Endogenous FGF1-induced Activation and Synthesis of Extracellular Signal-regulated Kinase 2 Reduce Cell Apoptosis in Retinal-pigmented Epithelial Cells*

Retinal-pigmented epithelial (RPE) cell survival is critical to the maintenance of the function of the neural retinal and in the development of various retina degenerations. We investigated molecular mechanisms involved in this function by assessing apoptosis in RPE cells following serum deprivation. Apoptosis induced by serum withdrawal is lower in aged RPE cells because of higher endogenous acidic fibroblast growth factor (FGF1) synthesis and secretion. These experiments examined several aspects of FGF signaling and the contribution of endogenous FGF1 to activation of the extracellular signal-regulated kinase 2 (ERK2). In aged RPE cells, FGFR1 was rapidly activated, and its autophosphorylation followed the kinetics of endogenous FGF1 secretion, before the onset of apoptosis. ERK2 phosphorylation, activity, and de novo synthesis increased at the same time. In marked contrast, no de novo JNK1 synthesis was observed. MEK1 inhibition resulted in lower levels of ERK2 activation and synthesis and higher levels of apoptosis. Treatment with neutralizing anti-FGF1 or blocking anti-FGFR1 antibodies mimics these effects. Thus, this study strongly suggests that the survival-increasing effect of FGF1 in aged RPE cells is because of an autocrine/paracrine loop in which the ERK2 cascade plays a pivotal role.


Retinal-pigmented epithelial (RPE) cell survival is critical to the maintenance of the function of the neural retinal and in the development of various retina degenerations. We investigated molecular mechanisms involved in this function by assessing apoptosis in RPE cells following serum deprivation. Apoptosis induced by serum withdrawal is lower in aged RPE cells because of higher endogenous acidic fibroblast growth factor (FGF1) synthesis and secretion.
These experiments examined several aspects of FGF signaling and the contribution of endogenous FGF1 to activation of the extracellular signal-regulated kinase 2 (ERK2). In aged RPE cells, FGFR1 was rapidly activated, and its autophosphorylation followed the kinetics of endogenous FGF1 secretion, before the onset of apoptosis. ERK2 phosphorylation, activity, and de novo synthesis increased at the same time. In marked contrast, no de novo JNK1 synthesis was observed. MEK1 inhibition resulted in lower levels of ERK2 activation and synthesis and higher levels of apoptosis. Treatment with neutralizing anti-FGF1 or blocking anti-FGFR1 antibodies mimics these effects. Thus, this study strongly suggests that the survival-increasing effect of FGF1 in aged RPE cells is because of an autocrine/paracrine loop in which the ERK2 cascade plays a pivotal role.
Fibroblast growth factors (FGFs) 1 are a family of at least 15 polypeptides that stimulate growth and differentiation in cells of various mesenchymal and ectodermal origins. (for reviews, see Refs. [1][2][3][4][5]. Acidic FGF (FGF1) and basic FGF (FGF2) are the prototype members of this family. FGF1 and 2 lack a classic signal peptide (6,7), implying that they are not secreted by the classical secretion pathway. There is evidence that FGF1 and 2 are exported from the cell and subsequently act as autocrine or paracrine factors (8,9). FGF1 and 2 exert their effects via high affinity tyrosine kinase receptors (FGFR1-FGFR4) (for reviews, see Refs. 10 -12) and via lower affinity heparan sulfate proteoglycan binding sites (13)(14)(15). FGFR activation causes tyrosine phosphorylation of the receptor itself, intracellular proteins including phospholipase C␥ (PLC␥) (16), extracellular signal-regulated kinases (ERKs) (17), and of several uncharacterized proteins of 80 -90 kDa (18,19). All cell layers in normal adult retina produce FGF1, as cells no longer differentiate or proliferate. The pattern of FGF1 production suggests that FGF1 may be involved in the regulation of specific spatiotemporal events, including proliferation, migration, differentiation and survival in the retina. In vivo, despite the presence of FGFs in the retinal interphotoreceptor matrix, RPE cells have a limited proliferation capacity consistent with the normal increase in retinal space associated with growth and age. In adult, RPE cells cannot divide in vivo, but they may still require survival factors to inhibit their apoptosis. Cell survival and proliferation may be determined by the amount of factor available and the amount of receptor produced (20). RPE cells produce FGF1 (21), FGFR (22,23), and low affinity binding sites (heparan sulfate proteoglycan) (24) in culture. In RPE cells, ERKs, also known as mitogen-activated protein (MAP) kinases, undergo rapid and biphasic activation in response to exogenous FGF1 and FGF2 (24).
