A Dominant Role for the c-Jun NH₂-Terminal Kinase in Oncogenic Ras-induced Morphologic Transformation of Human Lung Carcinoma Cells¹ Free

SCLC³represents ∼25% of all lung cancers and is characterized by early metastasis and initial responsiveness to chemotherapy and radiation. Despite the initial marked responsiveness to therapy, the vast majority of patients with SCLC relapse and develop drug/therapy resistance and eventually die from the disease (1, 2). Overall, the 5-year survival rate is 3–8% (3). Therefore, novel approaches to the treatment of this rapidly invasive tumor are urgently needed. It is likely that progress will require a better understanding of the intracellular molecular events that determine the malignant behavior of SCLC.

The biological mechanisms underlying tumor progression and the development of chemoresistance in SCLC are still not understood. Clinically, approximately one-third of the recurring chemoresistant tumors exhibit an apparent transition toward NSCLC histology, with cells resembling NSCLC partially or completely replacing SCLC cells (4, 5). A similar in vitro phenomenon has also been observed, involving the accrual of large cell undifferentiated phenotype when SCLC cell lines are maintained for extended periods of time in continuous culture (6, 7). These data suggest that one component of malignant progression may be a movement of the SCLC phenotype along a differentiation continuum linking SCLC with NSCLC,and the potential transition between these phenotypes may play an important role in the development of treatment resistance in patients (8). In support of this hypothesis, in vitrostudies have demonstrated that insertion and expression of a v-Ha-Ras oncogene into SCLC cell lines with amplified Myc family oncogenes induces a phenotypic transition with the acquisition of features typical of the NSCLC phenotype (9, 10). These features include a profound morphological change from SCLC cells in suspension to NSCLC-like adherent monolayer cells,reductions in expression of neuroendocrine markers, and induction of expression of the NSCLC-associated growth factor/receptor genes. Furthermore, this phenotypic transition is accompanied with an acquired resistance to 2-difluoromethylornithine typical of NSCLC carcinoma, but not SCLC, in vitro (11). However, the molecular mechanisms underlying the oncogenic Ras-mediated phenotype transition have not been elucidated.

Ras proteins function as a molecular switch that cycles between an active GTP-bound and an inactive GDP-bound form and play a pivotal role in regulating cell growth and differentiation (12). Activated GTP-bound Ras proteins interact with downstream effectors and promote the activation of at least two distinct serine/threonine kinase cascades, the Raf-dependent MAPK pathway and the Raf-independent JNK/stress-activated protein kinase pathway (13, 14, 15). Ras activation of the Raf-dependent MAPK pathway(Raf/MEK/ERK pathway) involves sequential activation of the Raf kinase,MEK1/2, and ERK1/2. Activated ERKs phosphorylate numerous substrates including other kinases, cytoskeletal elements, and nuclear transcription factors such as Elk-1, resulting in immediate-early gene expression, cell growth, and differentiation (15, 16). Ras also activates a Raf-independent serine/threonine kinase pathway leading to JNK activation. JNK activation up-regulates c-Jun and ATF-2 transcriptional activity (17, 18). Unlike ERKs that are primarily activated by mitogens, the JNKs are potently and preferentially activated by cellular stress (heat shock, osmotic shock,and UV and γ-irradiation) and by inflammatory cytokines (tumor necrosis factor α and interleukin-1; Ref. 19). The JNK pathway consists of MEK kinases, JNK kinases, and JNKs, which are sequentially activated by phosphorylation.

The Raf/MEK/ERK pathway is essential for Ras-mediated transformation (20, 21, 22). However, increasing evidence suggests that Raf-independent signaling pathway(s) are also important for Ras-induced cellular transformation. Studies using Ras effector domain mutants suggest that the coordinated activation of Raf-dependent and Raf-independent pathways by oncogenic Ras is necessary for establishing and maintaining the transformed phenotype (23, 24). Furthermore, it has been demonstrated recently that activation of the JNK pathway via a Raf-independent pathway is also essential for Ras transformation (25). Identification of multiple Ras downstream effectors raises the possibility that a very complex set of signals is generated by Ras that could cooperate to induce cellular transformation (21, 26, 27). Currently, the contribution of multiple downstream effector-mediated pathways to oncogenic Ras signaling and transformation has not been fully elucidated. In particular, the signaling mechanisms by which oncogenic Ras controls proliferation of human cancer cells remain poorly characterized,despite a high frequency of activating mutations of Rasfamily genes associated with many types of human cancer including lung cancer (28).

