Induction of Intercellular Adhesion Molecule-1 by Tumor Necrosis ...

Induction of Intercellular Adhesion Molecule-1 by Tumor Necrosis Factor-a Through the 55-kDa Receptor Is Dependent on Protein Kinase C in Human Retinal Pigment Epithelial Cells Brian D. Sippy,* Florence M. Hofman,*\ Albion D. Wright,* Jin Lin Wang,* Rayudu Gopalakrishna,% Usha Gundimeda,X Shikun He,* Stephen J. Ryan,\# and David R Hinton*§\\

Purpose. To determine second messenger signaling pathways associated with tumor necrosis factor-a (TNF)-mediated induction of intercellular adhesion molecule (ICAM)-l expression on human retinal pigment epithelial (HRPE) cells, a cell type known to express only the 55kDa TNF receptor (TNFR p55). Methods. SV40-immortalized HRPE (SVRPE) cells were exposed to TNF with and without pretreatment with the protein kinase C (PKC) inhibitor calphostin C or the protein kinase A (PKA) inhibitor H8. SV40-immortalized HRPE cells also were treated with the PKC activator phorbol 12myristate 13-acetate (PMA) or with the PKA activators forskolin plus 3-isobutyl-l-methyl-xanthine or dibutyryl cyclic adenosine monophosphate (cAMP) alone. Membrane fractions from untreated and treated SVRPE cells were assayed for PKC activity, and whole cell lysates were assayed for cAMP accumulation and PKA activity. Flow cytometry was performed on SVRPE cells using a monoclonal antibody specific to ICAM-1. Results. Activation of TNFR p55 on SVRPE cells with TNF resulted in a rapid increase of PKC activity at 1 minute, with a subsequent downregulation to baseline. There was no increase in intracellular cAMP accumulation or PKA activity within the first 10 minutes; however, both increased within 30 minutes and returned to baseline within 1 hour. SV40-immortalized HRPE cells treated with TNF for 1 hour showed maximal induction of ICAM-1 expression at 18 hours. ICAM-1 induction by TNF treatment was inhibited by calphostin C pretreatment and not by H8 pretreatment Protein kinase C activation with PMA for 3 hours was sufficient to induce ICAM-1 on SVRPE cells at 18 hours, whereas treatment with the PKA activators forskolin or dibutyryl cAMP did not induce ICAM-1 expression. Conclusions. Tumor necrosis factor sequentially activates the PKC and PKA pathways in SVRPE cells by way of the TNFR p55. The PKC pathway is necessary for TNF-mediated ICAM-1 upregulation, and specific activation of the PKC pathway with PMA is sufficient to induce ICAM-1 on these cells. SV40-immortalized HRPE cells may serve as a model in which to study further the functional signaling pathways associated with TNFR p55. Invest Ophthalmol Vis Sci. 1996; 37:597-606.

.Human retinal pigment epithelial (HRPE) cells re-

From the Defmrtments of * Pathology, f Ophthalmology, %Cell and Neurobiology, ^Neurology, and ^Neurological Surgery, University of Southern California School of spond to tumor necrosis factor-a (TNF) by proliferaMedicine, and the #Doheny Eye Institute, Los Angeles, California. tion, cytokine production, and adhesion molecule exSupported in part by National Institutes of Health grants EYO2061 and EYO3040. The, University of Southern Californoia Department of Ophthalmology is the recipient pression, including intercellular adhesion molecule of an award from Research to Prevent Blindness, Inc. (New York, New York). (ICAM)-!.1"6 Although many of the functions associSubmitted for publication May 22, 1995; revised December 4, 1995; accepted ated with TNF activation of HRPE cells are known, the December 5, 1995. Proprietary interest calagory: N. mechanisms through which TNF acts remain elusive. Reprint requests: David R Hinton, Department of Pathology, University of Southern California School of Medicine, HMR 209, 2011 Zonal Avenue, Las Angeles, CA 90033. Tumor necrosis factor activates human cells by trig-

Investigative Ophthalmology & Visual Science, March 1996, Vol. 37, No. 4 Copyright © Association for Research in Vision and Ophthalmology

597

598

Investigative Ophthalmology & Visual Science, March 1996, Vol. 37, No. 4

gering one or both of the two known surface receptors designated as tumor necrosis factor receptor (TNFR) type I (p55) and TNFR type II (p75). These two receptors have been identified in most human cells and tissues examined, and the pleiotropic nature of TNF action may be explained partially by the structural and functional heterogeneity in its receptors.7 Trefzer et al8 have reported that primary cultured human keratinocytes express only the p55 receptor and that TNF activation of these cells is specifically through this receptor. Recent studies in our laboratory9 have demonstrated the presence of p55 mRNA and protein, in the absence of p75, in primary cultured HRPE cells and SV40-immortalized HRPE (SVRPE) cells. Taken together, these studies suggest that particular human epithelial cells may respond to TNF specifically through the p55 receptor. It remains unclear, however, how a wide variety of cellular responses can be regulated so closely by only one receptor type. Intercellular adhesion molecule-1 (also called CD54) is a specialized cell surface glycoprotein that binds integrins present on leukocytes, specifically lymphocyte function-related antigen (LFA)-l (also called CDlla/CD18) and Mac-1.10" First identified on vascular endothelium, ICAM-1 appears to regulate the binding of leukocytes to endothelium and may influence extxavascular leukocyte trafficking by virtue of its presence on various cell types, particularly keratinocytes and fibroblasts.81213 Recently, ICAM-1 was identified on both freshly isolated and cultured human RPE cells and was upregulated by treatment with the proinflammatory cytokines interleukin (IL)-l, interferon (IFN)-gamma, and TNF.2 These studies suggest that RPE cells, an integral part of the blood-retinal barrier, may influence and regulate leukocyte movement into the diseased retina on activation by cytokines released during inflammation or other immunologic processes. Indeed, leukocytic infiltration of the choroid, retina, and vitreous is a hallmark of inflammatory ocular disease, but leukocytic infiltration also may be an important feature of other common, chronic retinal diseases such as proliferative vitreoretinopathy and age-related macular degeneration.l4~lfc> In addition, our laboratory9 has compared the ICAM-1 expression on primary cultured HRPE cells and SVRPE cells after treatment with TNF and found that SVRPE cells provide a consistent and reproducible model with which to study TNF responsiveness of HRPE cells. The resting RPE monolayer in vivo, or even freshly isolated RPE cells, may have a different phenotype when compared to cultured cells, especially transformed cells such as the SVRPE cell line. Cellular alterations may occur because of tissue disruption, cellular dispersion, and the in vitro culture environment. Indeed, normal retinal sections stained for ICAM-1 have

