Transforming growth factor-beta 1 suppresses serum deprivation-induced death of A549 cells through differential effects on c-Jun and JNK activities.

Transforming growth factor (TGF)-beta1, a pleiotropic cytokine involved in regulating growth and differentiation, can exert both pro-apoptotic and anti-apoptotic effects depending on the cell type or circumstances. We observed that TGF-beta1 blocked apoptosis resulting from serum withdrawal in A549 human lung carcinoma cells. This was associated with suppression of JNK activation that occurs concomitant with the onset of apoptosis in the absence of TGF-beta1, suggesting that JNK plays an active role in the death process and that TGF-beta1 exerts its protective influence by altering JNK activity. Overexpression of a dominant negative mutant form of SEK1, an upstream activator of JNK, likewise suppressed JNK activation and inhibited apoptosis. Investigation of early events following TGF-beta1 treatment revealed an early induction and phosphorylation of c-Jun that was absent in cells subjected to serum withdrawal alone. That TGF-beta1-induced expression of c-Jun is important for survival was supported by the finding that overexpression of non-phosphosphorylatable dominant negative mutant c-Jun, c-Jun(S73A), attenuated the protective influence of TGF-beta1. Our findings suggest that JNK activation is a late but essential event in serum deprivation-induced apoptosis in A549 cells. TGF-beta1 prevents apoptosis, in part, through the early induction and phosphorylation of c-Jun, which in turn results in attenuated JNK activation.

Transforming growth factor (TGF)-␤1, a pleiotropic cytokine involved in regulating growth and differentiation, can exert both pro-apoptotic and anti-apoptotic effects depending on the cell type or circumstances. We observed that TGF-␤1 blocked apoptosis resulting from serum withdrawal in A549 human lung carcinoma cells. This was associated with suppression of JNK activation that occurs concomitant with the onset of apoptosis in the absence of TGF-␤1, suggesting that JNK plays an active role in the death process and that TGF-␤1 exerts its protective influence by altering JNK activity. Overexpression of a dominant negative mutant form of SEK1, an upstream activator of JNK, likewise suppressed JNK activation and inhibited apoptosis. Investigation of early events following TGF-␤1 treatment revealed an early induction and phosphorylation of c-Jun that was absent in cells subjected to serum withdrawal alone. That TGF-␤1-induced expression of c-Jun is important for survival was supported by the finding that overexpression of non-phosphosphorylatable dominant negative mutant c-Jun, c-Jun(S73A), attenuated the protective influence of TGF-␤1. Our findings suggest that JNK activation is a late but essential event in serum deprivation-induced apoptosis in A549 cells. TGF-␤1 prevents apoptosis, in part, through the early induction and phosphorylation of c-Jun, which in turn results in attenuated JNK activation.
Cell death is as essential as cell proliferation for the regulation of mammalian development (reviewed in Refs. [1][2][3][4]. Growth factors play important roles in modulating such processes by transmitting their signals through specific receptor/ signaling pathways, thereby affecting the decision of cell survival or death. The removal of growth factors, hormones, or cytokines can trigger cell death depending on the particular cell's growth factor requirements (5)(6)(7). Insulin-like growth factor-1 (IGF-1), 1 platelet-derived growth factor, and nerve growth factor have all been shown to protect cells from death resulting from growth factor removal in certain cell types (8 -11). However, the signaling pathways leading to apoptosis after growth factor removal, as well as the mechanism(s) whereby specific growth factors rescue cells from death are poorly understood.
The c-Jun-N-terminal kinase (JNK), also known as stressactivated protein kinase (SAPK), is a member of the mitogenactivated protein kinase (MAPK) family of proteins that are activated upon exposure of cells to various cytokines and environmental stresses (12)(13)(14). JNK is activated by dual phosphorylation on Thr and Tyr residues by MAPK kinases (MAPKKs) that are in turn activated by various MAPKK kinases (MAP-KKKs) (13,15). SEK1 (also called MKK4) is one of the major MAPKK responsible for the phosphorylation and activation of JNK (16,17). Activated JNK plays a key role in regulating the activity of transcription factors. It enhances the activities of ATF-2, c-Jun, and Elk-1, but represses activity of NFAT4 (18 -20). Recent studies have also shown that activated JNK phosphorylates and stabilizes the transcription factor p53 protein under stress conditions (21,22). The extracellular signal-regulated kinases (ERK) constitute a separate subfamily of MAPK that are regulated in response to extracellular stimuli, most notably, growth factors (23)(24)(25)(26). ERK activity is regulated primarily through the Raf/MEK/ERK protein kinase cascade and its reduced activity during growth factor deprivation has been correlated with cell death (7,(23)(24)(25)(26).
c-Jun transcription, a major target of JNK, participates in various biological responses (18 -20). Its expression is induced in response to both proliferative and stressful stimuli. c-Jun can form homodimers or associate with other transcription factor partners, including members of the Jun, Fos, and ATF/ CREB subfamilies, to form heterodimeric complexes (19,20). Its activation through phosphorylation by JNK has been implicated in a variety of processes including embryonic development (27), cellular proliferation, and transformation (28,29), and the initiation of apoptosis in response to various stresses (18,30).