In this report, we investigate the signaling mechanisms by which an activated Ras oncogene induces a phenotype transformation of NCI-H82 SCLC cells. We demonstrate that, in NCI-H82 cells, oncogenic Ras preferentially activates JNK, and that activation of the JNK-dependent pathway, but not the ERK pathway, is required for the maintenance of the oncogenic Ras-induced transformed phenotype. Our results further suggest that one mechanism underlying JNK-mediated Ras transformation is stimulation of c-Jun/AP-1 activity. Our results indicate that the JNK pathway plays an important role in promoting lung carcinoma cell growth and suggest that the components of the JNK pathway may be possible targets for development of novel antineoplastic drugs in the treatment of carcinoma.

Plasmids pCDNA3-FLAG-JNK1 and pCDNA3-FLAG-dnJNK1 expressing a FLAG-epitope-tagged wild-type JNK1 and a FLAG-epitope-tagged dominant-negative mutant of JNK1, JNK1(APF), respectively, were kindly provided by R. J. Davis (University of Massachusetts,Worcester, MA). Expression plasmids for GAL4-c-Jun 1–223(1–223) and GAL4-c-Jun (1–223; A63/73) and 5×GAL4-LUC were a generous gift from M. Karin (University of California, San Diego, CA) (29). The 3×AP1-LUC plasmid was constructed by insertion of a double-stranded 40-mer deoxyoligonucleotide containing three copies of the AP-1 consensus sequence (30) in the upstream region of the SV40 promoter at the XhoI site of pGL3-promoter (Promega Corp., Madison, WI). GST-c-Jun 1–79(1–79) was purchased from Stratagene (La Jolla, CA). Rabbit polyclonal antibodies[anti-ERK1 (C-16), anti-JNK1 (C-17), and anti-c-Fos] and anti-cyclin D1 mouse monoclonal antibody were obtained from Santa Cruz Biotechnology (Santa Cruz, CA). The mouse monoclonal antibodies,anti-Ras and anti-c-Jun, were purchased from Transduction Laboratories(Lexington, KY). The phospho-specific c-Jun (serine 63) antibody was purchased from New England Biolabs (Beverly, MA). MBP and G418 sulfate were purchased from Life Technologies, Inc. (Gaithersburg, MD). PD98059 and hygromycin B were from Calbiochem (San Diego, CA).

The NCI-H82 SCLC cell line was maintained in RPMI 1640 supplemented with 9% (v/v) calf serum, 100 units/ml penicillin, and 100 units/ml streptomycin. Stably transfected cells were maintained in the same medium plus 400 μg/ml G418 or 250 μg/ml hygromycin B. Stable transfection of NCI-H82 cells was carried out using an electroporation method (31). Cells were selected in G418-containing growth medium. For double transfectants, Ras-expressing cell lines were secondarily transfected with pCDNA3-FLAG-dnJNK1 along with pSV-HygB2 at the molar ratio of 5:1 using Lipofectamine and selected in hygromycin-containing growth medium. Transient transfections were carried out using either Lipofectamine or Lipofectin, according to the procedures recommended by the manufacturer (Life Technologies). For transient transfection luciferase assays, plasmid pSV-β-gal (Promega)was cotransfected with the luciferase reporter constructs to monitor transfection efficiency. Cells were harvested 48 h after transfection and lysed in 1× Reporter Lysis Buffer (Promega). For serum starvation, cells were cultured in a serum-free medium supplemented with 1% BSA for 24 h before lysis. Assays for luciferase and β-galactosidase activities were performed as described using kits purchased from Promega and Clontech (Palo Alto, CA),respectively. Data shown are normalized luciferase activities as a ratio of luciferase activity (RLU/μl) to β-gal activity (unitβ-gal/μl).