shown no immunoreactivity on HRPE cells.2 Our laboratory and others have demonstrated2'9 that primary cultured HRPE cells basally express ICAM-1. We have also shown9 that SVRPE cells express little or no ICAM1 constitutively. These types of discrepancies demonstrate the inherent differences between cell lines, primary cultured cells and cells in vivo. Yet, when considering the quiescent RPE monolayer in vivo, SVRPE cells with lower basal activation in culture conditions may represent an appropriate system in which to study HRPE-specific cytokine responsiveness and intracellular signal transduction pathways. Protein kinase C (PKC) is one of the major intracellular signaling pathways that activated TNF receptors use to relay signals from the cell membrane to the nucleus. It comprises a family of functionally related isozymes with serine-threonine-specific kinase activity.17"19 Activated PKC enzymes may function by phosphorylating and activating other kinases and certain nuclear transactivating factors, leading to changes in cellular gene expression and to cellular responses.20"22 This signal transducing enzyme may play a critical role in mediating at least some of the cellular responses triggered by TNF. Indeed, TNFmediated PKC activation has been documented in a number of cell lines, and the specific activation pathway may involve the release of 1 '2'diacylglycerol from the plasma membrane phosphatidylcholine pool by phospholipase C (PLC).23"25 Another important signaling pathway for TNF may involve protein kinase A (PKA) activation through an increase in adenylate cyclase activity and subsequent cyclic adenosine monophosphate (cAMP) formation. Although PKA activity per se has not been linked directly to TNF signaling, several studies have shown adenylate cyclase activation with accumulation of increased cAMP and PKC-independent phosphorylation activity in response to TNF.26 Protein kinase A has been suggested to be involved in the activation of nuclear factors.22 Although much has been learned from these efforts to elucidate TNF signaling mechanisms, intracellular TNF signaling strategies appear to be specifically dependent on cell type. We initiated this study to understand better the role of TNF in HRPE cell activation by investigating potential p55-mediated signaling mechanisms and by examining how these pathways influence ICAM-1 expression in SVRPE cells. In this article, we demonstrate TNF-mediated activation of both PKC and cAMP-dependent PKA pathways in cultured SVRPE cells. Moreover, TNF increased surface ICAM-1 expression in SVRPE cells through a process dependent on the PKC signaling pathway.

TNF-Induced ICAM-1 Is Dependent on PKC in RPE Cells

MATERIALS AND METHODS Retinal Pigment Epithelial Cell Culture A human SV40-immortalized RPE cell line was obtained from the Coriell Institute for Medical Research (Camden, NJ), repository number AG06096. This cell line was originally generated by immortalization of a human fetal RPE-derived cell line with SV40 virus, strain RH911. These SVRPE cells, which are T-antigen positive, appear stable without crisis phase, do not pigment, and possess epithelioid morphology.27 They were grown in Dulbecco's modified Eagle's medium (DMEM; Fisher Scientific, Pittsburgh, PA) supplemented with 10% fetal bovine serum (FBS; Gibco BRL, Gaithersburg, MD), 2 mM glutamine (JRH, Lenexa, KS), 100 U/ml penicillin and 100 /zg/ml streptomycin (Gibco BRL). Purity of the SVRPE cells was evaluated by immunocytochemical staining of cytocentrifuge preparations using antibodies against cytokeratin (Dako, Carpinteria, CA).28 Absence of contaminating cells was confirmed using antibodies against endothelial cell antigen factor VIII (Dako) and macrophage antigen (CDllc) (Becton-Dickinson, San Jose, CA). Immunocytochemistry was performed by the immunoperoxidase method using the ABC Elite kit (Vector, Burlingame, CA) and aminoethylcarbizole as the chromogen as previously described.29 Negative controls included omission of primary antibody and use of an irrelevant primary antibody of the same isotype. After testing, SVRPE cultures were found to be free of Mycoplasma contamination (Mycoplasma detection kit; Boehringer-Mannheim, Indianapolis, IN). Throughout all experiments, cell number and viability were quantitated using a hemocytometer and the trypan blue (Gibco BRL) exclusion test. Retinal Pigment Epithelial Cell Treatment In all experiments, TNF stimulation of RPE cells was performed using recombinant human TNF (Boehringer-Mannheim) at a concentration of 10 ng/ml. This concentration has been shown to induce maximal proliferation of RPE cells.' Independent observations in our laboratory9 have confirmed the efficacy of this dose in inducing RPE proliferation (results not shown) and ICAM-1 induction. Because higher doses of TNF may result in significant cell death through apoptosis (results not shown), we chose not to complicate our experiments by using these higher TNF concentrations. For the PKC assay, SVRPE cells were grown to confluence in 75-cm2 cultureflaskscontaining DMEM supplemented with 10% FBS. Some experiments were performed with cells grown in DMEM containing 0.1% FBS to assess the role of serum factors in PKC activation. In the latter, SVRPE cells were allowed to

