. TNF-α protein was also detected in ex vivo corneal epithelium, primary corneal epithelial cells, and primary stromal fibroblasts using immunoprecipitation and Western blot analysis. TNF-α stimulated corneal fibroblast cell death when NF-κB activation was blocked with actinomycin D or SN50. Enhanced cell death was noted with dexamethasone, ketorolac tromethamine, or diclofenac sodium when used in the presence, but not in the absence, of actinomycin D. A gel shift assay revealed induction of NF-κB by TNF-α and suppression of induction in the presence of actinomycin D or SN50, but not by the control peptide SN50M.

conclusions. The TNF-α receptor system is expressed in the cornea, and NF-κB activation is an important regulator of TNF-α–mediated corneal fibroblast apoptosis. Nonsteroidal anti-inflammatory agents or corticosteroids may potentiate corneal fibroblast apoptosis in response to cytokine stimulation.

Apoptosis is a physiological form of cell death essential for the normal development and homeostatic maintenance of multicellular organisms. ^{1 2} Recently, this process has received attention because of its involvement in embryonic development, ³ autoimmune diseases, ⁴ T-cell depletion, ⁵ chemotherapeutic drug-induced killing of cancer cells, ⁶ and neurodegeneration. ⁷ Corneal studies have identified apoptosis of superficial keratocytes as an important response to herpes simplex virus infection, ⁸ scrape injury to the epithelium, ⁹ or epithelial damage associated with surgical procedures such as photorefractive keratectomy (PRK) and laser in situ keratomileusis (LASIK). ^{10 11} These studies suggested that keratocyte apoptosis has a pivotal role in modulating corneal function and led to hypotheses implicating keratocyte apoptosis in maintenance of normal corneal tissue organization and the pathophysiology of corneal diseases such as keratoconus. ^{9 11 12}  

Several cytokines, including interleukin (IL)-1, Fas ligand, bone morphogenic protein (BMP) 2 and BMP4, have been shown to trigger cytokine-mediated apoptosis of cultured corneal fibroblasts. ^{9 13 14 15 16} These studies have suggested that there is redundancy in the cytokine systems that modulate keratocyte apoptosis. The tumor necrosis factor (TNF)-α cytokine receptor system has been shown to modulate apoptosis in many cell types. ^{17 18 19 20 21 22} TNF-α has been detected in whole cornea. ^{23 24 25} Little investigation has been performed to determine the specific localization of TNF-α production in the cornea or the role of TNF-α in mediating keratocyte apoptosis. 

TNF-α is produced by neutrophils, activated lymphocytes, macrophages, and null killer cells but may be expressed by many nonimmune cell types. ²⁶ TNF-α was originally identified for its ability to initiate killing of cells, ²⁷ but it has a variety of cell type–dependent effects. Many cells are resistant to TNF-α–mediated programmed cell death, but undergo apoptosis in the presence of protein (cycloheximide) or RNA (actinomycin D) synthesis inhibitors. These include human fibrosarcoma (HT1080) cells ²⁸ and human fibroblasts (SV80). ²⁹ Such inhibitors are thought to block cytokine-stimulated expression of antiapoptosis factors. Holtmann et al. ²⁹ showed that a variety of cytokines, including TNF-α, activate both pro- and antiapoptotic mechanisms in different cell types. Thus, the fate of a particular cell exposed to TNF-α depends on the balance between these opposing pathways, which are regulated by a combination of factors. Recent studies have demonstrated that the transcription factor, nuclear factor (NF)-κB, is a key modulator of the antiapoptotic pathways triggered by TNF-α and that inhibitors such as cycloheximide and actinomycin D inhibit NF-κB activation. ^{28 30 31 32}  

In the present study, expression of TNF-α and its TNF receptor (R)I (55–60 kDa) and TNFR-II (75–80 kDa) receptors was examined in the human cornea. Studies were also performed to investigate NF-κB’s role in inhibiting apoptosis induced by TNF-α in human corneal fibroblasts. 

Materials and Methods

Unless otherwise specified all reagents were obtained from Sigma (St. Louis, MO). 

Cell Culture

Human donor corneas from infancy to 5 years of age were obtained from eye banks. These corneas were excluded from clinical use because of donor age. Primary cultures of corneal epithelial cells, fibroblasts, and endothelial cells were prepared using previously described methods. ¹³ Ex vivo corneal epithelium was collected by scraping a 7-mm area of central cornea with a scalpel blade at the time of PRK. These studies were approved by the investigational review boards at The University of Washington and The Cleveland Clinic Foundation, and the research followed the tenets of the Declaration of Helsinki. Corneal fibroblasts at passage 3 or less were used for experiments related to the in vitro induction of apoptosis. 