Unlike exogenous FGF1, endogenous FGF1 is not a mitogenic factor. It is a survival factor, both for nondividing epithelial cells (25) and for neuronal cells of the PC12 line (26). We have also shown that the FGF2-stimulated release of endogenous FGF1 is associated with lower levels of apoptosis in RPE cells (27), whereas FGF1 secreted by RPE cells and purified from culture medium causes RPE and retinal Mü ller glial (RMG) cell proliferation and survival via ERK2 activation (28). These data suggested these may be an FGF1 autocrine/paracrine pathway supporting RPE cell survival. FGF activity is regulated in RPE cell subcultures (24). So we used an in vitro approach to investigate several aspects of FGF signaling, including the production and excretion of endogenous FGF1, the expression and autophosphorylation of FGFR1 (the major FGF receptor in RPE cells), and the production and activation of ERK2 in quiescent, confluent RPE cell subcultures, as a function of cell survival. RPE cell cultures from bovine eyes were made to age in vitro by repeated culture passage, an aging strategy termed replicative senescence. This study shows that the amount of FGF1 controls cell survival by FGFR1 activation and by affecting the activation and the production of ERK2. Therefore, long term activation of MAP kinase and up-regulation of its production may be involved in integrating and transmitting transmembrane signals for cell survival.

EXPERIMENTAL PROCEDURES
Cell Culture-Bovine RPE cells were isolated as described previously (24). RPE cells were cultured in Dulbecco's modified Eagle's medium (Life Technologies, Inc.) containing 10% fetal calf serum (Life Technol-ogies, Inc.), 2.5 g/ml fungizone, 50 g/ml gentamycin, and 2 mM Lglutamine at 37°C in 5% humidified CO 2, 95% air. We used cells from the first to the eleventh passages for all experiments as described previously (24). Cells from the first to third subcultures were early passages, and cells from the sixth to eleventh subcultures were late passages. The density of cultured RPE cells in the presence of serum was similar for all passages studied. The production of cytokeratins, markers for differentiated RPE cells, was detected with a monoclonal anti-cytokeratin antibody (AE1/AE3 Boehringer Mannheim). RPE cells were cultured on plastic 6-well culture plates (Falcon) in 2 ml of culture medium for cell proliferation assays, Western blotting, ERK2 activity assay, and labeling experiments. RPE cells were cultured for FGF analysis by enzyme immunoassay (EIA), on 75-cm 2 vented tissue flasks (Falcon) containing 10 ml of culture medium. In some experiments, 200 g of anti-bovine FGF1 neutralizing antibody (R&D Systems) were added per ml of culture medium on days 1 and 3 (27), and monoclonal blocking anti-FGFR antibody (clone VBS1, Chemicon International) (29)  Cell Proliferation and Cell Death Assays-The status of proliferating and quiescent culture RPE cells was determined during the seven days of culture as described previously (27,28). The proliferation of RPE cells was assayed daily by counting the number of cells and by measuring [ 3 H]thymidine (Amersham Pharmacia Biotech, specific activity: 0.92 TBq/mmol) incorporation as described previously (27). The number of dead cells was determined by counting the cells floating on the culture medium and counting the cells remaining on the culture dish after staining with trypan blue and with the MTT (3-(4,5-dimethyl thiazol-2-yl)-2,5-diphenyl tetrazolium bromide) method (30). Cells undergoing programmed cell death (PCD) were detected each day by the terminal dЈUTP nick end labeling (TUNEL) technique (31) using the instructions of the manufacturer (PCD kit, Boehringer Mannheim) as described previously for RPE cells (27,28). The number of labeled cells was counted in three different fields, in three culture wells for each time point. The labeled cells were counted using an ocular grid and a ϫ25 objective. The grid was placed on the culture well, and at least 200 cells were counted per field.