599

acclimate overnight in low-serum DMEM before incubation with TNF. Before PKC assay, SVRPE cells were incubated with recombinant human TNF (10 ng/ml) for 1, 2, 5, 15, 60, 120, and 360 minutes. To demonstrate the efficacy of standard kinase activators and inhibitors in SVRPE cells, control PKC assays were performed after treating the SVRPE cells with the PKC activator phorbol 12-myristate 13-acetate (PMA, 50 nM; Sigma, St. Louis, MO) for 3 hours or after treating these cells with the PKC inhibitor calphostin C (50 nM; Calbiochem-Novabiochem, Lajolla, CA) photoactivated in light for 30 minutes before a 1-minute TNF treatment. Untreated cultures were incubated and processed in parallel to other treated groups. For the cAMP and PKA assays, SVRPE cells were grown to confluence in 60- mm2 petri dishes containing DMEM supplemented with 10% FBS. Before treatment, cells were conditioned for 3 hours with DMEM containing 0.1% FBS. For the cAMP assay, SVRPE cultures were incubated with recombinant human TNF (10 ng/ml) for 5, 10, 30, 60, and 120 minutes. Positive control cAMP cultures were treated with nicotine (1 fxM, Sigma) for 10 minutes to demonstrate assay efficacy. For the PKA assay, SVRPE cultures were incubated with TNF (10 ng/ml) for 1, 3, 5, 15, 30, 60, and 180 minutes. Protein kinase A activator and inhibitor experiments were used as controls to demonstrate the efficacy of standard reagents on cultured SVRPE cells. These control cultures were treated with the PKA activators forskolin (20 fiM; Sigma) plus 3isobutyl-1-methyl-xanthine (IBMX, 50 fjM; Sigma) for 10 minutes or were preincubated with the PKA inhibitor N-(2-[methylamino] ethyl)-5-isoquinolinesulfonamide (H8, 75 //M; Sigma) for 30 minutes before a 30-minute TNF treatment. For both cAMP and PKA assays, untreated samples were incubated in DMEM with 0.1% FBS in parallel to the treatment groups. Forflowcytometry experiments, SVRPE cells were grown to 60% to 70% confluence in T-25flasksbefore treatment. After experiments were initiated and reagents were applied, cultures were allowed to incubate for 18 hours before flow cytometry analysis of ICAM1 expression was completed. After initial treatment, reagents were removed (unless otherwise specified), and cultures were incubated in fresh control medium (DMEM with 10% FBS) for the remainder of the 18 hours. To establish a TNF-activation baseline in SVRPE cells, cultures were treated with TNF (10 ng/ ml) for 10 minutes or 1 hour before flow cytometry analysis; SVRPE cultures also were treated with PKC or PKA activators without TNF. Protein kinase C activation was achieved by treating cells with 100 nM PMA for 3 hours. Protein kinase A was activated with 20 fjM Forskolin plus 50 /iM'lBMX or with 100 fiM N(i, 2'-O-dibutyryladenosine 3\5'-dibutyryl-cAMP (Sigma)

600

Investigative Ophthalmology 8c Visual Science, March 1996, Vol. 37, No. 4

alone. The PKA agonists remained in the culture medium throughout the 18-hour incubation. For the kinase inhibition experiments, SVRPE cultures were pretreated with inhibitors for 1 hour, followed by TNF treatment for 1 hour. The PKC inhibition was accomplished by preincubating cells with 200 nM calphostin C photoactivated in light. As a control, SVRPE cells were incubated with calphostin C 3 hours after TNF treatment. For PKA inhibition, cells were preincubated with 75 fjM H8. Untreated cells were incubated with DMEM with 10% FBS in parallel to those treated widi kinase activators or to those treated with TNF in the presence or absence of kinase inhibitors. Protein Kinase C Assay Protein kinase activity was determined as previously described.30 Briefly, SVRPE cells were homogenized on ice in 1.6 ml of buffer A (20 mM Tris-HCL, pH 7.5; 1 mM ethylenediaminetetraacetic acid (EDTA); 0.1 mM dithiothreitol (DTT); 20 /zM leupeptin; 0.15 /iM pepstatin A; and 0.5 mM phenylmethylsulfonyl fluoride. Detergent-solubilized fractions were prepared as previously described17 and applied (1 ml) to equilibrated 2 diethylaminoethanol-cellulose columns packed (0.2 ml) into 1-ml tuberculin syringes and equilibrated with buffer B (20 mM Tris-HCL, pH 7.5; 1 mM EDTA; 0.1 mM DTT; 20 /uM leupeptin; and 0.15 fiM pepstatin A). After washing the columns with 0.8 ml of buffer B, the bound PKC was eluted with 0.5 ml of 0.1 M NaCl in buffer B. The first 0.1 ml of the elution was discarded, and the remaining 0.4 ml was collected. Protein kinase C activity present in the eluted fractions was determined by a multi-well filtration approach31 using histone HI as the substrate. Only the stimulated activity in the presence of Ca2+ and lipids was expressed as PKC activity in units, where 1 U of enzyme transfers 1 nmol of phosphate to histone HI per minute at 30°C. Experiments were conducted in triplicate, and samples were standardized to protein and expressed as units of PKC activity per milligram protein. Cyclic Adenosine Monophosphate Assay The cAMP-radioimmunoassay kit was purchased from Dupont NEN Research Products (Boston, MA), and the manufacturer's protocol was followed. After SVRPE cell treatment, culture medium was removed and 1 ml of 6% (wt/vol) trichloroacetic acid (TCA) was added. SVRPE cells were homogenized immediately on ice (10-15 strokes) with subsequent incubated at 4°C for 30 minutes. The whole cell lysates were centrifuged at 1500 g for 15 minutes at 4°C. The supernatant was collected, and trichloroacetic acid was removed by three washes with water-saturated ether. Samples were lyophilized and redissolved in 150 /il of