RT-PCR

For reverse transcription–polymerase chain reaction (RT-PCR) total cellular RNA was isolated from primary cultures of all three major cells of cornea and ex vivo corneal epithelium using Trizol reagent (Life Technologies, Gaithersburg, MD). RNA was treated with RNase-free DNase to eliminate genomic DNA contamination, and cDNA was generated as previously described. ¹³ The quality of cDNA was monitored using PCR with β-actin primers (Table 1) . cDNAs yielding a 350-bp product for β-actin mRNA without contamination with the 790-bp genomic amplification product were used for experimental amplifications. ¹³ PCR primers for TNF-α, TNFRI, and TNFRII (Table 1) were designed using primer analysis software (Oligo ver. 5.0; NBI, Plymouth, MN). PCR reactions were run using the cDNAs prepared from primary cultures of human corneal cells and ex vivo human corneal epithelium. Each 50-μl PCR reaction contained 200 ng of cDNA, 36 pg/μl of each primer, 400 μM of each dNTP, and 2.5 units of Taq polymerase (JumpStart; USB, Cleveland, OH) in a 10 mM Trizma-HCl (pH 8.3; Sigma, St Louis, MO), 50 mM KCl, 1.5 mM MgCl₂, and 0.001% gelatin. The cycling conditions were 95°C for 4 minutes, followed by 35 cycles of 95°C for 1 minute, 55°C for 30 seconds, and 72°C for 1 minute, with a final cycle at 72°C for 5 minutes. A 10-μl aliquot of PCR product was resolved on an agarose gel. ¹³ The amplification products for TNF-α, TNFRI, and TNFRII were cut from gels, cloned into PCR II cloning vector (Invitrogen, San Diego, CA), and sequenced (Sequenase 2.0; USB) according to a previously described method. ¹³  

RNase Protection Assay

Total cellular RNA was extracted from cultured cells using Trizol (Gibco, Rockville, MD). RNA was dissolved in diethyl pyrocarbonate–treated water, and the concentrations were measured with a spectrophotometer. cDNA probes for human TNF-α and β-actin were amplified using PCR with the primers listed in Table 1 . The amplification products were cloned into the pCR2.1 TA cloning vector (Invitrogen) and sequenced using standard methods to confirm the sequence. ³²P-labeled RNA probe for TNF-α andβ -actin were prepared with an RNA transcription kit (Stratagene, San Diego, CA). 

The RNase protection assay was performed using a commercially available kit (Boehringer–Mannheim, Indianapolis, IN), according to the manufacturer’s protocol. Two microliters RNA probe (1 × 10⁶ cpm/μl) were used in each assay with 50μ g RNA from a particular cell or tissue type. Test and probe RNA were precipitated in the presence of 0.3 M sodium acetate with ice-cold ethanol. RNA was recovered by centrifugation at 15,000 rpm for 15 minutes at 4°C and dissolved in 30 μl of hybridization buffer. After denaturation for 5 minutes at 90°C, the samples were incubated overnight at 42°C. Each hybridization mixture was digested with 40 units of RNase T1 and 10 units of RNase T2 for 50 minutes at 30°C, then digested with 3 μl proteinase K (20 μg/μl) in the presence of 0.5% sodium dodecyl sulfate (SDS) for 20 minutes at 37°C. The protected RNA fragments were precipitated by adding 5 μg of yeast transfer RNA and 1 ml ethanol and then extracted with 400 μl phenol-chloroform (1:1). The RNA pellet was resuspended in 7 μl loading buffer, heated for 5 minutes at 90°C, and subjected to SDS-polyacrylamide gel electrophoresis (SDS-PAGE) in a 4% polyacrylamide–7 M urea gel at 300 V in 1× TBE (0.089 M tris, 0.089 M borate, and 0.002 M EDTA) buffer. The gel was fixed with 10% acetic acid and 10% methanol and dried with a vacuum gel dryer (Bio-Rad, Richmond, CA). Dried gels were exposed overnight to film (BioMax; Eastman Kodak, Rochester, NY). The actual sizes of the protected RNA fragments were confirmed using a ³²P-labeled RNA ladder that was included on each gel. These markers were revealed with a brief exposure, and then the marker lane was cut from the dried gel to prevent overexposure of adjacent RNase-protected lanes. 

Immunocytochemistry

Corneoscleral rims were obtained from enucleated eyes of patients with orbital tumors not involving the cornea or choroidal melanomas. Corneoscleral rims were excised, embedded in HistoPrep (Fisher, Fairlawn, NJ), snap frozen in liquid nitrogen, and stored at −85°C. Informed consent was obtained from each patient before surgery. Seven-micrometer-thick sections were cut with a cryostat (Reichert–Jung; Leica, Deerfield, IL), placed on slides (Superfrost plus; Fisher), and frozen at −85°C until they were used for staining. 

Goat polyclonal antibodies (Santa Cruz Biotechnology, Santa Cruz, CA) for TNF-α (sc-1350), TNFRI (sc-1067), and TNFRII (sc-1071) were used at a final concentration of 1 μg/ml. TNF-α (sc-1350P), TNFRI (sc-1067P), and TNFRII (sc-1071P) blocking peptides (Santa Cruz) were used for preabsorption control. Peptide neutralization was performed overnight at 4°C by combining 10 μg/ml of peptide with 1 μg/ml of the corresponding antibody. This mixture was used in control staining. Control procedures were performed by omitting any primary antibody. 