Analysis of FGF1 and FGF2 Production by Enzyme Immunoassay-A second antibody solid phase EIA, as described previously (27,32,33), was used to quantify FGF1 and FGF2 levels in the RPE cell extracts. This was a classic competitive immunoassay using an FGF1-acetylcholinesterase conjugate as tracer. Standard curves were obtained with purified human recombinant FGF1 and FGF2. Mouse anti-rabbit IgG monoclonal antibodies (2 g/well, a generous gift from Dr. J. Grassi, CEA) were adsorbed onto 96-well Immuno plates (Nunc, Denmark). 1% bovine serum albumin in EIA buffer (0.1 M Na-phosphate buffer containing 0.4 M NaCl, 1 mM EDTA, 0.1% bovine serum albumin, 0.001% NaN 3 ) was added to saturate the wells. Cell extract (50 l), purified as described previously (27) diluted in EIA buffer, or 50 l of FGF standard and 50 l of affinity purified rabbit anti-human FGF1 antibody (1:10 4 dilution) or anti-human FGF2 serum (1:10 3 dilution) (35) was added to the wells, and the mixture was incubated for 5 h. 50 l of FGF1 or FGF2 coupled to acetylcholinesterase (1 Ellman unit/ml) was added as a tracer. After an overnight incubation at 25°C and intensive washes, 200 l of Ellman's reagent (10 Ϫ2 M sodium phosphate buffer, pH 7.4, containing 7.5 ϫ 10 Ϫ4 M acetylcholine and 2.5 ϫ 10 Ϫ4 M dithionitrobenzoic acid) was added to each well, and the mixture was incubated for 1-3 h. Absorbance was measured at 412 nm using a Bio-Tek EIA plate reader. Nonspecific binding was determined using an incubation mixture in which the second antibody was replaced with 50 l of EIA buffer. The results are expressed as B/B 0 ϫ 100, where B is the color reaction of the bound fraction in the presence of FGF1. We detected no cross-reaction of the antibody with FGF2 at the dilution used. The amounts of endogenous FGF1 and FGF2 in RPE cells are expressed in ng/mg of protein. Amounts of FGF1 and of FGF2 secreted are expressed in ng/10 ml of culture medium. 10 6 RPE cells contained 1-1.2 mg of protein. The protein concentration of the lysates was determined using a BCA kit (Pierce).
ERK2 Activity Assay-Quiescent, confluent RPE cells were incubated in serum-free medium for 7 days. Each day, cells were lysed in Triton X-100 lysis buffer (50 mM Tris-HCl, pH 7.5, 100 mM NaCl, 50 mM NaF, 5 mM EDTA, 40 mM ␤-glycerophosphate, 0.2 mM sodium orthovanadate, 1 g/ml leupeptin, 1 M pepstatin A, and 1% Triton X-100). Cell lysate containing equivalent amount of protein was then allowed to immunoprecipitate overnight at 4°C with 0.1 g of polyclonal antibody against ERK2 (Santa Cruz). Protein G-Sepharose beads (50 l v/v) were added, and the mixture was incubated for 1 h at 4°C. Immune complexes were collected by centrifugation at 12,000 ϫ g and washed three times in lysis buffer and once with kinase buffer (20 mM Hepes, pH 7.4, 10 mM MgCl 2 , 1 mM dithiothreitol, and 10 mM pnitrophenylphosphate). They were then suspended in 40 l of kinase buffer containing 10 g of myelin basic protein, 50 M unlabeled ATP, and 3 Ci of [␥ 32 P]ATP (Amersham, 5,000 cpm/pmol) per sample. The reaction was stopped after 10 min at 30°C by adding 40 l of 2ϫ Laemmli's sample buffer, and samples were subjected to SDS-PAGE (12% polyacrylamide gel) (34).