cold cAMP assay buffer. A portion (100 //I) of the rehydrated samples was acetylated, and the cAMP content was measured by radioimmunoassay. Experiments were performed in triplicate, and data were expressed as femtomole of cAMP per 1000 cells. Protein Kinase A Assay The PKA assay kit was purchased from Gibco-BRL, and the manufacturer's protocol was followed. The SVRPE cells were homogenized on ice (10 to 15 strokes), and whole cell lysates were measured for PKA activity. The PKA activity was determined as the difference between phosphorylation of a PKA-specific substrate (kemptide) without and with a PKA-specific inhibitor (provided by Gibco-BRL). The reaction mixture contained enzyme, [32P]-ATP, substrate, and either inhibitor or extra buffer. Sample mixtures were held at 37°C for 15 minutes and spotted onto phosphocellulose paper. Unbound [32P]-ATP was washed away, and adherent radioactive substrate was measured in a /^-counter (Beckman). Experiments were performed in triplicate, and all samples were standardized to protein (Bio-Rad Protein Assay; Bio-Rad, Hercules, CA) and expressed as picomole phosphate transferred per minute per microgram protein. Flow Cytometry Analysis All SVRPE cell samples were stained for ICAM-1 expression using a monoclonal antibody purchased from AMAC (Westbrook, MA) at 1:500 in phosphate-buffered saline (PBS). The isotypic (IgGl) control staining was performed on TNF-treated cells with a monoclonal anti-CD4 antibody from Becton-Dickinson at 1:10. A secondary phycoerythrin-conjugated goat antimouse antibody from Sigma was used at 1:40 for all ICAM-1 and CD4 staining. Briefly, adherent cells were washed and detached with a 0.02 M EDTA-PBS solution. Cells were washed widi DMEM, centrifuged, and transferred to microfuge tubes (1.5 ml). Normal goat serum (Vector), diluted to 5% in PBS, was used as a blocking reagent for 15 minutes before a 30-minute incubation with the primary antibody (100 (A). Cells were washed with PBS twice before the application of secondary antibody (50 //I) for 30 minutes. Again, cells were washed twice with PBS and then fixed with a 1 % paraformaldehyde solution in PBS. All reagents and cells were kept on ice during the staining procedure. Fixed cells were stored at 4°C until flow cytometry was performed, no less than 1 hour and no longer than 24 hours. Flow cytometry analysis and mean channel fluorescence was determined with a Becton-Dickinson Facscan using Consort 30 software. Mean channel fluorescence of the isotypic control sample was subtracted from the values obtained for unstimulated, TNF-stimulated, kinase-inhibited, and kinase-activated


601

RPE cells before quantitative comparisons were made between the mean channel fluorescence of each group.

300n

Statistical Analysis Protein kinase C assay samples were processed in duplicate, and error was expressed as the coefficient of variation. Data from triplicate PKC or PKA experiments were compared using the paired Student's ttest; significance was determined for P values < 0.01. RESULTS Time-Dependent Protein Kinase C Activation After Tumor Necrosis Factor Treatment of SVRPE Cells Treated with TNF in the presence or absence of PKC inhibition, SVRPE cells were assayed for PKC activity and were represented as percent of control in Figure 1. After TNF treatment, PKC activity in the detergentsoluble membrane fraction showed a 2.6-fold increase at 1 minute (P< 0.01) that recovered toward baseline level at 60 minutes (Fig. 1). Basal activity was maintained out to 6 hours after recovery to baseline (not shown). Cells incubated with TNF in low-serum medium (0.1% FBS) showed a similar pattern of PKC activation; however, the TNF-induced PKC activation observed at 1 minute was maximal in control medium containing 10% FBS. Preincubation of the SVRPE cells with the specific PKC inhibitor calphostin C (50 nM) prevented the TNF-mediated PKC activation observed at 1 minute and decreased PKC activity to below control level (Fig. 1). As a positive control, SVRPE cells were treated with the PKC activator PMA (100 nM) for 3 hours, which significantly increased PKC activity 4.4-fold (not shown). Untreated SVRPE cells, depicted at time zero (Fig. 1), represented 100% of PKC activity and maintained consistent levels throughout the duration of the experiments. Significance was determined from the average of the three experiments. Time-Dependent Cyclic Adenosine Monophosphate Accumulation After Tumor Necrosis Factor Treatment of SVRPE Cells Accumulation of cAMP was enhanced in SVRPE cells treated with TNF as demonstrated in Figure 2. There was no evidence of cAMP accumulation 5 minutes after TNF treatment (Fig. 2); however, cAMP levels began to rise at 10 minutes (Fig. 2). Intracellular cAMP levels showed a 1.6-fold increase (P < 0.01) from control 30 minutes after TNF treatment (Fig. 2). Levels of cAMP returned to baseline by 60 minutes (Fig. 2) and maintained this level out to 2 hours (not shown). The positive control SVRPE cultures were incubated with forskolin (20

|

200-

^^M

^^M

_[_

illIlL 0

1

2

5

15

60

1

Time (min)

FIGURE 1. A representative protein kinase C (PKC) experiment after tumor necrosis factor-alpha (TNF) treatment of simian virus-40 immortalized human retinal pigment epithelial cells (SVRPE) cells. Detergent-soluble membrane extracts were isolated from SVRPE cells after TNF treatment at 1, 2, 5, 15, and 60 minutes. Tumor necrosis factor-activated SVRPE cells showed a rapid (1 minute) activation of PKC with subsequent reversion (60 minutes). Protein kinase C inhibition with calphostin C (50 nM) prevented TNF activation of PKC at 1 minute and reduced PKC activity below baseline (Fig. 1, Calph-C). Untreated cultures, represented at time zero by 100% PKC activity, showed consistent basal levels of PKC activity. Membrane PKC activity was quantitated, expressed in units of activity per mg protein, compared to the control level, and expressed as percent of control. Experiments were performed in triplicate, and each showed a similar pattern of PKC activity after treatment with TNF. Within each experiment, samples were processed twice, which allowed for averaging of PKC activity at each time point. Error bars represent the coefficient of variation for each sample. *Significant difference from control after averaging three experiments; P =s 0.01. for 10 minutes, which increased cAMP levels approximately 5.4-fold (not shown). Represented at time zero (Fig. 2), cAMP values of untreated cultures were consistent throughout all experiments. Significance was determined from the average of the three experiments.

Time-Dependent Protein Kinase A Activation After Tumor Necrosis Factor Treatment of SVRPE Cells Protein kinase A was activated in SVRPE cells treated with TNF as demonstrated in Figure 3. It showed a 1.6-fold increase (P < 0.01) after a 30-minute incubation with TNF. Protein kinase A activity returned to

602


treated cells, represented at time zero (Fig. 3), maintained consistent PKA activity throughout all experiments. Significance was determined from the average of the three experiments.