Two normal human corneas from different adults were used to perform immunocytochemistry for TNF-α and its receptors. Tissue sections were fixed with acetone at −20°C for 10 minutes, and immunocytochemistry was performed using standard methods with a commercial kit (Universal LSAB+; Dako, Carpinteria, CA) according to a previously described method. ¹³  

Immunoprecipitation and Western Blot Analysis

Cell pellets or ex vivo epithelial tissue was lysed in 5 ml of lysis buffer (50 mM Tris/Cl (pH 8.0), 0.5% Triton X-100, 10% glycerol, 0.2 mM EDTA, 150 mM NaCl, 1 mM dithiothreitol (DTT), 0.5 mM phenylmethylsulfonyl fluoride (PMSF), 3 μg/ml aprotinin, 2 μg/ml pepstatin, and 1 μg/ml leupeptin) on ice for 20 minutes. The extracts were centrifuged at 15,000 rpm in a microcentrifuge for 10 minutes at 4°C. Supernatants were decanted into a fresh tube, and the protein concentration of each extract was determined (Protein Assay Kit; Bio-Rad). Five hundred micrograms of lysate at 1 mg/ml was incubated with preimmune serum (2.5 μl) containing Protein A Sepharose 6MB (Pharmacia, Piscataway, NJ) for 1 hour, and the lysate was clarified by brief centrifugation in a microcentrifuge at 15,000 rpm. The lysate was incubated with 10 μg of antibody and 50 μl Protein A Sepharose 6MB overnight at 4°C with continuous mild agitation. Sepharose beads were washed three times in cell lysis buffer, and the bound proteins were eluted in SDS gel loading buffer by boiling. SDS-PAGE was performed. Immunoblotting was performed by chemiluminescence (ECL System; Amersham, Arlington Heights, IL) according to the manufacturer’s instructions. The anti-TNF-α antibody sc-1350 was used for immunoprecipitation and Western blot analysis. 

Trypan Blue Exclusion Assay

Human corneal fibroblast cells (passages 1–3) were seeded at a density of 1 × 10⁴ cells/well in standard six-well plates (Falcon, Franklin Lakes, NJ) in Eagle’s modified essential medium (EMEM) supplemented with 10% fetal bovine serum (FBS; Life Technologies, Waltham, MA). The optimum time and dose of TNF-α and inhibitors used to trigger cell death were selected by preliminary experiments (data not shown). In the reported experiments, the medium was changed to EMEM with 0.5% serum at 24 hours after plating. TNF-α (10 ng/ml, R&D Systems, Minneapolis, MN) in presence or absence of actinomycin D (5 μg/ml) alone or actinomycin D (5 μg/ml) with dexamethasone (5 μM; Sigma), diclofenac sodium (1 μg/ml; CIBA Vision Ophthalmics, Atlanta, GA), or ketorolac tromethamine (25μ g/ml; Allergan, Irvine, CA) were added to the wells. Control wells without TNF-α, but with the inhibitors, were also included. 

Six wells were used for each treatment, and cells were incubated for 8 hours before analysis. Cells were trypsinized, resuspended in 0.5 ml of Hanks’ balanced salt solution, and stained by adding 0.5 ml of 0.4% trypan blue for 5 minutes. All the stained and unstained cells were counted in 10 squares of a hemocytometer and the percentage of cell death calculated. The assay was repeated three times. 

DNA Fragmentation Assay

The SN50 cell-permeable peptide (Biomol, Plymouth Meeting, PA) that acts as a specific inhibitor of translocation of the NF-κB active complex into the nucleus ¹⁶ and the nonfunctional SN50M control peptide (Biomol) were used according to the manufacturer’s protocol. Second- or third-passage corneal fibroblasts (70–80% confluent) were exposed to vehicle, TNF-α (30 ng/ml), actinomycin D (5 μg/ml), TNF-α (30 ng/ml) with SN50M control peptide (100 μg/ml), TNF-α (30 ng/ml) with SN50 peptide (100μ g/ml), or TNF-α (30 ng/ml) with actinomycin D (5 μg/ml). The DNA fragmentation assay was performed (Apoptotic DNA Laddering Kit; Boehringer–Mannheim) according to the manufacturer’s instructions. Isolated DNA was analyzed on ethidium bromide–stained agarose gels. A positive control DNA ladder generated in U937 cells was also included on the gel. The experiment was repeated five times. 

Apopain/CPP32 Assay

An apopain assay kit (FluorAce; Bio-Rad) that detects activation of the caspase CPP32 was used according to the manufacturer’s instructions. For preparation of cell extract, 1 × 10⁶ second- or third-passage corneal fibroblasts were seeded in 100-mm culture plates using EMEM supplemented with 10% FBS. When the cells were 70% to 80% confluent, they were stimulated with vehicle or TNF-α (10 ng/ml) in EMEM with 0.5% FBS containing actinomycin D (5 μg/ml), SN50 peptide (100 μg/ml), SN50M control peptide (100 μg/ml), or additional vehicle. SN50 peptide, but not the control SN50M peptide, has been shown to be a specific inhibitor of NF-κB activation. ¹⁶ Incubation was performed in a 37°C humidified incubator for 22 to 24 hours in 5% CO₂. Floating cells were harvested by transferring the media into a 15-ml conical tube, centrifuging at 500 rpm for 5 minutes, and discarding the supernatant. These cells were washed twice with 5 ml of cold 1× phosphate-buffered saline (PBS) and pelleted at 500 rpm. The cells attached to the plates were lysed with 15 ml apopain lysis buffer containing 10 mM HEPES (pH 7.4), 2 mM EDTA, 5 mM DTT, 1 mM PMSF, 0.1% 3-([3-cholamidopropyl] dimethylammonio)-1-propane sulfonate (CHAPS), 10 μg/ml pepstatin A, 10 μg/ml aprotinin, and 20 μg/ml leupeptin; scrapped with a plastic cell scraper; and added to the matched floating cell pellets. Tubes were vortexed gently to resuspend the cell pellets and freeze-thawed five to six times by alternatively transferring between an isopropanol dry-ice bath and a 37°C water bath. The cell extracts were transferred to microfuge tubes and spun at 14,000 rpm for 30 minutes at 4°C. The supernatant was either used immediately for assay or stored at −80°C for later analysis. The protein content of each sample was determined by the Bradford method. ³³  