Western Blot Analysis-RPE cells were incubated in serum-free medium for 7 days. They were washed twice in phosphate-buffered saline and lysed in ice-cold Triton X-100 lysis buffer and centrifuged at 4°C for 10 min at 10,000 ϫ g. Monoclonal antibody directed against ␤-actin was used as an internal standard for control of protein loading. For ERK2, JNK1, and FGFR1 analysis, 30, 80, and 100 l of Triton X-100 cell lysate, respectively, were mixed with 5ϫ Laemmli's buffer and heated for 5 min at 95°C. The soluble protein of the cell lysates was separated by SDS-PAGE (12% polyacrylamide gel for ERK2 and 7% polyacrylamide gel for JNK1 and FGFR1 analysis), transferred by electroblotting onto nitrocellulose filters, and probed with polyclonal antibodies raised against p42 ERK2, p46 JNK1, and FGFR1 (Santa Cruz). The primary antibodies were detected using a horseradish peroxidase-conjugated goat anti-rabbit second antibody. enhanced chemiluminescence substrates were used to detect positive bands, according to the manufacturer's instructions, and the membrane was used to expose Hyperfilm TM enhanced chemiluminescence (Amersham Pharmacia Biotech). The protein bands detected on the autoradiography were quantified by using an LKB Ultrascan XL laser densitometer (Amersham Pharmacia Biotech).
Statistics-Each figure shows the results of experiments repeated at least three times. All data are expressed as the mean Ϯ S.E.. Statistical comparisons were performed using the two-tailed Student's t test (Gaussian populations with equal S.D.) and the Wilcoxon or Mann and Whitney test (nonparametric).

Aged RPE Cells Are More Resistant to Apoptosis and Synthesize More FGF1 after Serum Withdrawal
Apoptosis Is Lower in Aged RPE Cell Cultures in Serum-free Conditions-Later passages (5th to 11th) of RPE cells were less sensitive to apoptosis after serum withdrawal than early passages (Fig. 1A). We used the TUNEL method to identify PCD at the single cell level and found that, until day 3 of culture, there was less than 1.1% TUNEL-labeled nuclei in late RPE cells. This is similar to the value obtained for cells cultured in the presence of serum. In aged RPE cells, the number of cells undergoing PCD doubled on day 5 and was 5.2-fold higher on day 7. Early passages of RPE cell cultures in the absence of serum underwent apoptosis more rapidly than later passages. In early passages, the number of cells undergoing PCD increased and was 2.2-fold higher (2% PCD) on day 3, 13-fold higher (10.5% PCD) on day 5, and 21-fold higher (16.8% PCD) on day 7 than the basal level of PCD in RPE cells on day 1 of culture (0.8% PCD).
Higher Levels of FGF1 Accumulation and Secretion in Aged RPE Cell Cultures in Serum-free Conditions-We quantified FGF1 production during RPE cell serial passage. Later passages of RPE cells accumulated and secreted more FGF1 than early passages (Fig. 1, B and C). The kinetics of accumulation and secretion of FGF1 also differed between late and early passages. Analysis of FGF protein level by EIA showed that the amount of FGF1 accumulated on day one in cells of late passages (45 Ϯ 1 ng/ml of total protein) was 3.7-fold higher than that in cells of early passages (12 Ϯ 0.5 ng/ml of total protein). The amplitude of the peak of accumulated FGF1 (87 Ϯ 5 ng/mg of total protein) on day 3 was eight-fold higher in late passage RPE cells than in early passage cells (10 Ϯ 2 ng/mg of total protein). Thereafter, the amount of FGF1 decreased on day 4 and then reached a plateau (48 Ϯ 4 ng/mg of total protein). In contrast, the amount of endogenous FGF1 in early passage RPE cells did not change during the first 3 days of culture (Fig.  1C). The concentration of FGF1 then increased significantly and was 2.5-fold higher than the basal level on the 4th day of culture (29 Ϯ 2.5 ng/mg of total protein). It reached a plateau by the 7th day of culture.