Transforming growth factor ␤1 (TGF-␤1) is a member of the TGF-␤ family of multifunctional factors that regulate growth * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18  and differentiation (reviewed in Refs. 31 and 32). It stimulates proliferation of certain cell types (e.g. WI38 human fetal lung fibroblasts, osteoblast, and chondrocytes) while inhibiting the growth of others (e.g. T and B lymphocytes, keratinocytes, and bronchial epithelial cells) (reviewed in Refs. [31][32][33][34]. TGF-␤1mediated growth inhibition has been shown to correlate with dephosphorylation of the retinoblastoma (Rb) protein (35)(36)(37) and induction of the cyclin-dependent kinase inhibitors, p27 Kip1 (38,39), p15 (40), and p21 Waf1 (41,42) in different systems. A role for TGF-␤1 in the induction of cell death has also been established in a variety of cell types (43)(44)(45). On the other hand, several recent studies have provided evidence that TGF-␤1 can inhibit Fas antigen-mediated apoptosis in T lymphocytes and rheumatoid synovial cells, and serum deprivation-induced cell death in macrophages (46 -48). Thus, as is the case for cell proliferation, TGF-␤1 may have dual roles in regulating cell death. TGF-␤1 signals are transmitted through a heteromeric complex between two transmembrane serine/ threonine kinase receptors, the type I and type II receptors (30 -32). Smad-related proteins are thought to be the essential intracellular components of TGF-␤1 signaling pathways that serve to transduce the TGF-␤1 signal from the cell membrane to the nucleus to regulate downstream genes (33,34,49).
Previously we reported that human lung carcinoma A549 cells undergo apoptosis upon serum removal, an effect associated with elevated JNK activity (50). Here, we demonstrate that TGF-␤1 can prevent this apoptosis induced by serum withdrawal. The prevention of cell death by TGF-␤1 in these cells is associated with an early induction and sustained phosphorylation of c-Jun protein but diminution of JNK activation which occurs later during the response to serum removal. Through additional experiments utilizing dominant-negative mutant SEK1 and non-phosphorylatable c-Jun mutants, we provide evidence that modulations of c-Jun and JNK activities are essential events in the ability of TGF-␤1 to influence cell survival during serum deprivation.

EXPERIMENTAL PROCEDURES
Cell Culture and Transfection Conditions-The human lung carcinoma cell line A549 was cultured in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum. For serum deprivation, exponentially growing cells were washed once with phosphatebuffered saline and recultured in Dulbecco's modified Eagle's medium without serum. Growth factors were added from concentrated stock solutions to the serum-free medium to achieve the following final concentrations: five to 25 ng/ml EGF (Life Technologies, Inc.), 5 ng/ml IGF, 5 g/ml insulin, and 5 ng/ml TGF-␤1 (purchased from Upstate Biotechnology and Life Technologies, Inc.). The anti-human TGF-␤1 neutralizing antibody (final concentration of 3 g/ml) was purchased from R & D Systems (Minneapolis, MN).
Transfections were performed using either the calcium phosphate precipitation method as described (51) or LipofectAMINE reagent according to the protocol suggested by the manufacturer (Life Technologies, Inc.). Transfected plasmids included the empty vector pLHCX, the expression vector pLHCX-c-Jun mutant 73 (Ser 3 Ala) (52), a Neo-CMV empty vector, a GST-SEK1(K-R) dominant negative expression vector (provided Dr. J. Kyriakis, Massachusetts General Hospital, MA), and a Ϫ79/ϩ170 Jun-luciferase reporter plasmid containing the first 79 base pairs of the c-Jun promoter (kindly provided by Dr. M. Karin, University of California, San Diego, CA) (53). The pEGFP c1-wild type-SEK1 and mutant-SEK1 (K-R) expression vectors were generated by inserting the full-length wild type-or mutant (K-R)-SEK1 cDNAs into the pEGFP-C1 expression vector (CLONTECH, CA). For generation of the stable transfectants, G418 (700 g/ml) or hygromycin (250 g/ml) was added into the growth medium 48 h after transfection for selection of antibiotic resistant clones. For transient transfection with the c-Jun-luciferase reporter construct, cells were treated with TGF-␤1 (5 ng/ml) 24 h after transfection and were assayed for luciferase activity at the indicated times as described previously (54).