To determine apopain activity, 96-well plates were used to perform the apopain calibration curve and apopain activity assay on cell-derived samples, according to the manufacturer’s instructions. In brief, 8μ l 25× reaction buffer and 2 μl caspase inhibitor ZDEVD-AFC were added to 75 μg protein (30–40 μl cell extract). The final volume of the reaction was brought to 200 μl by adding sterile water. The blank (no protein) and positive control (apopain provided in the kit) reactions were performed at the same time. The plate was incubated at room temperature for 1 hour, and absorbance was determined in a reader (Cyto Fluor II; PerSeptive Biosystems, Framingham, MA) by excitation at 360 nm and measurement of emission at 530 nm. The apopain-CPP32 activity expressed in apopain units was calculated using the manufacturer’s method. 

Transmission Electron Microscopy

Approximately 80% confluent corneal fibroblasts were exposed to TNF-α (10 ng/ml), actinomycin D (5 μg/ml), or TNF-α (10 ng/ml) with actinomycin D (5 μg/ml) for 8 hours. Floating cells in the medium were collected by centrifugation at 500 rpm. The monolayers of remaining cells were trypsinized, transferred to the corresponding tubes containing floating cell pellets, and washed with medium containing 10% FBS. The cells were pelleted in 1.5-ml tubes (Eppendorf, Fremont, CA) and fixed in a solution of 3% glutaraldehyde and 1% paraformaldehyde. TEM sections were cut at 70 nm and stained with 3% uranyl acetate for 15 minutes, followed by 3 minutes in Reynold’s lead citrate. TEM was otherwise performed as previously described. ¹³  

Colorimetric Cell Assay

The assay was performed by nonradioactive cell proliferation assay kit (Celltiter 96 Aqueous Assay; Promega, Madison, WI) according to the manufacturer’s instructions. Briefly, 2000 corneal fibroblasts (passage 2 or 3) per well were seeded in a 96-well plate. At 80% confluence, the cells were stimulated with either vehicle or TNF-α (10 ng/ml), with each also containing actinomycin D (5 μg/ml), SN50 peptide (100 μg/ml), SN50M control peptide (100 μg/ml), or vehicle, in a total volume of 100 μl EMEM containing 0.5% FBS. Incubation was performed at 37°C in a humidified 5% CO₂ incubator. Six wells were used for each treatment. After 6 hours, 50μ l of freshly prepared assay solution was added to each well. The plate was further incubated for 2 to 3 hours at 37°C and examined with an inverted microscope. Absorbance was measured at 490 nm using a microplate reader (Thermomax; Molecular Devices, Sunnyvale, CA). 

Gel Shift Assay

Activation of NF-κB was monitored (Gel Shift Assay Core System; Promega) according to the manufacturer’s instructions. Seventy to 80% confluent corneal fibroblasts in EMEM containing 0.5% FBS were treated with vehicle, TNF-α (10 ng/ml), or TNF-α (10 ng/ml) with actinomycin D (5 μg/ml) for 1 to 3 hours at 37°C. The actinomycin D treatment was performed 20 minutes before TNF-α stimulation. A nuclear extract of the cells was then prepared. Briefly, cells were washed twice with cold 1× PBS, harvested with 1 ml cold PBS, and spun at 13,000 rpm at 4°C for 2 minutes. Supernatants were removed, cells were extracted on ice for 5 minutes in 100 μl of buffer A (10 mM HEPES [pH 7.9], 1.5 mM MgCl₂, 10 mM KCl, 0.5 mM DTT, and 0.5 mM PMSF), and resuspended completely by adding 0.9 ml of buffer A. The mixture was centrifuged at 13,000 rpm for 2 minutes at 4°C. The resultant cell pellet was washed with 0.5 ml of buffer A, resuspended in 100 μl of buffer B (20 mM HEPES [pH 7.9], 0.42 M NaCl, 1.5 mM MgCl₂, 0.5 mM DTT, 0.5 mM PMSF, 0.1 mM sodium vanadate, 25% glycerol, 0.5 μg/ml pepstatin, 0.5 μg/ml leupeptin, and 0.5 μg/ml aprotinin), and extracted on ice for 30 minutes with vortexing at 5-minute intervals. The resultant supernatant was aliquoted and either used immediately or stored at −80°C until analysis was completed. The protein content was measured by the method of Bradford. ³³  

The DNA binding reaction was performed by incubating 4 μg of the extract with 16 femtomoles of the ³²P-labeled NF-κB consensus sequence oligonucleotide ³⁴ for 20 minutes at room temperature. The incubation mixture included 1 μg of poly (dI.dc) from the kit in binding buffer (25 mM HEPES [pH 7.9], 50 mM NaCl, 1 mM DTT, and 1 mM EDTA). The DNA-protein complex that formed was separated from free oligonucleotide on 4% nondenaturing polyacrylamide gels at 150 V for 4 hours using buffer containing 50 mM Tris (pH 8.5), 200 mM glycine, and 1 mM EDTA. After electrophoresis, the gels were dried, and labeled DNA was detected by autoradiographic exposure for 16 hours. The specificity of binding was examined by competition with the unlabeled oligonucleotide. 