In addition, the amount of secreted FGF1 on day 1 in the absence of serum by late passage RPE cells (10 ng/10 ml of culture medium) was higher than that detected in early RPE cell passages (3.2 ng/10 ml of culture medium) (Fig. 1B). Secretion of FGF1 from aged RPE cells was fairly constant throughout the 7-day culture period (10 -13 ng/10 ml of culture medium) and did not follow the kinetics of endogenous FGF1 production. The secretion of FGF1 in early passage cells followed a similar pattern to that of the synthesis of FGF1 (Fig.  1C). Note that the amount of accumulated FGF1 in the cells and in the culture media at the later time points are identical in the late and early passage cells and that the early passage cells begin to accumulate significant amounts of FGF1 as they begin to undergo apoptosis. The amount of FGF2 produced and secreted by RPE cells was low (11 Ϯ 1 ng/mg of total protein produced and Ͻ 0.1 ng/10 ml of culture medium secreted) (27). It did not vary significantly during the 7-day culture period and was unaffected by passage number (data not shown).

Serum-deprivation Causes More Rapid FGFR1 Activation by Secreted FGF1 in Late RPE Cells
Cell survival may be controlled by the amount of receptor produced. We analyzed the de novo synthesis of FGFR1 (the only FGF tyrosine-kinase receptor at the bovine RPE cell surface) during serial passage and investigated its production during the 7-day culture period after serum withdrawal (Fig. 2,  A and B). The de novo synthesis of FGFR1 was constant in both early and late passages during the culture period in the absence of serum, and there was no difference in the level of the de novo synthesis of FGFR1 during the subcultures ( Fig. 2A). In addition, FGFR1 production did not vary significantly during the 7-day culture period in early and late RPE cell passages and was unaffected by passage number (Fig. 2B). Because autophosphorylation of FGFR plays a key role in the interactions between activated tyrosine-kinase receptors and downstream signal transduction pathways, we investigated the autophophorylation of FGFR1 in RPE cells during subcultures in the absence of serum (Fig. 2, C and D). FGFR1 in early passages (Fig. 2C) underwent tyrosine phosphorylation to a lesser extent than in late passages (Fig. 2D). In addition, FGFR1 tyrosine autophosphorylation in early passage RPE cells was weak, did not change during the first three days of culture, and then increased to reach a plateau by the 4th day of culture (Fig.  2C). In contrast, FGFR1 tyrosine autophosphorylation in late passage RPE cells was fairly constant throughout the culture period. Neutralization of secreted FGF1 in early and late passages resulted in reduction of FGFR1 activation to a constant basal level.

Sustained Phosphorylation, Activity, and de Novo Synthesis of ERK2 over a 7-Day Culture Period after Serum Deprivation in Late Passage Cells
Exogenous FGF1 Causes a Sustained ERK2 Activation in Late Passage Cells-We found that there was a relationship between cell survival, FGF1 synthesis and secretion, and FGFR1 activation. This led us to explore the downstream pathways which may be involved in cell survival in late RPE cell passages. We began by investigating the activation of ERK2 by exogenous FGF1 during serial passage (Fig. 3). Activation of FGFR1 resulted in phosphorylation of ERK2 within 5 min, reaching a peak at 15 min in both early and late passages, (Fig.  3, A and B). In early and late passage cells, the first phase of activation was followed by a second sustained phase of activity lasting 12 h. ERK2 activity was determined by measuring phosphorylation of myelin basic protein as substrate (Fig. 3, C  and D). As expected, there was a direct correlation between ERK2 phosphorylation and ERK2 activity in both early and late passages. No change in JNK1 expression was detected after FGF1 stimulation in either early or late passage cells (Fig. 3, E and F). Sustained Activation and de Novo Synthesis of ERK2 over the 7-Day Culture Period in Late Passage Cells-As exogenous FGF1 induced sustained, high level ERK2 activation in late passage cells, we investigated ERK2 phosphorylation and activity in both early and late RPE cell passages over the 7-day culture period in the absence of serum in which various kinetics and amounts of FGF1 excretion were observed (Fig. 4). There was only weak phosphorylation and activity of ERK2 during the 7-day culture period of cells in early passage (Fig.  4A). Contrary to expectation, we found that the total amount of ERK2 was in late passage cells 4 -5-fold higher than that in early passage cells (Fig. 4A). Phosphorylation of ERK2 detected after 24 h was still detected after 7 days of culture, and ERK2 activity was also 4 -6-fold higher than in early passage cells (Fig. 4B).