Soft agar assays were carried out in six-well plates as described by Clark et al. (32). Briefly, cells were seeded in 0.3% agarose in growth medium overlaid on a base of 0.6% agarose. Cultures were fed weekly. Colonies were scored 14 days after plating. For growth curves, 1×10⁵ cells/well were plated in six-well plates in complete medium or low serum medium containing 0.5% calf serum. Media were changed every 2 days. Cell number was determined by trypan blue exclusion. Triplicate cultures of each cell clone were prepared and processed in all experiments.

Cells were lysed in a modified RIPA buffer [50 mm HEPES(pH 7.5), 150 mm sodium chloride, 1% Triton X-100, 2 mm EDTA, 2 mm EGTA, 50 mm sodiumβ-glycerophosphate, 5 mm sodium pyrophosphate,50 mm sodium fluoride, 1 mm sodium orthovanadate, 1 mm DTT, 1 mmphenylmethylsulfonyl fluoride, 10 μg/ml leupeptin, and 10 μg/ml aprotinin]. Endogenous ERK1/2 and JNK1 were immunoprecipitated from 50μg of cleared cell lysate using the appropriate rabbit polyclonal antibodies in conjunction with protein A-agarose (Life Technologies,Inc.). The immunoprecipitates were washed twice with lysis buffer and twice with kinase assay buffer. Kinase reactions were carried out by incubating the immunoprecipitates with substrate in 30 μl of kinase assay buffer [25 mm HEPES (pH 7.5), 20 mmmagnesium chloride, 0.1 mm EGTA, 50 mm sodiumβ-glycerophosphate, 0.1 mm sodium orthovanadate, 1 mm DTT, and 1 μm okadaic acid] supplemented with 20 μm ATP and 5 μCi[γ-³²P]ATP at room temperature (22°C) for 25 min. Ten μg of MBP and 1 μg of GST-c-Jun 1–79(1–79) were used as substrates for ERK1/2 and JNK activities, respectively. Kinase reactions were terminated by the addition of 30 μl of 2× Laemmli sample buffer. The phosphorylated proteins were resolved on a 10%polyacrylamide-SDS gel and analyzed by autoradiography and InstantImager (Packard Instrument, Meriden, CT).

Whole-cell lysates (50–100 μg) each sample were resolved on a polyacrylamide-SDS gel and electroblotted onto a nitrocellulose membrane. Protein expression was determined by immunoblotting with appropriate antibodies. Briefly, the membrane was blocked in TBST [20 mm Tris-HCl (pH 7.5), 500 mm sodium chloride,and 0.05% Tween 20] containing 5% nonfat dry milk at 37°C for 2 h and then incubated with primary antibody in TBST containing 1% BSA (fraction V) at room temperature for 2 h, followed by incubation with secondary antibody conjugated with horseradish peroxidase in TBST containing 5% nonfat dry milk. Visualization of the protein was by the enhanced chemiluminescence detection system(Amersham, Arlington Heights, IL).

NCI-H82 is an established human SCLC line that contains no Ras gene mutations (33), which has been shown to undergo changes toward NSCLC-like phenotype in response to v-Ha-Ras oncogene expression (9). In this study, a constitutively activated Ras oncogene,c-Ha-Ras(V12) with a point mutation at codon 12 (Gly→Val) (34), was introduced into NCI-H82 cells by stable transfection. Clonal populations were generated from individual G418-resistant colonies. Northern blot analysis was used to determine steady-state levels of Ras mRNA. Elevated p21-Ras protein expression was observed in those clonal populations that have increased steady-state levels of Ras mRNA (Fig. 1,B and data not shown). The Ras-expressing clonal cell lines underwent profound morphological changes from floating aggregates of loosely packed cells to adherent monolayer cells, whereas the morphology of vector-transfected cells [Vector(Neo)], a pool of G418-resistant colonies from the vector transfected NCI-H82 cells, was indistinguishable from the parental cells (Fig. 1,A). The morphological changes in Ras-expressing cells were accompanied by marked alterations in growth parameters. Ras-expressing cells showed greatly reduced serum dependence, proliferating well in medium containing 0.5% calf serum (Fig. 1,D), and had an increased colony-forming activity on soft-agar (Fig. 1 C). Consistent with a previous report (9), these data indicate that expression of an activated Ras oncogene in NCI-H82 cells promotes transforming activities and induces changes in cell morphology.