Kinase Assays-JNK and ERK kinase assays were performed as described previously (50,54). Briefly, endogenous JNK1 and ERK2 were immunoprecipitated from cell lysates with anti-p46 JNK and anti-p42 ERK antibodies, respectively (Santa Cruz Biotechnology, Santa Cruz, CA). Extensively washed immunoprecipitates were assayed for kinase activity with buffer supplemented with [␥-32 P]ATP (10 Ci/ml) at 30°C for 20 min and either GST-c-Jun (provided by J. Woodgett) (55) or myelin basic protein (Sigma) as substrates for JNK and ERK, respectively. 32 P-Labeled GST-c-Jun and myelin basic protein products were electrophoresed on 12 or 14% SDS-polyacrylamide gel electrophoresis, respectively, and exposed to x-ray films or Phosphor-Imager for quantitative analysis.
Western Blot Analysis-JNK1 and ERK2 protein levels were analyzed from the same cell lysates used for kinase assays. Briefly, 25 g of total protein from each sample was electrophoresed through 10% SDS-polyacrylamide electrophoresis gel and transferred to supported nitrocellulose membranes. Anti-JNK1 (Santa Cruz Biotechnology), anti-ERK2 (Transduction Laboratories, Lexington, KY), anti-MEK4 (SEK1; Santa Cruz Biotechnology), and anti-c-Jun (Transduction Laboratories and Santa Cruz Biotechnology) antibodies were used to detect JNK, ERK2, SEK1, and c-Jun protein levels, respectively. Phosphorylated ERK2 was examined with an anti-ACTIVE MAPK antibody (which only reacts with phosphorylated forms of ERK1 and ERK2; Promega, Madison, WI). Phosphorylated p38 was examined with an anti-phospho-specific p38 antibody (New England BioLabs, Beverly, MA). Phosphorylated forms of JNK1 and JNK2 were analyzed by Anti-ACTIVE JNK antibody, which specifically reacts with phosphorylated JNK1 and JNK2 (Promega Corp., Madison, WI). Phospho-specific c-Jun (Ser 73 ) (New England BioLabs) and phosph-specific-Jun (Ser 63 ) antibodies (Santa CruZ Biotechnology) were used to detect c-Jun protein phosphorylated at residues Ser 73 and Ser 63 . SEK1(K-R)-GST protein was detected using an anti-GST antibody (Santa Cruz Biotechnology). All proteins were visualized using the enhanced chemiluminescence (ECL) system (Amersham Pharmacia Biotech).
Nuclear Staining and DNA Fragmention Assays for Apoptosis-For DAPI (4,6-diamidino-2-phenylindole) nuclear staining (56), cells grown on either 100-or 60-mm plates were washed three times with phosphate-buffered saline and fixed with 4% paraformaldehyde. Fixed cells were again washed three times with phosphate-buffered saline and stained with DAPI (1 g/ml; Sigma) for 30 min. Stained cells were examined with a fluorescence microscope to identify apoptotic cells. DNA fragmentation was analyzed as described previously (50,57). Briefly, cells grown in the presence or absence of 10% serum were harvested and lysed in a hypotonic lysis buffer (10 mM Tris, pH 7.5, 1.0 mM EDTA, 0.2% Triton X-100) for 30 min on ice. Cell lysates were centrifuged at 14,000 rpm for 10 min to pellet the high molecular weight DNA. Supernatants were collected and extracted with phenol/chloroform followed by precipitation in isopropyl alcohol. The DNA was electrophoresed on a 1.5% agarose gel and visualized by ethidium bromide staining.
Statistical Analysis-An unpaired Student's t test was used to assess statistical significance of differences. A p value of Ͻ0.05 was considered significant.

TGF-␤1 Suppresses Apoptosis Induced by Serum Deprivation in A549
Cells-In keeping with our previous studies (50), serum withdrawal resulted in apoptosis of A549 cells in a timedependent manner (Fig. 1A). This was evident by DAPI nuclear staining as well as by assaying for DNA fragmentation (Fig. 2,  A and B). In order to gain insight into what particular growth factor(s) were important for A549 cell survival, a panel of growth factors, including IGF, EGF, insulin, and TGF-␤1 were examined for their abilities to overcome the serum deprivationinduced cell death. With serum withdrawal alone about 38% of the cells appeared apoptotic by 72 h (Figs. 1 and 2). Neither IGF (5 ng/ml) nor EGF (5-25 ng/ml) altered the response to serum deprivation, but insulin (5 g/ml) supplementation modestly augmented cell death (Fig. 1B).