Protein synthesis was monitored using radiolabeled amino acids on each day of the 7 days of culture, to determine whether the increase in ERK2 during serial passage was because of its de novo synthesis, (Fig. 4C). ERK2 synthesis was low in early passage cells and did not significantly change during the 7-day culture period. De novo synthesis of ERK2 was detected on day 1 of culture, at similar levels in early and late-passage cells. After 24 h in late passage RPE cells, de novo synthesis of ERK2 increased and was 3-fold higher on day 2 and 5-fold higher on day 5, whereas de novo synthesis of JNK1 was unaffected (Fig. 4D).
ERK2 Synthesis and Activation Is Required for RPE Cell Survival-In a control experiment, 10 M PD098059 completely inhibited activation of ERK2 in RPE cells by exogenous FGF1 (data not shown). Inhibition of MEK 1 activity resulted in inhibition of ERK2 synthesis and phosphorylation (Figs. 5, A  and B). ERK2 synthesis was 7-fold lower than the basal levels in early and late passage cells on day 5, and its activation was completely inhibited in early passage cells and 5-fold lower than basal levels in late passage cells. A single dose of PD098059 led to a rapid 8-fold increase in apoptosis after only 24 h of culture in both early and late cell passages (Fig. 5, C and  D). On day 7, cell survival in the presence of the MEK 1 inhibitor was 2-fold lower in early passage cells and 4-fold lower in late passage cells on day 7. There were no nonspecific cytotoxic effects of PD098059 because we detected no significant increase in the number of apoptotic cells in either early or late passages in the presence of serum (Fig. 5, C and D). Furthermore, 10 M PD098059 did not inhibit total protein synthesis (data not shown).  A and B). Cell lysates were immunoprecipitated with the anti-ERK2 antibody, and ERK2 activity was measured as described under "Experimental Procedures" (C and D). Activation of JNK1 was measured by Western blot with anti-JNK1 antibody (E and F). These data are representative of at least three separate experiments.

Lower Levels of Apoptosis via ERK2 Activation Depend on FGF1 Secretion and FGFR1 Activation
We examined the effects of secreted FGF1 on the activation and de novo synthesis of ERK2 (Fig. 6). Neutralization of secreted FGF1 resulted in inhibition of ERK2 phosphorylation after 3 days of culture in both early and late passage cells (Fig.  6A). It also reduced de novo synthesis of ERK2 by 70% after 5 days of culture in both early and late passage cells (data not shown).
We added a blocking anti-FGFR1 antibody to the culture medium to check that the effects of secreted FGF1 on activation of the MAP kinase pathway and on reduction of RPE cell apoptosis were mediated by an autocrine/paracrine stimulation of the tyrosine-kinase receptor FGFR1. We analyzed the effects of FGFR1 blocking on PCD of RPE cells by the TUNEL method (Fig. 6B). Inactivation of FGFR1 had no significant effect during the first 3 days of culture. On day 5, the number of TUNELlabeled nuclei doubled in early passage cells and increased by a factor of 12 in late passage cells in the presence of neutralizing antibody. On day 7, TUNEL labeling of cells in both early and late passages indicated levels of PCD similar to those obtained in the presence of the MEK 1 inhibitor (Figs. 5, C and D, and  6B). Inhibition of the binding of secreted FGF1 to FGFR1 in early RPE cell passages resulted in a 3-fold reduction in the de novo synthesis of ERK2 after 3 days of culture (Fig. 6C). In late RPE cell passages, the blocking anti-FGFR1 antibody reduced FIG. 4. Effects of serum depletion in long-term cultures on ERK2 phosphorylation, ERK2 activity and de novo synthesis of ERK2 and JNK1. Early and late passage RPE cell cultures were incubated in serum-free conditions for 7 days as described in Fig. 1. Phosphorylation of ERK2 (A) and ERK2 activity (B) were measured as described in Fig. 3. For measurement of de novo synthesis of ERK2 (C) and of JNK1 (D), RPE cells were cultured in the absence of serum for 7 days and were given 150 Ci/ml of a mixture containing [ 35 S]cysteine and [ 35 S]methionine daily. Samples of 150 g of proteins were prepared daily with RIPA buffer, and ERK2 and JNK1 were immunoprecipitated with anti-ERK2 and anti-JNK1 antibodies. C and D, time of exposure of autoradiograph was 2 days. Similar results were achieved in three independent experiments.