The involvement of the Raf/MEK/ERK pathway in Ras transformation has been well documented. Thus, it was determined whether the Raf/MEK/ERK pathway was activated after oncogenic Ras expression. To control for any potential variation among clonal populations, Fig. 1. Oncogenic Ras causes morphological transformation of NCI-H82 cells. A, morphology of clones stably expressing an activated c-Ha-Ras(V12) oncogene (×200). B, Western blot analysis of p21-Ras protein expression. Whole-cell lysate (50 μg) from representative cell clones was separated on a 12% polyacrylamide-SDS gel, transferred to nitrocellulose membranes, and immunoblotted with a monoclonal antibody against rat Ha-Ras. C, anchorage-independent growth. Cells were seeded in six-well plates in growth medium containing 0.3%agarose over a base layer of 0.6% agarose. Colonies were scored 14 days after seeding. The efficiency of colony formation (%) represents the ratio of the number of colonies formed to the total number of cells seeded. D, growth curves in low serum (0.5% calf serum)medium. Cells were plated at 1×10⁵ cells/well in six-well plates. Cell number was determined by trypan blue exclusion at day 3,7, and 10. Data shown are the means of triplicate cultures from one representative experiment; bars, SD. At least three independent experiments were done with similar results. RasV12-9,RasV12-11, and RasV12-27 are clonal populations of c-Ha-Ras(V12)-transfectants of NCI-H82 cells. Vector(Neo) is a pool of vector-transfected NCI-H82 cells.\. three Ras-expressing clonal cell lines (RasV12-27, RasV12-11, and RasV12-9) that express roughly equivalent levels of activated Ras (Fig. 1,B) were used in all studies. Raf-1 kinase activity was assayed using the immunoprecipitated Raf-1 from whole-cell lysates to measure its ability to phosphorylate the Raf-1 substrate, a kinase-dead MEK-1, in the presence of [γ-³²P]ATP (35). Surprisingly, the Ras-expressing cells showed no significant increase in the Raf-1 kinase activity (data not shown). Moreover, ERK activity was only slightly elevated (1.5–2.2-fold)relative to parental and vector-transfected control cells (Fig. 2 A), consistent with the lack of up-regulated Raf-1 kinase activity in these cells. These data indicate a modest activation of the Raf/MEK/ERK pathway by oncogenic Ras in NCI-H82 cells.

Oncogenic Ras also activates JNKs through Raf-independent pathway(s) (18, 21, 23). Furthermore, several lines of evidence have suggested a functional role for JNK in Ras-induced and other oncogene-induced transformation (19, 25). Therefore, the JNK activity of Ras-expressing cells was examined in immunocomplex kinase assays. Endogenous JNK was immunoprecipitated from whole-cell lysates using an anti-JNK polyclonal antibody specific against JNK1(p46), and kinase activity of the immunoprecipitated JNK was measured using GST-c-Jun as a substrate in the presence of[γ-³²P]ATP. As shown in Fig. 2 B,JNK activity was robustly increased (26–33-fold induction) in response to oncogenic Ras expression. All of three Ras-expressing clonal cell lines display a very high level of sustained JNK activity.