In contrast, treatment with TGF-␤1 markedly suppressed serum deprivation-induced apoptosis (Figs. 1B and 2A). The TGF-␤1-treated cells were slightly flattened in appearance ( Fig. 2A, lower panel) but continued to divide and remained viable even after 110 h of serum deprivation. The ability of TGF-␤1 to suppress cell death was dose-dependent with marked protection occurring with concentrations of 3 ng/ml or higher (Fig. 1C). That the protective influence was indeed due to TGF-␤1 was evidenced by the fact that it was seen with TGF-␤1 from two different sources and was antagonized by addition of an antibody known to neutralize TGF-␤1 activity (Fig. 1D). This anti-apoptotic effect of TGF-␤1 was not limited to A549 cells, as the cytokine (5 ng/ml) also partially inhibited serum withdrawal-induced cell death in both Rat1a-c-myctransformed fibroblasts and NIH-3T3 mouse fibroblasts (Fig.  2C). Importantly, however, the ability of TGF-␤1 to prevent cell death was not pervasive to all death inducers as similar concentrations of TGF-␤1 (5-25 ng/ml) had no effect on etoposideinduced cell death in A549 cells (data not shown).
Suppression of Cell Death by TGF-␤1 Is Associated with Inhibition of JNK Activation-ERK and JNK have been implicated in regulating the induction of apoptosis triggered by different stresses, with high JNK activity and reduced ERK activity being associated with cell death (7, 58 -61). To investigate the possible mechanism(s) whereby TGF-␤1 exerts its protective effect on serum deprivation-induced cell death, we examined the activities of ERK and JNK in the presence or absence of TGF-␤1. As shown in Fig. 3A, serum withdrawal alone resulted in a modest decrease in ERK activity in A549 cells (Fig. 3A). EGF treatment transiently increased ERK activity bringing it back to the level seen in cells grown in 10% FCS. TGF-␤1 treatment, on the other hand, not only prevented the loss in ERK activity seen with serum withdrawal, but further increased ERK activity relative to that seen in cells grown in 10% FCS. No differences in ERK protein levels were observed with any of the treatments (Fig. 3B).
Examination of JNK activity revealed that it was markedly increased (8-fold) by 48 h following serum withdrawal and remained elevated through 96 h (Fig. 4A). EGF treatment did not alter JNK activity in these serum-deprived cells, but TGF-␤1 treatment significantly reduced the activation of JNK in response to serum withdrawal (Fig. 4A). Not only was maximum JNK activity reduced in the presence of TGF-␤1, but the attenuation of JNK activation was accelerated such that JNK was essentially back to basal levels by 72 h post-treatment. Western analysis of JNK1 protein expression using the same set of cell lysates revealed no change in the levels of JNK protein with TGF-␤1 treatment (Fig. 4B). These effects were not limited to JNK1 as similar activation of JNK2 by serum removal and its suppression by TGF-␤1 was seen in a separate experiment using anti-phospho-specific JNK1 and JNK2 antibodies (Fig. 4C).
Overexpression of Dominant-Negative SEK1 Reduces JNK Activity and Blocks Serum Deprivation-induced Cell Death-To investigate whether the alterations in ERK and JNK activities noted above are important for the suppression of apoptosis by TGF-␤1 we employed other strategies to inhibit their activation. PD98059, which inhibits the activity of MEK (the kinase which lies directly upstream of ERK) (26), was used to suppress ERK activity. Phosphospecific antibodies were used to assess the relative activity of ERK in the presence or absence of the inhibitor. As shown in Fig. 5A, treatment of the cells with PD98059 (20 M) effectively eliminated the phosphorylated (active) ERK regardless of the treatment conditions (Fig. 5A). Interestingly, however, PD98059 treatment did not result in increased apoptosis during serum removal, but rather decreased cell death (Fig. 5B). The mechanism responsible for this effect of PD98059 is not clear. Importantly, however, treatment of cells with PD98059 plus TGF-␤1 did not alter the protective influence seen with TGF-␤1 alone (Fig. 5B). These experiments. B, TGF-␤1, but not IGF, EGF, or insulin, suppresses apoptosis following serum withdrawal. TGF-␤1 (5 ng/ml), IGF (5 ng/ml), EGF (5-25 ng/ml), and insulin (5 g/ml) were added to serum-free medium. Cells were fixed after 72 h serum deprivation followed by DAPI nuclear staining to assess apoptosis. The numbers shown are the mean Ϯ S.E. of three independent experiments. * (p Ͻ 0.01) and ** (p Ͻ 0.05) denote values significantly different from those of cells subjected to serum deprivation alone. C, dose-dependent protective effect of TGF-␤1 on serum-deprived A549 cells. Cells were treated with or without TGF-␤1 in serum-free medium for 72 h, after which, the percentage of apoptotic cells was determined. The data presented are the mean Ϯ S.E. derived from three independent experiments. *, significantly different (p Ͻ 0.01) from cells subjected to serum deprivation alone. D, neutralizing antibody to TGF-␤1 blocks its suppressive effect on serum deprivation-induced cell death. The anti-human TGF-␤1 antibody (3 g/ml final concentration) was added along with TGF-␤1 to serum-free medium. Cells were fixed 72 h after serum withdrawal and cell death was determined as described above. The numbers shown are the average of two independent experiments. TGF-␤1, JNK, c-Jun, and Serum Deprivation-induced Apoptosis data indicate that the enhanced ERK activity seen with TGF-␤1 treatment does not contribute to the cytokines protective effect against serum deprivation-induced cell death.