Statistical Analysis

Statistical analyses were performed using an analysis of variance program (StatView 4.5; Abacus Concepts, Berkeley, CA). The Bonferroni–Dunn test was performed for multiple comparisons. 

Results

Messenger RNA coding for TNF-α, TNFRI, and TNFRII were detected in primary human corneal epithelial cells (HCE), ex vivo human corneal epithelium (ex vivo HCE), primary corneal fibroblasts (HSF), and primary endothelial cells (HCN) using RT-PCR (Fig. 1) . Nucleic acid sequencing of the PCR products of TNF-α, TNFRI, and TNFRII conclusively demonstrated that the amplified products were derived from TNF-α, TNFRI, and TNFRII, respectively. An additional band at 350 bp in TNFRI (Fig. 1B) was found to be PCR artifact through DNA sequencing. TNF-α messenger RNA was also confirmed to be present in ex vivo corneal epithelium, cultured corneal epithelial cells, and cultured stromal fibroblasts using the RNase protection assay (Fig. 2) . 

The proteins for TNF-α, TNF-RI, and TNFRII were detected in normal human corneas by immunocytochemistry. Figure 3 demonstrates localization of TNF-α (Figs. 3A 3B) , TNFRI (Figs. 3C 3D) , and TNFRII (Figs. 3E 3F) proteins in corneal epithelial cells, keratocytes, and corneal endothelial cells. Figures 3G and 3H show control immunostaining in which the primary antibody was omitted. The individual control immunocytochemistries for TNF-α, TNFRI, and TNFRII were also performed using preabsorption of primary antibody with control peptide. These preabsorption controls (data not shown) were indistinguishable from controls shown in Figures 3G and 3H with omitted primary antibody. 

TNF-α protein of the expected size of 19 kDa was detected in ex vivo corneal epithelium, cultured corneal epithelial cells, and cultured stromal fibroblasts by immunoprecipitation and Western blot analysis (Fig. 4) . TNF-α protein of the expected size was also detected by this method when Western blot analysis was performed with control TNF-α (Fig. 4) . 

Cell death induced by TNF-α was first monitored by trypan blue exclusion assay. The optimal dosages of TNF-α and other agents were established by performing preliminary dose- and time-dependent cell toxicity assays (data not shown). Cell death of 24% ± 5% (P < 0.01) was noted in corneal fibroblasts stimulated with TNF-α (10 ng/ml) for 8 hours in the presence of actinomycin D (5μ g/ml). Cell death was not induced by TNF-α in the presence of the nonsteroidal anti-inflammatory drug ketorolac tromethamine or diclofenac sodium or the corticosteroid dexamethasone if actinomycin D was not included in the incubation. However, statistically significant enhanced cell death was noted when ketorolac tromethamine (25 μg/ml; 39% ± 6%, P < 0.001), diclofenac sodium (1 μg/ml; 83% ± 11%, P < 0.001), or dexamethasone (5 μM; 40% ± 8%, P < 0.001) was added to stromal fibroblast cells treated with TNF-α and actinomycin D. A minimum of six wells were tested for each condition in each experiment. The experiment was repeated with similar results each time. 

Laddering consistent with apoptosis was detected in DNA isolated from cultures of corneal fibroblasts exposed to 30 ng/ml TNF-α with 100μ g/ml SN50 for 8 hours or 30 ng/ml TNF-α with 5 μg/ml actinomycin D for 8 hours, but not with vehicle, 30 ng/ml TNF-α alone, 5 μg/ml actinomycin D alone, or 30 ng/ml TNF-α with 100 μg/ml SN50M control peptide (Fig. 5) . DNA laddering was also detected when corneal fibroblasts were treated with ketorolac tromethamine, diclofenac sodium, or dexamethasone in the presence of TNF-α and actinomycin D (not shown). The DNA laddering assay is qualitative and therefore an increase in DNA laddering was not noted when these agents were added along with TNF-α and actinomycin D. 

No cell morphologic changes were observed with TEM in cultures stimulated with either TNF-α (10 ng/ml) or actinomycin D (5 μg/ml) alone (Figs. 6A 6B 6C 6D) . The morphologic features of apoptosis that include cell shrinkage, chromatin condensation, chromatin fragmentation, and cell blebbing were detected by TEM in approximately 20% of corneal fibroblast cells exposed to TNF-α (10 ng/ml) and actinomycin D (5μ g/ml) for 8 hours (Figs. 6E 6F 6G 6H) . Identical morphologic changes were observed by TEM in cultures exposed to TNF-α and actinomycin D with ketorolac tromethamine, diclofenac sodium, or dexamethasone (data not shown). 