FIG. 5. Effects of MEK1 inhibition on de novo ERK2 synthesis, ERK2 phosphorylation, and on programmed cell death in early and late RPE cell passages. RPE cell subcultures were incubated in serum-free conditions, in the presence or the absence of the MEK1 inhibitor (Inh), PD098059 added on day 1 of the culture at a concentration of 10 M. De novo ERK2 synthesis (A) and ERK2 phosphorylation (B) were analyzed on days 1, 3, 5, and 7 similarly as described in Fig. 3. Similar results were obtained in three independent experiments. C and D, PCD was measured on days 1, 3, 5, and 7 as described under "Experimental Procedures." Differences between means were analyzed using the Mann and Whitney test. Values are the means of four independent measurements. *, p Ͻ 0.05; **, p Ͻ 0.005.
de novo synthesis of ERK2 by a factor of 3 after 5 days and by a factor of 8 after 7 days (Fig. 6D). Neutralizing anti-FGF1 and blocking anti-FGFR1 antibodies did not affect PCD in RPE cells cultured in the presence of serum and did not significantly alter total protein synthesis (data not shown). This shows there was no nonspecific cytotoxic effect of the antibodies. These results support the notion that the activation of FGFR1 by secreted FGF1 and the subsequent activation and synthesis of ERK2 were required for secreted FGF1 to increase the survival of RPE cells. In this report we show that, in the absence of serum, confluent stationary aged RPE cells, which overproduced FGF1 and secreted a sustained high level of FGF1, were more resistant to apoptosis than cells of early RPE cell passages. FGF1 detected in the RPE cell culture medium was not because of release of this growth factor during cell death because 1) the secretion of FGF1 and its increase preceded RPE cell death, 2) no increase in FGF2 secretion was observed during RPE cell culture when PCD increased greatly (27), 3) RPE cells were intact and without swelling or rupture of membranes when viewed by light microscopy, and 4) the programmed cell death observed on days 5 and 7 did not involve cell lysis.
In a previous study, the direct role of endogenously secreted FGF1 in delaying RPE cell death was demonstrated using an anti-FGF1 neutralizing antibody which caused RPE cell apoptosis (27). However, no mechanistic connection between excretion of FGF1 and RPE cell survival was characterized. Several recent studies show that endogenous FGF1 increases the survival of nondividing cells. Injection of specific FGF1 antisense oligonucleotides into confluent quiescent epithelial lens cells leads to cell death (25). In PC12 cells with a high level of FGF1 production achieved by transfection, there is an increase in the number of surviving cells (26). However, these studies do not distinguish between the potential roles of endogenous FGF1 acting via an intracrine pathway and secreted FGF1 acting through tyrosine-kinase receptors. Fox and Shanley (37) demonstrated an autocrine mechanism for FGF2, mediating survival of vascular smooth muscle cells independent of cell proliferation, and highlighted different roles of exogenous and endogenous FGF2.