To address the possibility that TGF-␤1 influences survival through suppression of JNK activity, we established an A549 cell line which constitutively overexpresses a dominant inhibitory mutant of SEK1, SEK1(K-R), fused to GST. This mutant has been shown to effectively block the activity of SEK1, a major kinase responsible for the activation of JNK (16,17). Overexpression of SEK1(K-R) in the A549 transfectants was confirmed by Western analysis using an anti-GST antibody (Fig. 6A). The level of JNK activation seen in SEK1(K-R)expressing cells relative to parental cells and cells containing the empty vector at various times following serum withdrawal is shown in Fig. 6B. The magnitude of JNK activation following serum withdrawal was significantly reduced (35-60% lower than control) in the SEK1(K-R) cells compared with either parental cells or cells transfected with the empty vector. Most importantly, overexpression of SEK1(K-R) effectively blocked apoptosis induced by serum withdrawal (Fig. 6C). These observations were confirmed using another set of A549 cells stably expressing either wild-type SEK1 (WT) or SEK1(K-R) in a different expression vector, pEGFP-C1, where the SEK1 protein is fused to green fluorescent protein (Fig. 6D). While overexpression of SEK1(WT) resulted in a slight increase in apoptotic cells, overexpression of SEK1(K-R) again greatly reduced serum deprivation-induced apoptosis. Together, these findings support the view that JNK is an important mediator of serum deprivation-induced apoptosis and suggest that TGF-␤1 inhibits death, at least in part, by preventing JNK activation.
p38 kinase is a third member of the MAPK family whose activity is regulated by conditions of stress. Although SEK1 predominantly regulates JNK activity, it can also contribute to p38 activation in some circumstances (15,62,64). Therefore, it was of interest to determine whether p38 was also activated during serum withdrawal in A549 cells and if so, whether modulation of its activity would alter survival. Using a phospho-specific p38 antibody, we were unable to detect the pres-  upper panels), serum-free medium for 96 h (middle panels); and serum-free medium supplemented with 5 ng/ml TGF-␤1 (lower panels). Left panels: phase-contrast photomicrographs, magnification ϫ400. Right panels: fluorescence photomicrographs of A549 cells stained with DAPI nuclear dye. Note the presence of condensed nuclei in cells cultured in serum-free medium alone, but not in those cultured in medium containing 10% FCS or serum-free medium supplemented with TGF-␤1. B, internucleosomal DNA fragmentation in serum-deprived A549 cells. DNA extracted from cells cultured in either 10% FCS (left lane) or serumfree medium for 96 h (right lane) was fractionated on 1.5% agarose ethidium bromide-stained gel. C, TGF-␤1 suppresses apoptosis following serum withdrawal in Rat1a-myc and NIH-3T3 cells. Photomicrographs were taken with a phase-contrast microscope at a magnification of ϫ200. Cells were cultured in serum-free medium for 48 h with or without the addition of 5 ng/ml TGF-␤1. Results shown are representative of at least two independent experiments.

TGF-␤1, JNK, c-Jun, and Serum Deprivation-induced Apoptosis
ence of activated p38 in serum-deprived A549 cells, regardless of TGF-␤1 treatment (data not shown). Further evidence for the lack of involvement of p38 in mediating serum deprivationinduced death in A549 cell was obtained using the pharmacologic inhibitor of p38, SB202190. Consistent with the lack of p38 activity in serum-deprived A549 cells, SB202190 treatment did not alter cell death following serum withdrawal nor did it interfere with the protective influence of TGF-␤1 (data not shown).