The role of JNK signaling in Ras transformation of NCI-H82 cells was explored further using a FLAG-epitope-tagged dominant-negative JNK1 mutant called JNK1(APF) (Ref. 17). Transient expression of this mutant has been shown to inhibit JNK-mediated transcriptional regulation (17, 36). To evaluate the long-term effect of inhibition of JNK signaling, an expression plasmid for FLAG-JNK1(APF)was introduced into Ras-expressing cells by stable transfection. Multiple clones expressing different levels of FLAG-JNK1(APF) were generated (Fig. 3,A). JNK1(APF)-expressing cell clones were morphologically reverted, growing as floating aggregates of packed cells similar to the parental NCI-H82 cells (compare Figs. 1,A and 3,B;data not shown). Furthermore, JNK1(APF)-expressing cell clones showed a complete loss of their ability to form colonies on soft agar (Fig. 3,C) and a marked reduction in their ability to grow in low serum medium (Fig. 3 D). Expression of JNK1(APF) had little effect on cell growth in complete medium (data not shown). Therefore,the inhibitory effect of JNK1(APF) on Ras-induced changes in cell growth properties appears to be specific.

Because JNK1(APF) is indicated by the slower migrating bands detected with the anti-JNK1 antibody. B, morphology of JNK1(APF)-expressing cell clones (×200). C, anchorage-independent growth. Colony formation on soft agar was determined with 3000 cells/well as described in “Materials and Methods.” Results are expressed in a bar graph as the number of colonies formed. Data are presented as the average of triplicate cultures from one representative experiment. Two independent experiments were performed with similar results. D,growth curves in the low serum medium. Log-phase cells(1×10⁵) were seeded in a low-serum medium containing 0.5%calf serum. Cell number was determined by trypan blue exclusion on days 2, 4, and 6. Data are presented as the means from two independent experiments performed in triplicate; bars, SD. APF-1, 8,11, and 22 are clonal populations of JNK1(APF)-transfected RasV12-11 cells.\. Ras-expressing cells display a modest up-regulation of the Raf/MEK/ERK pathway, experiments were performed to determine the effects of inhibition of ERK activation on Ras transforming activities. In contrast to the effects of dominant-negative JNK1(APF), inhibition of ERK activation using the MEK inhibitor PD98059 (Refs. 37and 38; 50 μm, up to 72 h) did not lead to morphological reversion of Ras-transformed cells (Fig. 4,A). However, PD98059 inhibited the low-serum growth of Ras-transformed cells (Fig. 4,B). The reduction in cell growth in low serum medium was 60–80%, compared with cells treated with the vehicle (0.1% DMSO). The inhibitory effect of PD98059 on Ras-induced low-serum growth seems to be less specific, because PD98059 also causes a significant reduction (50–60%) in cell growth when they were cultured in complete medium (Fig. 4 B).

A major target of the JNK signaling pathway is the AP-1 transcriptional activator, which is composed of homo- or heterodimers of the Jun and Fos family proteins (19, 39). In particular, AP-1 activity is induced by the Ras GTPases and phorbol esters that activate the MAPKs (40). Therefore, the AP-1 transcriptional activity was examined in Ras-expressing cells using an AP-1-dependent luciferase gene expression system (see “Materials and Methods”). As shown in Fig. 5 A, AP-1-dependent luciferase activity was elevated∼3–6-fold in the Ras-expressing cells, indicating a stimulation of AP-1 activity by oncogenic Ras.