20

ICAM-1 Expression and Regulation in SVRPE Cells After TNF treatment, SVRPE cells showed increased surface expression of the ICAM-1 antigen as represented by flow cytometry analysis in Figure 4. Untreated SVRPE cells exhibited negligible ICAM-1 ex-

o o o 10

4 0 -i

en

t |

"o

E

>-> 10

30

60

Time FIGURE 2. A representative cyclic adenosine monophosphate (cAMP) assay on simian virus-40 immortalized human retinal pigment epithelial cells (SVRPE) cells after tumor necrosis factor-alpha (TNF) treatment. Crude lysates were isolated from untreated SVRPE cells and from those treated with TNF at 5, 10, 30, and 60 minutes. Tumor necrosis factoractivated SVRPE cells showed a transient and significant accumulation of cAMP at 30 minutes. Untreated cells, depicted at the zero time point, maintained consistent cAMP levels throughout. Levels of cAMP were quantitated, standardized to cell number, and expressed in femtomoles of cAMP per 1000 cells (fmol/1000 cells). Experiments were performed in triplicate, and each displayed a similar time course of cAMP accumulation. Individual samples were processed three times to determine average cAMP levels and significance within each experiment. *Significant difference from control after averaging three experiments; P =s 0.01. baseline levels at 60 minutes (Fig. 3) and maintained this profile out to 3 hours (not shown). Pretreatment with the specific PKA inhibitor H8 (75 /xM) prevented TNF from evoking a PKA response at the 30-minute time point (Fig. 3). As a positive control, SVRPE cells were incubated for 10 minutes with the PKA activators forskolin (20 //M) plus IBMX (50 /xM), which demonstrated a significant 5.2-fold increase (P < 0.001) in PKA activity to 119.89 ± 5.52 pmol///g • minute. Un-

15

30

60

30

Time (min) FIGURE 3.

A representative protein kinase A (PKA) assay on simian virus-40 immortalized human retinal pigment epithelial cells (SVRPE) cells after tumor necrosis factor-alpha (TNF) treatment. Crude lysates were isolated from SVRPE cells untreated and treated with TNF at 15, 30, and 60 minutes. Tumor necrosis factor-activated SVRPE cells showed a transient and significant increase in PKA activity at 30 minutes. Protein kinase A inhibition with H8 (75 /JM) prevented the TNF-mediated activation of PKA at 30 minutes (H8). Untreated cells, depicted at the zero time point, maintained consistent PKA activity. Protein kinase A activity was quantitated, standardized to protein, and expressed in picomoles of phosphate transferred per minute and per microgram of protein. Forskolin plus 3-isobutyl-l-methyl-xanthine treatment represented a positive control and increased PKA activity to 119.89 ± 5.52 pmo\/fj,g • minute. Experiments were performed in triplicate, and each displayed a similar pattern of PKA activation. Individual samples were processed three times to determine average PKA activity and significance within each experiment. *Significant difference from control after averaging three experiments; P =s 0.01.

603


= Untreated = Forskolin/IBMX

10°

102

10 3

10°

10 2

103

4. A flow cytometry representation of intercellular adhesion molecule-1 (ICAM-1) expression and regulation on simian virus-40 immortalized human retinal pigment epithelial cells (SVRPE) cells. Cultured SVRPE cells constitutively express little or no surface ICAM-1 compared to isotypic control stain against the CD4 antigen (A). In SVRPE cells, ICAM-1 expression was increased dramatically from a low basal level after tumor necrosis factor-alpha (TNF) treatment for 10 minutes (B). Preincubation with the protein kinase C (PKC) inhibitor calphostin C (200 nM) before a 1-hour TNF treatment partially inhibited the TNF-mediated ICAM-1 upregulation (C). In contrast, pretreatment with the PKA inhibitor H8 (75 /xM) did not suppress ICAM1 induced by a 1-hour TNF treatment (D). Protein kinase C activation with an incubation with phorbol 12-myristate 13-acetate (100 nM) for 3 hours was sufficient to induce ICAM-1 expression on SVRPE cells (E). However, PKA activation with forskolin (20 //M) plus 3-isobutyl-l-methyl-xanthine (50 ^M) did not induce ICAM-1 expression (F). Increased binding of anti-ICAM-1 antibodies on TNF-activated cells is depicted by a rightward shift in mean fluorescent intensity from the mean observed using unstimulated control cells. FIGURE

pression compared to die isotypic control (Fig. 4A). Tumor necrosis factor treatment of these cells for 1 hour, followed by an 18-hour incubation with control medium (DMEM with 10% FBS), resulted in increased (13.4-fold) ICAM-1 expression as depicted by a rightward shift in mean channel fluorescence intensity compared to that of control (Fig. 4B). Furthermore, pulse treatment of die cells widi TNF for 10 minutes increased ICAM-1 expression 8.6-fold at 18 hours (not