Protective Influence of TGF-␤1 Is Associated with Early Induction and Phosphorylation of c-Jun-Although the experi-ments described above indicate that JNK activation is an important event for serum deprivation-induced death of A549 cells, it is a relatively late event. TGF-␤1 prevented this JNK activation and it seemed likely that this was due to some earlier effect of the cytokine. Therefore, we sought to identify early signaling events that might account for TGF-␤1's prosurvival influence. Recent studies have provided evidence that c-Jun acts as a downstream effector of JNK-mediated apoptosis (16,20,30), but it is also well documented that c-Jun induction occurs as an immediate early response to growth factor stimulation and plays an important role in supporting cell growth (19,20,28,29). Since TGF-␤1 has been shown to regulate c-Jun expression and AP-1 binding activity in some situations (65)(66)(67)(68), we investigated whether TGF-␤1 affected c-Jun expression in our model system. Examination of c-Jun by Western blot analysis revealed no change in c-Jun protein levels in response to serum withdrawal alone (Fig. 7A). Addition of EGF to serumdeprived cells (which did not alter survival) resulted in only a slight, transient increase in c-Jun protein. In contrast, TGF-␤1-treated cells showed a strong induction of c-Jun protein as early as 3 h after cytokine addition. In addition, the induction of c-Jun protein by TGF-␤1 in serum-deprived cells was not limited to A549 cells, but was also observed in NIH-3T3 cells in which TGF-␤1 also protects against serum deprivation-induced apoptosis (data not shown). The expression of the protein remained highly elevated throughout the first 48 h of treatment. The early c-Jun induction by TGF-␤1 was not affected by overexpression of the dominant-negative mutant SEK1 (Fig. 7B).
A shift in the mobility of the c-Jun protein was also evident in the TGF-␤1-treated cells, suggestive of enhanced phospho- FIG. 3. Effect of TGF-␤1 on ERK activity following serum removal. A, ERK activity of cells cultured in 10% FCS, or serum-free medium with or without growth factor supplementation. ERK2 kinase activity was assayed by an immunocomplex kinase assay using myelin basic protein as substrate. All ERK activities are expressed relative to the level of kinase activity seen in cells grown in 10% FCS. B, the expression of ERK2 was detected by Western blot analysis using an anti-ERK2 antibody. Results shown are representative of at least two independent experiments for each treatment.

FIG. 4. TGF-␤1 inhibits JNK activation following serum withdrawal in A549 cells. A, JNK activity in serum-deprived A549 cells.
Cells were placed in serum-free medium containing TGF-␤1 or EGF for the indicated times. Endogenous JNK was immunoprecipitated using an anti-JNK1 antibody and kinase activity was assayed using GST-c-Jun as substrate. JNK activities are expressed relative to that seen in cells grown in 10% FCS. B, JNK protein expression was analyzed from the same cell lysates by Western blot analysis using an anti-JNK1 antibody. C, phosphorylated JNK1 and JNK2 proteins were detected in a separate experiment utilizing phospho-specific JNK1 and JNK2 antibodies. Results shown are representative of at least two independent experiments for each treatment condition.

FIG. 5. Effect of TGF-␤1 and/or PD98059 treatment on ERK1/ ERK2 phosphorylation and cell death in serum-deprived A549 cells.
A, cells were placed in serum-free medium with or without TGF-␤1 (5 ng/ml) and/or PD98059 (PD) (20 M) for the indicated times. Cell lysates were analyzed by Western immunoblotting using a phospho-specific anti-ACTIVE MAPK antibody. B, effects of TGF-␤1 and PD98059 on cell death in serum-deprived cells. Cells were treated with serum-free medium containing the indicated reagents for 72 h and assayed for apoptosis as described in the legend to Fig. 1. The values shown are the mean Ϯ S.E. derived from three independent experiments. p Ͻ 0.01, comparing differences between apoptosis of serumdeprived cells and serum-deprived cells treated with TGF-␤1, PD, or TGF-␤1 ϩ PD.