One possible mechanism for the increased AP-1 activity is that the Ras-mediated kinase activation increases the expression of Fos and Jun family proteins. Structure/function studies of the Jun and Fos proteins have demonstrated that induction of AP-1 activity is dependent upon increases in the abundance of these AP-1 components as well as stimulation of their transcriptional activity (39, 40, 41). Therefore, Western blot analysis was performed to determine whether the levels of c-Fos and c-Jun expression were altered in response to oncogenic Ras expression (Fig. 5,B). Although no significant changes in c-Fos protein levels were observed, the c-Jun level was substantially increased in Ras-expressing cells. More importantly,perhaps, the level of phosphorylated c-Jun on the serine 63 was significantly increased in these cells. JNKs can phosphorylate c-Jun on serine 63/73 to stimulate c-Jun transcriptional activity that may also contribute to AP-1 induction (39, 41). Thus, the transcriptional activity of c-Jun was determined in a transient transfection assay. In this assay, GAL4-dependent luciferase reporter gene expression is mediated by the coexpression of either GAL4-c-Jun 1–233(1–233) (wild type) or GAL4-c-Jun (1–233; A63/73) (mutant). The mutant GAL4-c-Jun fusion protein contains serine to alanine mutations at residues 63 and 73 in the transactivation domain of c-Jun and thus cannot be phosphorylated by JNK (29). Consistent with the increased level of phosphorylated c-Jun on serine 63, transient transfection results (Fig. 5,C) indicate that the c-Jun transcriptional activity is elevated in the Ras-expressing cells transfected with wild-type GAL4-c-Jun (3-fold induction over Vector(Neo) cells). However, c-Jun transcriptional activity was not increased when measured in cells transfected with the mutant GAL4-c-Jun fusion protein (Fig. 5 C), suggesting that the increased c-Jun transcriptional activity involves phosphorylation of c-Jun on serine 63/73. Additionally, Northern blot and nuclear run-on transcription analysis demonstrate that the steady-state level of c-Jun mRNA as well as rates of c-Jun gene transcription were increased in Ras-expressing cells (data not shown). These results indicate that Ras activation of JNK results in c-Jun phosphorylation in its NH₂-terminal transactivation domain, which stimulates its transcriptional activity and leads to enhanced c-Jun gene expression. The data suggest that the elevated c-Jun in association with constitutively expressed c-Fos may contribute to the Ras-induced stimulation of AP-1 activity.

Because c-Jun is a downstream target of the JNK pathway and its induction is correlated with JNK activation in Ras transformed NCI-H82 cells, the effect of JNK1(APF) on c-Jun expression was also examined. c-Jun expression (Fig. 3 A) was lower in the JNK1(APF)-expressing cell clones 1, 8, 11, and 22 than in the vector-transfected and untransfected RasV12-11 cells. As expected,c-Fos expression was not changed in these cells (data not shown). These results suggest that c-Jun induction by Ras is mediated by JNK.

Genetic evidence demonstrates that AP-1 activity is crucial for Ras transformation and that c-Jun expression is required for activation of AP-1 by Ras (42). Because Ras transformation of NCI-H82 cells results in c-Jun induction and JNK1(APF) suppresses this effect,it is predicted that JNK1(APF) may also suppress AP-1 induction by oncogenic Ras. Consistent with this prediction, Fig. 6,A shows that AP-1-dependent luciferase activity in JNK1(APF)-expressing cells is 70% lower in clones 1 and 8 and 36%lower in clone 22 than that in the vector-transfected cells(RasV12-11 + Vector). The degree of decreases in the AP-1 activity correlates well with the ability of these cells to grow in low-serum medium (Figs. 3,D and 6 A). These results suggest that c-Jun is the key component of the AP-1 complex, and its induction can directly contribute to increased AP-1 transcriptional activity.

Because Ras-expressing cells display modest up-regulation of the Raf/MEK/ERK pathway, experiments were performed to determine whether inhibition of ERK activation would inhibit the oncogenic Ras-induced stimulation of AP-1 activity. For this analysis, the MEK inhibitor PD98059 was used to specifically block ERK activation. Transiently transfected cells were cultured in serum-free medium for 18 h prior to treatment and then treated with 50 μm PD98059 for 4 or 8 h. Results indicate that the elevated AP-1 activity in Ras-expressing cells was not affected by PD98059 (Fig. 6,B),suggesting that the activation of ERKs is not required for the induction of AP-1 activity. Consistent with this data, treatment of Ras-transformed NCI-H82 cells with PD98059 (50μ m, up to 72 h) did not cause down-regulation of c-Jun expression (Fig. 6 C). Thus, the JNK pathway is the major Ras-dependent signaling pathway leading to stimulation of c-Jun/AP-1 activity.