shown). Tumor necrosis factor-induced ICAM-1 expression after a 1-hour TNF treatment was suppressed dramatically by pretreatment widi die PKC inhibitor calphostin C (200 nM) for 1 hour (Fig. 4C). Specifically, calphostin C pretreatment allowed TNF to induce ICAM-1 expression to only 25% of that observed with TNF alone for 1 hour (Fig. 4C). As a control to confirm the specificity of calphostin C, SVRPE cells were treated with calphostin C 3 hours after the 1hour TNF treatment. ICAM-1 expression was dien determined at 18 hours. Calphostin C incubation after TNF treatment did not suppress die ICAM-1 expression observed widi calphostin C pretreatment (not shown). Protein kinase A inhibition with H8 (75 fjM) pretreatment did not prevent the TNF-induced increase in ICAM-1 expression (Fig. 4D). SVRPE cells treated with the PKC activator PMA (100 nM) alone for 3 hours, followed by a 15-hour incubation widi control medium (total incubation of 18 hours), increased ICAM-1 expression 6.8-fold beyond control levels (Fig. 4E). However, as shown in Figure 4F, treatment of SVRPE cells widi the PKA activators forskolin (20 (JLM) plus IBMX (50 /JM) did not enhance ICAM1 expression in the absence of TNF. Moreover, PKA activation widi dibutyryl cAMP (100 fiM) did not induce ICAM-1 expression (not shown). DISCUSSION Experiments have been performed in many cell systems to ascertain the functional significance of each TNFR, and they have suggested that particular functions are associated with p55 or p75.32"34 However, little is known about the intracellular mechanisms that TNF triggers to obtain cellular activation. Recendy, we reported that primary cultured HRPE cells and SVRPE cells express only the p55 TNFR and, therefore, represent an ideal cell model to elucidate signaling mechanisms activated through diis TNFR.9 In the current study, we demonstrate die TNF-activated serine-threonine kinase signaling mechanisms of SVRPE cells. Specifically, the PKC and PKA pathways are activated sequentially and transiendy by TNF, suggesting that PKC is the initial signaling event. Furdiermore, based on die current data, it is likely that PKA activation is not an early, immediate signaling pathway but a downstream event. Our experiments widi kinase activators and inhibitors suggest diat PKC is necessary for TNF-activated ICAM-1 expression and that PKC activation without TNF is sufficient to induce ICAM1 expression in SVRPE cells. In contrast, PKA activators are unable to induce ICAM-1 expression in these cells. Although TNF has been shown to activate a number of second and third messenger padiways in a vari-

604


ety of human cell types,35'36 compilation of these findings into one model is difficult because signaling mechanisms are highly dependent on cell type. There has been no clear documentation of one TNFR being involved with only one specific signaling pathway in multiple cell types. Wiegmann et al37 reported transfection of human TNFR p55 into murine 70Z/3 pre15 cells and found that signaling occurred via PKC, ceramide, and PLA2. Protein kinase A activity was not assessed. Activated cytoplasmic second messengers such as these may regulate nuclear transcription indirecdy through NK-KB or other as yet undefined transcription factors.22'23 The regulatory domains of human ICAM-1 contain binding sites for NF-KB and other nuclear factors.38'39 Thus, TNF-activated signaling mechanisms that ultimately induce these nuclear factors could potentially influence ICAM-1 expression.40 Activation of PKC and PKA with specific reagents has demonstrated the importance of these pathways for NF-KB or ICAM-1 induction in cell types other than RPE cells.36'4M2 However, activation of signaling pathways known to be associated with the induction of NF-KB and ICAM-1 occasionally have proven ineffective at upregulating ICAM-1.43 It remains unclear why particular second messengers are activated at any time by a single TNF molecule. Perhaps it is related to cell type, cellular environment, intracellular activation, or other cytokine interactions. For our control experiments, the efficacy of the kinase inhibitors in the SVRPE cell model was accessed because inhibitor concentrations that give 50% effective inhibition (IC50) are determined from purified reagents and often do not account for physiological processes in specific cell types, such as reagent uptake.44'45 The PKC inhibitor, calphostin C, was used at concentrations far below that needed to inhibit other kinases, particularly PKA.45 H8 also was titrated to allow for specific inhibition of PKA with no effect on the PKC system.44 Each inhibitor was titrated in the culture system specifically to block the corresponding kinase as assessed by the kinase activity assays. Tumor necrosis factor-mediated activation of the PKA pathway in SVRPE cells suggests a role for this kinase system in cellular responsiveness to TNF, but our results show that PKA is not an immediate activation pathway and is not responsible for TNF-mediated ICAM-1 induction. These data are supported by the similar time course for the accumulation of cAMP after TNF stimulation. Perhaps PKA activation in RPE cells is necessary for other cellular responses triggered through p55 by TNF. Sequential activation of PKC and PKA after TNF treatment of SVRPE cells suggests a cross-talk or interaction between the two kinase pathways. Direcdy or indirectly, PKC may enhance PKA activity through as yet undefined mechanisms.46 Fur-

ther studies of this temporal relationship may explain biochemical mechanisms linking the two kinase pathways in RPE cells. The PKC pathway is activated rapidly by TNF in SVRPE cells; consequently, a short 10-minute exposure to TNF is sufficient to initiate the signals necessary for ICAM-1 induction. These observations, together with the fact that PKC activation is necessary for ICAM-1 expression, suggest that brief TNF exposure of the RPE layer in vivo could create a cascade of activation leading to increased adhesion molecule expression and the potential of leukocytic infiltration. However, it remains unknown whether PKC or PKA activation in SVRPE cells becomes refractory to intermittent or repetitive TNF stimulation. Detailed studies of these mechanisms may offer a better understanding of existing pharmacologic agents47 or may lead to novel strategies for attenuating RPE ICAM-1 expression in the clinical setting. Strategically positioned at the blood-retinal barrier, the RPE may participate actively in pathologic processes and influence the interactions between circulating leukocytes and the retina.48 Cytokine-activated RPE cells are capable of attracting leukocytes and expressing HLA-DR antigens, suggesting their role as antigen-presenting cells.49"52 Elner et al2 demonstrated that cytokine-activated human RPE cells preferentially expressed ICAM-1 on their apical surfaces. Our study demonstrates the importance of the rapidly activated PKC pathway in TNF-mediated ICAM-1 induction in HRPE cells. Together, these studies suggest a mechanism by which RPE cells could aid in the migration of leukocytes from the choriocapillaris into the inflamed retina. Ocular diseases known to involve TNF53 may be aggravated by TNF-induced ICAM-1 on RPE cells and an amplified leukocyte infiltration. Therefore, it will be important to consider therapeutic strategies to prevent TNF production by infiltrating, pro-inflammatory cells54 or perhaps strategies to protect the retina from soluble TNF that may penetrate a faulty blood-retinal barrier. Key Words CD54, intercellular adhesion molecule-1, protein kinase A, protein kinase C, tumor necrosis factor-a, tumor necrosis factor-a receptor Acknowledgments The authors thank Christine Spee at the Doheny Eye Institute for her administrative and technical support, Karen Louise for her support and patience, the staff of the University of Southern California School of Medicine Flow Cytometry Laboratory (Dr. John Parker, Director), and the graphic design staff of the University of Southern California Biomedical Communication Office.