FGFR1 de novo synthesis was constant during subcultures of RPE cells and was not modified by serum deprivation. This is consistent with previous data demonstrating that FGFRs were constantly turning over in the presence or in the absence of ligand (38). In contrast, we showed that FGFR1 was more rapidly and to a greater extent activated by secreted FGF1 in aged RPE cells than in early passage RPE cells. This suggests that, FGFR1 activation mediated by an autocrine/paracrine loop was the mechanism involved in reducing apoptosis in late passage cells because aged cells accumulated and secreted a higher amount of FGF1 and because inactivation of FGFR1 resulted in an increase in PCD. This is consistent with previous data showing that autocrine stimulation of growth factor receptors makes it possible for cells to resist PCD (39), whereas lower levels of FGFR1 signal transduction are involved in the inability of senescent endothelial cells to respond to exogenous FGF1 (40).
ERK2 Activation Is Required for FGF1-dependent Inhibition of Apoptosis-At least two pathways are used to transduce FGF1 signals generated by FGFR1 stimulation, one dependent on Ras and the other is dependent on PLC-␥ (41, 42). Raf 1 integrates signals from these two pathways to activate the MAP kinase cascade. ERK2 is vital for integration and trans- mission of the FGF1 signal in RPE cells (28), JNK1 is implicated in the induction of apoptosis after growth factor deprivation (43), and ERK2 and JNK1 have opposing effects on cell death (44). PD098059, which prevents MEK1 activation by Raf 1, does not inhibit JNK1 (45). In our culture system, treatment with PD098059 effectively inhibited ERK2 activation in both early and late passage cells. This led to a rapid increase, after 1 day of culture, in the loss of cell viability, consistent with previous reports on ERK2 inactivation by PD098059 leading to rapid cell death (45) and inhibition of FGF2-protection against apoptosis (46). The effects we observed were specific because the MEK1 inhibitor had no effect on RPE cell apoptosis in the presence of serum although ERK2 was still inactivated (data not shown).
We attempted to define further the molecular mechanisms involved in the regulation of ERK2 expression in serial cultures of RPE cells after serum depletion. We demonstrated that both neutralization of secreted FGF1 and inhibition of FGFR1 inhibited ERK2 synthesis and activity resulting in rapid apoptosis. PCD induced by addition of anti-FGF1 neutralizing or anti-FGFR1 blocking antibodies occurred later than PCD induced by MEK1 inhibition, suggesting that ERK2 activation is presumably an early step in cell activation. This is consistent with previous reports of higher de novo synthesis of FGF1 on days 4 and 5 of the culture period after serum withdrawal (27).
Two mechanisms may be involved in producing the high levels of ERK2 protein detected in later passages. The subculture itself may be important because the amount of ERK2 protein in aged RPE cells cultured in the presence of serum was 4 -5-fold higher than in early passage RPE cells. The depletion of serum may also be involved because de novo synthesis of ERK2 in aged RPE cells increases by a factor of 5, 5 days after serum withdrawal. This suggests that both phosphorylation and synthesis of ERK2 may be required for cell survival. Our results are consistent with recent data demonstrating that up-regulation of MAP kinase expression was a critical element of the metastatic potential of various forms of human breast cancer (47). In our model, the increase in ERK2 synthesis was not because of a general increase in total protein synthesis because synthesis of FGF2 and JNK1 was not affected by serum withdrawal.
The mechanisms by which the ERK2 pathway regulates apoptosis in RPE cells are unclear. Studies with the MEK1 inhibitor PD098059 have shown that activation of ERK isoforms is required for growth factor-induced protein synthesis (48). It has been demonstrated that FGF2-induced up-regulation of Bcl2 gene expression delays apoptosis (49). Moreover, in vivo models of retinal degeneration in which FGFs are upregulated, overexpression of Bcl2 reduces apoptotic photoreceptor cell death (50). This suggests that FGF1 may increase cell survival by activating the ERK2 pathway, leading to transcription of anti-apoptotic genes. This notion is supported by very recent data showing that long-term survival is strongly correlated with the induction of Bclx gene expression, which is dependent on MAP kinase activation (51). Activation of JNK, which antagonizes the anti-apoptotic action of Bcl2 (52) and has opposing effects to ERK2, is also consistent with this. Such mechanisms may account for the effects on survival of FGFs in quiescent and differentiated cells. The synthesis of FGFs and autocrine FGFs signaling may be involved in the survival of the terminally differentiated cells.