Cell death triggered in corneal fibroblasts by TNF-α in the presence of the specific inhibitor of NF-κB activation SN50, compared with the inactive control SN50M peptide or actinomycin D alone (vehicle), was determined using the colorimetric cell assay (Fig. 7) . Wells were inspected with an inverted microscope to confirm that a decrease in dehydrogenase enzyme release (associated with lower color development in the Celltiter 96 Aqueous Assay; Promega) correlated with an increase in corneal fibroblast cells that had died, rounded up, and dissociated from the plate (25–30% of total cells in wells containing TNF-α with actinomycin D or TNF-α with SN50). Cell death was quantitated by determining the absorbance of the wells. These experiments confirmed cell death induced by TNF-α with either the nonspecific (actinomycin D) or specific (SN50) inhibitor of NF-κB activation. 

The apopain assay detects activation of the apoptosis-specific caspase CPP32. This assay was used to confirm that corneal fibroblast cell death in response to TNF-α with actinomycin D or SN50 was apoptosis (Fig. 8) . A statistically significant increase in apopain activity was noted in cultures stimulated with TNF-α in the presence of actinomycin D (approximately fourfold, P < 0.05) or SN50 (approximately threefold, P < 0.05) over TNF-α alone. No significant increase in apopain activity was noted in cells stimulated with TNF-α in the presence of SN50M. The positive control apopain provided by the manufacturer induced an approximate fourfold increase in activity over the negative control. 

The effect of TNF-α or TNF-α with actinomycin D, SN50 peptide, or SN50M control peptide on the nuclear translocation (activation) of NF-κB was evaluated in human corneal fibroblast cells by electrophoretic mobility gel shift assay (Fig. 9) . TNF-α induced nuclear translocation of NF-κB in corneal fibroblast cells. A significant decrease in TNF-α–induced NF-κB activation was noted in the presence of actinomycin D or SN50 peptide. The control peptide SN50M had no effect on the activation of NF-κB induced by TNF-α. HeLa cell extract was used as a positive control for this experiment. 

Discussion

TNF-α modulates apoptosis in many cell types ^{17 18 19 20 21 22} through two receptors, TNFRI and TNFRII, which have sequence similarities only in the extracellular domains and are coded from different genes. ^{35 36} TNFRI is predominant in mediating TNF-α effects on apoptosis. ³⁷ At least one of these TNFR types is expressed by virtually every cell type. ^{35 36} Thus, it is not surprising that we detected expression of TNFRI and TNFRII mRNAs and proteins in corneal epithelial, keratocyte, and endothelial cells, suggesting that TNF-α can modulate functions in these cells. 

TNF-α cytokine production is commonly associated with immune cells such as neutrophils, activated lymphocytes, macrophages, and null killer cells. ²⁶ Thus, it is likely that TNF-α derived from immune cells that invade the cornea could modulate parenchymal cell functions associated with inflammatory, infectious, or wound-healing conditions. Expression of TNF-α by immune cells that invade the cornea during infection by herpes simplex virus has recently been demonstrated. ³⁸ A previous study did not detect TNF-α in corneal cells, ³⁹ although others found TNF-α in whole cornea or corneal fibroblasts. ^{22 23 24} Therefore, we were surprised to detect TNF-α mRNA in cultured corneal epithelial, fibroblast, and endothelial cells using RT-PCR (Fig. 1) . This may have been attributable to changes in expression after culturing the cells in vitro, but TNF-α mRNA was also detected in ex vivo corneal epithelium that was transferred directly from the cornea into RNA extraction reagent (Fig. 1) . In subsequent experiments, TNF-α mRNA was also detected in corneal cells, including ex vivo corneal epithelium, using the RNase protection assay (Fig. 2) . TNF-α protein was also detected in the major cell types of the cornea using a sensitive immunocytochemical technique (Fig. 2) and immunoprecipitation with Western blot analysis (Fig. 3) . Thus, our data suggest that TNF-α is produced, at least at low levels, by each of the parenchymal cells of the normal unwounded cornea. TNF-α expression has been detected recently in other parenchymal cells, including those in smooth muscle, ⁴⁰ skin, ⁴¹ and apocrine glands. ⁴² Therefore, it may be that TNF-α has a broader range of expression than has been generally appreciated, because few studies have been performed to detect the cytokine in parenchymal cells of organs other than the immune system. Further study is needed to determine whether altered expression occurs after corneal injury and the subsequent wound-healing response. 

No effect of TNF-α on corneal fibroblast apoptosis was noted in previous studies. ⁹ Those studies, however, were performed before the discovery that the effect of TNF-α on a particular cell was influenced by the activation status of the transcription factor NF-κB. ^{28 31 32} In the present study, TNF-α triggered apoptosis of corneal fibroblasts in the presence of a nonspecific (actinomycin D) or specific (SN50) inhibitor of NF-κB activation. The electrophoretic mobility gel shift experiments (Fig. 9) demonstrate that both of these inhibitors effectively inhibited NF-κB activation in corneal fibroblasts in response to TNF-α. Apoptosis was confirmed by characteristic DNA fragmentation, cellular morphologic changes noted by TEM, and activation of the caspase CPP32 by assay. In other experiments, it was noted that TNF-α had no effect on corneal fibroblast proliferation, with or without NF-κB activation inhibitors, but had a positive chemotactic effect on stromal fibroblasts in the absence of inhibitors of NF-κB activation (data not shown). 