References 1. Burke JM. Stimulation of DNA synthesis in human and bovine RPE by peptide growth factors: The response to TNF-alpha and EGF is dependent upon cul16. ture density. Curr Eye Res. 1989;8:1279-1286. 2. Elner SG, Elner VM, Pavilack MA, et al. Modulation and function of intercellular adhesion molecule-1 17. (CD54) on human retinal pigment epithelial cells. Lab Invest. 1992; 66:200-211. 3. Elner VM, Scales W, Elner SG, Danforth D, Kunkel 18. SL, Strieter RM. Interleukin-6 (IL-6) gene expression and secretion by cytokine-stimulated human retinal 19. pigment epithelial cells. ExpEyeRes. 1992; 54:361-368. 4. Jaffe GJ, Le LV, Valea F, et al. Expression of interleukin-1 a, interleukin-1 /?, and an interleukin-1 receptor 20. antagonist in human retinal pigment epithelial cells. ExpEyeRes. 1992; 55:325-335. 5. Tanihara H, Yoshida M, Yoshimura N. Tumor necrosis factor-alpha gene is expressed in stimulated retinal 21. pigment epithelium cells in culture. Biochem Biophys Res Comm. 1992; 187:1029-1034. 6. Planck SR, Huang XN, Robertson JE. Rosenbaum JT. Retinal pigment epithelial cells produce interleukin22. 1/3 and granulocyte-macrophage colony-stimulating factor in response to interleukin-la. Curr Eye Res. 1993; 12:205-212. 23. 7. Tartaglia LA, Goeddel DV. Two TNF receptors. Immunol Today. 1992; 13:151-153. 8. Trefzer U, Brockous M, Loetscher H, et al. 55-kd tumor necrosis factor receptor is expressed by human 24. keratinocytes and plays a pivotal role in regulation of human keratinocyte ICAM-1 expression. J Invest Dermatol. 1991;97:911-916. 9. Sippy BD, Hofman FM, He S, et al. SV40-immortalized 25. and primary cultured human retinal pigment epithelial cells share similar patterns of cytokine-receptor expression and cytokine responsiveness. Curr Eye Res. 1995; 14:495-503. 26. 10. Dustin ML, Springer TA. Lymphocyte associated function antigen-1 (LFA-1) interaction with intercellular adhesion molecule-1 (ICAM-1) is one of at least three mechanisms for lymphocyte adhesion to cultured endothelial cells. / Cell Biol. 1988; 107:321-331. 27. 11. Springer TA. Adhesion receptors of the immune system. Nature. 1990; 346:425-434. 28. 12. Dustin ML, Singer KH, Tuck DT, Springer TA. Adhesion of T lymphoblasts to epidermal keratinocytes is regulated by interferon-7 and is mediated by intercel29. lular adhesion molecule 1 (ICAM-1). / Exp Med. 1988; 167:1323-1340. 13. Kawai M, Nishikomori R, Jung EY, et al. Pyrrolidine 30. dithiocarbamate inhibits intercellular adhesion molecule-1 biosynthesis induced by cytokines in human fibroblasts. J Immunol. 1995; 154:2333-2341. 14. Penfold PL, Killingsworth MC, Sarks SH. Senile macu31. lar degeneration: The involvement of immunocompetent cells. Graefes Arch Clin Ophthalmol. 1985; 223:6976. 15. Weller M, Heimann K, Wiedemann P. Immunochemi-

605

cal studies of epiretinal membranes using APAAP complexes: Evidence for macrophage involvement in traumatic proliferative retinopathy. Int Ophthalmol. 1988;11:181-186. Jerdan JA, Pepose JS, Michels RG, et al. Proliferative vitreoretinopathy membranes: An immunohistochemical study. Ophthalmology. 1989;96:801-810. Nishizuka Y. The molecular heterogeneity of protein kinase C and its implications for cellular regulation. Nature. 1988; 334:661-665. Stabel S, Parker PJ. Protein kinase C. Pharmacol Ther. 1991;51:7l-95. Nishizuka Y. Intracellular signaling by hydrolysis of phospholipids and activation of protein kinase C. Science. 1992;258:607-614. Brenner DA, O'Hara M, Angel P, Chojkier M, Karin M. Prolonged activation of jun and collegenase genes by tumour necrosis factor-alpha. Nature. 1989; 337: 661-663. Shirakawa F, Mizel SB. In vitro activation and nuclear translocation of NF-kappa B catalyzed by cyclic AMPdependent protein kinase and protein kinase C. Mol Cell Biol. 1989;9:2424-2430. Hohmann HP, Remy R, Kolbeck R, et al. Involvement of protein kinases A and C in NF-kappa B activation by TNF and IL-1. / Cell Biochem. 1990; 14B:288-293. Schiitze S, Nottrott S, Pfizenmaier K, Kronke M. Tumor necrosis factor signal transduction: Cell-type-specific activation and translocation of protein kinase C. J Immunol. 1990; 144:2604-2608. Matsubara N, Fuchimoto S, Orita K. Tumor necrosis factor-alpha induces translocation of protein kinase C in tumor necrosis factor-sensitive cell lines. Immunology. 1991; 73:457-459. Schiitze S, Berkovic D, Tomsing O, Unger C, Kronke M. Tumor necrosis factor induces rapid production of 1'2'diacylglycerol by a phosphatidylcholine-specific phospholipase C. /Exp Med. 1991; 174:975-988. Zhang Y, Lin JX, Yip YK, Vilcek J. Enhancement of cAMP levels and of protein kinase activity by tumor necrosis factor and interleukin 1 in human fibroblasts: Role in the induction of interleukin 6. Proc Natl Acad Sci USA. 1988;85:6802-6805. Aronson JF. Human retinal pigment cell culture. In Vitro. 1983;8:642-650. Hunt RC, Davis AA. Altered expression of keratin and vimentin in human retinal pigment epithelial cells in vivo and in vitro. J Cell Physiol. 1990; 145:187-199. Schneider J, Hofman FM, Apuzzo MLJ, Hinton DR. Cytokines and immunoregulatory molecules in malignant glial neoplasms. J Neurosurg. 1992; 77:265-273. Gopalakrishna R, Chen ZH, Gundimeda U. Nonphorbol tumor promoters okadaic acid and calyculin-A induce membrane translocation of protein kinase C. Biochem Biophys Res Comm. 1992; 189:950-957. Gopalakrishna R, Chen ZH, Gundimeda U, Wilson JC, Anderson WB. Rapid filtration assays for protein kinase C activity and phorbol ester binding using multiwell plates with fitted filtration discs. Anal Biochem. 1992;206:24-35.