TGF-␤1, JNK, c-Jun, and Serum Deprivation-induced Apoptosis rylation of the protein. To provide more direct evidence for the phosphorylation of c-Jun protein, phospho-specific antibodies were utilized. Phospho-specific c-Jun antibody (KM-1), which reacts with phosphorylated c-Jun at Ser 63 , identified phosphorylated c-Jun protein only in TGF-␤1-treated samples (Fig. 7C). Since c-Jun activation is commonly associated with the activation of JNK (18 -20), it was surprising that no c-Jun phosphorylation and/or activation was detected in A549 cells under circumstances where JNK kinase activity was strongly stimulated (serum removal alone for 48 and 72 h) (Figs. 4, A and C, and 6B). To further confirm the phosphorylation status of c-Jun protein in cells treated with or without TGF-␤1 during serum removal, we utilized another c-Jun antibody, which recognizes phosphorylated c-Jun at Ser 73 . Again, this antibody recognized phosphorylated c-Jun species from lysates of cells treated with TGF-␤1, but not cells subjected to serum deprivation alone (Fig. 7D).
To characterize the nature of the c-Jun induction, we transiently transfected A549 cells with a c-jun luciferase (LUC) reporter, containing the first 79 base pairs of the c-jun promoter (53) and examined its responsiveness to TGF-␤1 treatment during serum removal. As shown in Fig. 8, c-jun promoter activity was greatly increased within 3 h of TGF-␤1 addition, suggesting that transcriptional activation of the c-jun promoter contributes to the induction of c-Jun protein in response to TGF-␤1 treatment.
Mutant c-Jun (Ala 73 ) Attenuates TGF-␤1-mediated Protection against Serum Deprivation-induced Cell Death-Given the correlation between the expression and phosphorylation of c-Jun in response to TGF-␤1 treatment and enhanced survival of TGF-␤1-treated cells exposed to serum deprivation, we sought to examine more directly whether active c-Jun contributes to the protective influence of TGF-␤1. Phosphorylation of the c-Jun protein at Ser 73 in its activation domain has been shown to be critical for its function (69,70). Mutation of Ser 73 to alanine abolishes the response to signals generated by growth factors, transforming oncoproteins, and UV irradiation (53,70). These mutants when expressed in otherwise normal cells also act in a dominant negative fashion to block c-Jun activity (52,70). Therefore, A549 cell lines that overexpress the dominant negative mutant c-Jun(S73A) were established through stable transfection (Fig. 9A). Two clones, 11 and 4, which expressed a relatively high level of the mutant protein were selected and compared with vector alone-transfected cells for their sensitivity to serum withdrawal in the absence or presence of TGF-␤1. As shown in Fig. 9B, the c-Jun mutant clones displayed sensitivity to serum deprivation similar to that of their wild-type counterparts. However, the expression of mutant c-Jun markedly attenuated the protective influence of TGF-␤1. The attenuation effect was greatest for clone 4 in which the protection by TGF-␤1 was almost completely nullified. While a significant level of protection by TGF-␤1 was still seen in clone 11 (p Ͻ 0.05), it was greatly reduced relative to vector control cells. DISCUSSION TGF-␤1 is a pleiotropic cytokine that acts in a cell type-dependent manner to regulate growth and differentiation (31,32). It can both stimulate and inhibit proliferation and can exert both pro-apoptotic and anti-apoptotic effects (31)(32)(33)(34)(43)(44)(45)(46)(47)(48). It is, however, most commonly associated with growth arrest and cell death (43)(44)(45). In the present study we have shown that TGF-␤1 suppresses apoptosis induced by serum withdrawal in human lung carcinoma A549 cells and have provided further evidence that modulation of the expression/ activities of c-Jun and JNK contribute to this effect. Recently it was reported that TGF-␤1 also suppresses apoptosis of mouse macrophages induced by serum withdrawal (71). In that study, however, activation of ERK MAPK was implicated as a critical survival factor, as inhibition of TGF-␤1-induced ERK expression abrogated TGF-␤1's protective effect. Although we too observed an elevation in ERK activity in response to TGF-␤1 treatment, it was modest in magnitude, and did not appear to contribute to the anti-apoptotic effect of the cytokine, as blocking ERK activation did not reduce the protective influence of TGF-␤1. It has also been reported that TGF-␤1 can suppress FasL-mediated apoptosis in certain cell types, and one study has suggested that it acts through suppression of p38 MAPK (46 -48). In our model system, p38 was found not to be involved, as p38 was not activated in response to serum deprivation, nor did treatment of cells with the pharmacologic inhibitor of p38, SB202190, influence survival.