NF-κB is a member of the rel family. ⁴³ It is known to regulate a wide variety of genes involved in diverse biologic processes such as cell growth, adhesion, and apoptosis. ^{43 44 45} For example, NF-κB inhibits apoptosis in mice, and transgenic mice that show no expression of the NF-κB p65 Rel-A gene die embryonically of extensive liver apoptosis. ⁴⁶ Release of NF-κB from I-κB in the cytoplasm (activation) results in the translocation of NF-κB to the nucleus where it regulates the expression of specific genes that control these responses of the cell. NF-κB is activated by cytokines such as TNF-α, Fas ligand, or IL-1α ^{28 30 31 43} that have been implicated in the regulation of keratocyte apoptosis in the cornea. ^{9 11 13 15} The downstream effectors controlled by NF-κB compete with the proapoptotic effectors triggered by these same cytokines. ^{29 31 32 47 48} If antiapoptotic influences predominate, then programmed cell death is the result. Thus, the balance between NF-κB activation and other effects of the cytokines probably determines the fate of keratocytes exposed to TNF-α, IL-1α, and other cytokines that mediate their effects through pathways regulated by NF-κB. 

NF-κB could be an attractive target in acutely regulating cytokine-mediated apoptosis of keratocyte cells to clinical advantage. For example, if the hypothesis ^{10 11} that keratocyte apoptosis is an initiator of the corneal wound-healing response is ultimately found to be correct, then pharmacologic agents that promote NF-κB activation could inhibit the corneal wound-healing response associated with corneal surgery. Studies are in progress to examine this possibility. 

This study demonstrated that corticosteroids or nonsteroidal anti-inflammatory drugs (NSAIDs) such as diclofenac sodium and ketorolac tromethamine could augment TNF-α’s effects on apoptosis of corneal fibroblasts. NSAIDs ⁴⁹ and corticosteroids ^{50 51} have been shown to inhibit NF-κB activation. NSAIDs and corticosteroids could also promote the proapoptotic effects of other cytokines that stimulate transcription factor NF-κB activation such as IL-1 and Fas ligand. Thus, NSAIDs or corticosteroids could augment keratocyte apoptosis triggered by cytokines in response to corneal epithelial injury by inhibiting the antiapoptotic cascades triggered by these cytokines and tipping the balance toward apoptosis in the affected cells. Anecdotal clinical observations have suggested that NSAIDs can be used to promote regression after PRK and other corneal surgical procedures. This effect could be mediated by promotion of keratocyte apoptosis and, therefore, an augmented wound-healing response. ¹¹ Surprisingly, prospective randomized clinical studies have demonstrated no effect of corticosteroids on regression after injury produced during PRK. ⁵² One explanation for this observation could be that the marked anti-inflammatory effects of the corticosteroids are balanced by the proapoptotic effect on keratocytes. Shimoyama et al. ⁵³ showed that dexamethasone inhibited NF-κB activation in corneal endothelial cells. Thus, corticosteroids could influence the apoptotic process in other cell types in the cornea besides keratocytes. Further studies are warranted to explore the potential effects of NSAIDs and corticosteroids after corneal surgical procedures. 

 Presented at the annual meeting of the Association for Research in Vision and Ophthalmology, Fort Lauderdale, Florida, May 1998 and May 1999.

 Supported in part by US Public Health Service Grant EY-10056 from the National Eye Institute and an unrestricted grant from Research to Prevent Blindness.

 Submitted for publication June 17, 1999; revised December 9, 1999; accepted December 28, 1999.

 Commercial relationships policy: N.

 

Table 1.

 

Primers Used in RT-PCR to Detect TNF-α and TNFR mRNAs

Modulator	PCR Length	Upstream Primer	Downstream Primer	Accession Number
TNF-α	423	TGTAGCCCATGTTGTAGCAAA (Exon 3)	CAAAGTAGACCTGCCCAGACT (Exon 4)	X01394
TNFRI	418	TCCTTCACCGCTTCAGAAAA (Exon 3)	GGGATAAAAGGCAAAGACCAA (Exon 7)	M33294
TNFRII	444	AACTGGGTTCCCGAGTGCTTG (Exon 3)	AGTGCTGGGTTCTGGAGTTGG (Exon 6)	M55994
[Beta]-Actin	350	AGGCCAACCGCGAGAAGATGACC (Exon 3)	GAAGTCCAGGGCGACGTAGCAC (Exon 4)	X00351

Figure 1.

 

Detection of TNF-α (top), TNFRI (middle), and TNFRII (bottom) mRNAs in human corneal cells and tissues using RT-PCR. Amplification was performed with two independent cDNA samples from human ex vivo corneal epithelium removed at the time of PRK (ex vivo HCE), human primary (1°) corneal epithelial cells (HCE), human primary corneal fibroblasts (HSF), and human corneal endothelial cells (HCN). Amplification products of the expected size for TNF-α, TNFRI, and TNFRII were detected in each sample. Amplification products of the expected sizes are indicated by arrows labeled with the size of the product in base pairs. The additional band at approximately 350 bp for TNFRI was noted to be an artifact by nucleic acid sequencing. C, negative controls with water as the target rather than cDNA; L, a 100-bp (BP) DNA marker with base pairs indicated to the left.