606

Investigative Ophthalmology 8c Visual Science, March 1996, Vol. 37, No. 4

32. Thoma B, Grell M, Pfizenmaier K, et al. Identification of a 60-kD tumor necrosis factor (TNF) receptor as the major signal transducing component in the TNF response. / Exp Med. 1990; 172:1019-1023. 33. Tartaglia LA, Pennica D, Goeddel DV. Ligand passing: The 75 -kDa tumor necrosis factor (TNF) receptor recruits TNF for signalling by the 55-kDa TNF receptor. JBiol Chem. 1993; 268:18542-18548. 34. Laegreid A, Medvedev A, Nonstad U, et al. Tumor necrosis factor receptor p75 mediates cell-specific activation of nuclear factor kB and induction of human cytomegalovirus enhancer. J Biol Chem. 1994; 269: 7785-7791. 35. Schiitze S, Machleidt T, Kronke M. Mechanisms of tumor necrosis factor action. Semin Oncol. 1992; 19:1624. 36. Hansen AB, Andersen CB. Comparison of the effects of tumor necrosis factor a stimulation and phorbol ester treatment on the immunocytochemical staining of intercellular adhesion molecule 1 in human renal carcinoma cell cultures. Virchows Archiv B Cell Pathol.

1993;63:107-113. 37. Wiegmann K, Schiitze S, Kampen E, Himmler A, Machleidt T, Kronke M. Human 55-kDa receptor for tumor necrosis factor coupled to signal transduction cascades. JBiol Chem. 1992;267:l7997-18001. 38. Stade BG, Messer G, Riethmuller G,JohnsonJP. Structural characteristics of the 5' region of the human ICAM-1 gene. Immunobiobgy. 1990; 182:79-87. 39. Voraberger G, Schaffer R, Stratowa C. Cloning of the human gene for intercellular adhesion molecule 1 and analysis of its 5'-regulatory region. / Immunol. 1991; 147:2777-2786. 40. Mackay F, Loetscher H, Stueber D, Gehr G, Lesslauer W. Tumor necrosis factor a (TNF-a)-induced cell adhesion to human endothelial cells is under dominant control of one TNF receptor type, TNF-R55.JExpMed. 1993; 177:1277-1286. 41. Meichle A, Schutze S, Hensel G, Brunsing D, Kronke M. Protein kinase C-independent activation of nuclear factor kB by tumor necrosis factor. / Biol Chem. 1990; 265:8339-8343. 42. Ghersa P, van Huijsduijnen RH, Whelan J, Cambet Y, Pescini R, DeLamarter JF. Inhibition of E-selectin gene transcription through a cAMP-dependent protein kinase pathway. / Biol Chem. 1994; 269:2912929137.

43. Deisher TA, Garcia I, Harlan JM. Cytokine-induced adhesion molecule expression on human umbilical vein endothelial cells is not regulated by cyclic adenosine monophosphate accumulation. Life Sd. 1993; 53: 365-370. 44. Hidaka H, Inagaki M, Kawamoto S, Sasaki Y. Isoquinolinesulfonamides, novel and potent inhibitors of cyclic nucleotide dependent protein kinase and protein kinase C. Biochemistry. 1984; 23:5036-5041. 45. Gopalakrishna R, Chen ZH, Gundimeda U. Irreversible oxidative inactivation of protein kinase C by photosensitive inhibitor calphostin C. FEBSLett. 1992; 314: 149-154. 46. Kuo JF. Protein Kinase C. New York: Oxford University Press; 1994. 47. Iiversidge J, Thompson AW, Forrester JV. FK 506 modulates accessory cell adhesion molecule expression and inhibits CD4 lymphocyte adhesion to retinal pigment epithelial cells in vitro: Implications for therapy of uveoretinitis. Transplant Proc. 1991;23:3339-3342. 48. Liversidge J, Sewell HF, Forrester JV. Interactions between lymphocytes and cells of the blood-retina barrier: Mechanisms of T lymphocyte adhesion to human retinal capillary endothelial cells and retinal pigment epithelial cells in vitro. Immunology. 1990;71:390-396. 49. Detrick B, Newsome DA, Percopo CM, Hooks JJ. Class II antigen expression and gamma-interferon modulation of monocytes and retinal pigment epithelial cells from patients with retinitis pigmentosa. Clin Immunol Immunopathol. 1985; 36:201-211. 50. Elner VM, Strieter RM, Elner SG, Baggiolini M, Lindley I, Kunkel SL. Neutrophil chemotactic factor (IL-8) gene expression by cytokine-treated retinal pigment epithelial cells. Am J Pathol. 1990; 136:745-750. 51. Elner SG, Strieter RM, Elner VM, Rollins BJ, Del Monte MA, Kunkel SL. Monocyte chemotactic protein gene expression by cytokine-treated human retinal pigment epithelial cells. Lab Invest. 1991;64:819-825. 52. Gabrielian K, Osusky R, Sippy BD, Ryan SJ, Hinton DR. Effect of TGF-/? on IFN-y-induced HLA-DR expression in human retinal pigment epithelial cells. Invest Ophthalmol Vis Sd. 1994; 35:4253-4259. 53. Hofman FM, Hinton DR. Tumor necrosis factor-alpha in the retina in acquired immune deficiency syndrome. Invest Ophthalmol Vis Sd. 1992; 33:1829-1835. 54. Raine C. Multiple Sclerosis: TNF revisited, with promise. Nature Med. 1995;1:211-214.