Based on the findings reported here, activation of JNK appears to be a critical event leading to cell death in serumdeprived A549 cells, and TGF-␤1 exerts its anti-apoptotic effect, in large part, through inhibition of JNK activation. The observations supporting this conclusion are the following: 1) serum deprivation led to an elevation in JNK activity that preceded the onset of cell death and persisted for the remainder of the treatment period associated with extensive cell death; 2) TGF-␤1 treatment, which prevented cell death, attenuated the magnitude and duration of JNK activation seen in the serum-deprived cells; 3) suppression of JNK activation was specific to TGF-␤1 and did not occur with other growth factor treatments, such as EGF, that failed to inhibit apoptosis; and 4) inhibition of JNK, through expression of a dominant inhibitory mutant of SEK1, was also effective in preventing serum deprivation-induced apoptosis in A549 cells. In our studies, the inhibition of JNK activation by SEK1(K-R) was only partial. This likely reflects the contribution of other kinases, such as MKK7, in the activation of JNK (72). Despite this moderate inhibition of JNK activity, apoptosis was almost completely abolished in cells expressing SEK1(K-R). We have noted a similar phenomenon in a previous study in which partial inhibition of JNK activity also greatly inhibited the level of apoptosis seen in hydrogen FIG. 8. TGF-␤1 stimulates c-jun promoter activity in serumdeprived A549 cells. Cells seeded at 50 -60% confluency were transfected with 1.5 g of DNA/plate of a Ϫ79/ϩ170 jun-LUC reporter plasmid using LipofectAMINE reagent. Twenty-four h later, cells were exposed to TGF-␤1 in serum-free medium and assayed for luciferase activity at the indicated times thereafter (54). Luciferase activity is expressed relative to that seen in untreated cells at time 0. Results represent the means (ϮS.E.) of three independent experiments. peroxide-treated cells (73). It is possible that a critical threshold of JNK activity is required to induce cell death, and that the inhibition achieved with SEK1(K-R) brings JNK activity below this level (74). Our findings are also consistent with studies in PC12 cells in which an inhibition of JNK was shown to suppress apoptosis initiated by withdrawal of nerve growth factor (7). Presumably, suppression of JNK activity blocks the activation of one or more downstream targets required for cell death. While the identity of these JNK effectors is presently unknown, c-Jun can be ruled out as discussed below.
The inhibition of apoptosis by TGF-␤1 in serum-deprived A549 cells is associated with an early induction and phosphorylation of c-Jun protein. c-Jun is an immediate early activator of transcription, and part of the AP-1 complex. This transcription factor serves an important role in regulating genes associated with proliferation and transformation (19,20,28,29), but a number of recent studies have also implicated it in the control of cell death (17,30,61). In such situations, JNK activation, c-Jun phosphorylation, and AP-1 activation are tightly coupled. Indeed, prevention of c-Jun activation through use of dominant negative c-Jun mutants that cannot be phosphorylated has been shown to prevent JNK pathway-mediated cell death in certain instances (17,30). However, other recent studies have shown that c-Jun expression and/or activation is not necessary for apoptosis to occur (75) and a recent study has suggested that c-Jun protects cells against UVC-induced cell death (76). The induction of c-Jun protein and its phosphorylation in response to TGF-␤1 treatment clearly precedes and therefore does not rely on JNK activation. Such discordance between c-Jun expression and JNK activation is also apparent at later times points following serum withdrawal (72 and 96 h), where JNK activation is apparent, but no phosphorylated c-Jun is detectable regardless of TGF-␤1 treatment. Although other studies have shown that Jun induction can occur independent of JNK (63), phosphorylation of c-Jun protein on Ser 63 and Ser 73 is believed to be specific to JNK. Therefore, it appears that either basal levels of JNK activity are sufficient to allow this phosphorylation to occur in TGF-␤-treated cells, or some other as yet unidentified kinase is responsible for this phosphorylation. That induction of c-Jun by TGF-␤1 constitutes a survival signal in serum-deprived A549 cells was strongly supported by the observation that cells expressing c-Jun(S73A), a dominant inhibitory mutant form of c-Jun, sensitized cells to serum deprivation-induced apoptosis and greatly diminished the protective influence of TGF-␤1 (Fig. 9). Two likely explanations for the inability of mutant c-Jun to totally reverse the protective influence of TGF-␤1 are that the mutant c-Jun cannot completely block c-Jun activity, and/or other, as yet undetermined, effects of TGF-␤1 contribute to its ability to influence survival. Although the downstream targets of this early induction of c-Jun expression remain to be determined, it is clearly important for preventing serum deprivation-induced apoptosis in these cells and paradoxically is associated with an attenuation of the later JNK activation that ultimately leads to cell death. These findings contribute to our understanding of the pleiotropic effects of TGF-␤1 and suggest a novel role of this cytokine in regulating cell death signaling through its ability to differentially modulate the expression and activity of components of the JNK signaling pathway.