Figure 2.

 

Detection of TNF-α mRNA in corneal cells using the RNase protection assay. The protected band of the expected size of 423 bases for TNF-α was detected in two independent samples each of ex vivo corneal epithelium (Ex vivo HCE-1 and Ex vivo HCE-2), primary stromal fibroblast cells (HSF-1 and HSF-2), and primary cultured human corneal epithelial cells (HCE-1 and HCE-2). The 350-bp protected band in each sample is β-actin as an indicator of mRNA loading. HeLa cells were included as positive controls.

Figure 3.

 

Immunocytochemical detection of TNF-α, TNFRI, and TNFRII proteins in fresh-frozen human corneal tissue. (A, B) TNF-α protein was detected in corneal epithelium (e), keratocytes (arrowheads) in the stroma (s), and endothelial cells (arrows). (C, D) TNFRI protein was detected in all three corneal cell types. High expression of this protein was present throughout the corneal epithelium. Arrowheads in (C) indicate keratocytes and arrows in (D) indicate endothelial cells. (E, F) TNFRII protein was detected in epithelium (e, arrows in E), in keratocytes (arrowheads) in the stroma (s), and in the endothelium (arrows in F) of human cornea. (G, H) Cornea stained by omitting the primary antibody (identical with preabsorption controls that were run individually for TNF-α, TNFRI, and TNFRII). Arrows in (H) indicate endothelial cells. Magnification, ×400.

Figure 4.

 

Immunoprecipitation and Western blot analysis for TNF-α in human corneal tissue and cells. TNF-α protein of the expected size of 19 kDa was detected in ex vivo corneal epithelium (ex vivo HCE-1 and ex vivo HCE-2), primary cultured corneal epithelial cells (HCE-1 and HCE-2), primary cultured stromal fibroblasts (HSF-1 and HSF-2), and HeLa cells. Ab, antibody used for immunoprecipitation alone as a control. Control protein (TNF-α Con) was also included as a positive control but was included in the Western blot without prior immunoprecipitation.

Figure 5.

 

DNA fragmentation assay. Fragmentation consistent with apoptosis was detected in corneal fibroblast cells exposed to 30 ng/ml TNF-α with NF-κB inhibitor SN50 at 100 μg/ml or 30 ng/ml TNF-α with 5μ g/ml actinomycin D. No fragmentation was noted in cultures exposed to TNF-α vehicle, 30 ng/ml TNF-α alone, 5 μg/ml actinomycin D alone, or 30 ng/ml TNF-α with 100 ng/ml SN50M inactive control peptide. A 100-bp DNA marker was included on the gel with sizes of representative bands indicated in base pairs to the left. Control ladder indicates a positive control apoptosis ladder generated in U937 cells.

Figure 6.

 

TEM of sections of stromal fibroblasts exposed to (A, B) TNF-α alone (30 ng/ml) or (C, D) actinomycin D alone (5 μg/ml) did not show signs of apoptosis. (E, F) Cells exposed to TNF-α (30 ng/ml) along with actinomycin D (5 μg/ml) had evidence of cell death with the hallmarks of apoptosis, including chromatin condensation, fragmentation, and cell blebbing. Arrows in (E, F) indicate condensed chromatin consistent with apoptosis. c, normal chromatin pattern; n, nucleolus in some cell nuclei. Magnification, ×1650.

Figure 7.

 

Cell death stimulated by TNF-α in the presence of actinomycin D, SN50, or SN50M was evaluated using the colorimetric cell assay. Graphic presentation of the percentage absorbance decreases for different treatment conditions compared with the vehicle-treated control.* Significant difference compared with TNF-α with SN50M or vehicle.

Figure 8.

 

Detection of apoptotic death by apopain assay. The activation of CPP32 was measured in corneal fibroblasts exposed to TNF-α in the presence of actinomycin D, SN50 peptide, and control SN50M peptide compared with the vehicle containing TNF-α alone. The apopain provided in the kit was used as a positive control. Significantly higher apopain activity was noted in cultures treated with TNF-α with actinomycin D or SN50, but not in TNF-α with SN50M, over the control. *Significantly increased activity compared with vehicle with TNF-α alone or TNF-α with SN50M.

Figure 9.

 

Evaluation of NF-κB activation by electrophoretic mobility gel shift assay in human corneal fibroblast cells stimulated with TNF-α (30 ng/ml), TNF-α (30 ng/ml) with SN50 (100 μg/ml), TNF-α (30 ng/ml) with actinomycin D (5 μg/ml), or TNF-α (30 ng/ml) with SN50M control peptide (100 μg/ml). Comp probe: specific competition with unlabeled NF-κB consensus oligonucleotide in a sample derived from corneal fibroblast cells stimulated with 30 ng/ml TNF-α. HeLa cells served as a positive control. NFkB AC and arrow: the activation complex with NFκB bound to the radiolabeled probe. Arrowhead: the probe alone.