This study investigates the role of two distinct signal transduction pathways, the JNK pathway and the Raf/MEK/ERK pathway, in the oncogenic Ras-induced phenotype transformation of NCI-H82 human lung carcinoma cells. In NCI-H82 cells, oncogenic Ras preferentially activates JNK. Stable expression of the dominant-negative mutant JNK1(APF) inhibits JNK signaling, reverses transformed morphology, and inhibits anchorage-independent and low-serum growth of oncogenic Ras-expressing cells. These results demonstrate that activation of the JNK pathway is absolutely required for the maintenance of the oncogenic Ras-induced transformed phenotype. A direct correlation of c-Jun/AP-1 activity and JNK-dependent Ras transforming activity suggests that AP-1 is a downstream component of the JNK pathway in Ras transformation. Our data clearly demonstrate a primary role for JNK in the oncogenic Ras-induced phenotype transformation of NCI-H8 human SCLC cells.

The role for JNK in Ras transformation has been suggested by several studies. Expression of oncogenic Ha-Ras leads to phosphorylation of c-Jun on the same sites (serines 63 and 73)phosphorylated by JNK and a c-Jun transdominant inhibitor protein,c-Jun (S63A, S73A) in which two serine sites were mutated to alanines,blocks Ras transformation (43). Expression of a dominant-negative mutant of JNK kinase, an upstream activator of JNKs,causes a significant inhibition of oncogenic Ras-induced focus-forming activity (25). Studies have demonstrated that members of the Rho family proteins, Rac1/2 and Cdc42, selectively regulate the activity of the JNK pathway and play essential roles in mediating Ras transformation (29, 44). Furthermore, a role for JNK in cell transformation induced by other oncoproteins has also been suggested (19). It has been shown that the JNK pathway is constitutively activated in cells transformed by Bcr-Abl,and expression of dominant negative c-Jun or JIP-1, a cytoplasmic inhibitor of JNK, markedly inhibits transformation of pre-B cells by Bcr-Abl (45, 46). Additionally, transformation of NIH 3T3 cells mediated by v-Crk is inhibited by dominant negative JNK kinase 1 (47). These data suggest that JNK may be a universal mediator of cell transformation in response to different transforming agents. However, whether the JNK pathway contributes to oncogenic Ras transforming activity in human cancer cells is unclear. Our study provides mechanistic evidence that suggests a role for JNK in mediating oncogenic Ras action in human lung cancer cells. This is supported by recent studies from Bost et al.(48), which demonstrate a significant growth-promoting role for JNK in human lung cancer cells.

In contrast to the JNK activation, overexpressing the activated Ras oncogene does not efficiently activate the Raf/MEK/ERK pathway in NCI-H82 cells. One explanation might be that the ERK activity in the immortal cancer cells may be already elevated, and therefore expression of an activated Ras has no additional effect. Indeed, we observed a high basal level of ERK activity in serum-starved NCI-H82 cells. Alternatively, the lack of ERK activation may be attributable to down-regulation of ERK activity by modulating the expression of dual specificity phosphatase(s).⁴The existence of negative regulatory mechanisms that repress ERK activity has been suggested in fibroblast cell lines transformed by different oncoproteins including oncogenic Ras (49). Furthermore, the lack of Raf/ERK activation in Ras-transformed NCI-H82 cells suggests that effectors other than Raf may mediate the oncogenic Ras function. The observation that inhibition of ERK activation via a pharmacological approach using a MEK inhibitor PD98059 is unable to fully reverse the Ras-induced transformed phenotype suggests that constitutively active ERK is not essential for maintenance of the Ras-induced transformed state in NCI-H82 cells. Supporting our findings, recent works from Luo and Sharif (50) and Yip-Schneider et al. (51) show that expression of oncogenic Ki-Ras does not lead to constitutive ERK activation in human astrocytoma cells and pancreatic carcinomas. However, that the administration of PD98059 leads to inhibition of cell growth under conditions of exponential growth or serum starvation (Fig. 4 B) suggests that ERK activity appears to be required for normal proliferation of Ras-transformed NCI-H82 cells.

Our results reveal distinct differences in the signaling pathways that Ras uses to cause transformation of fibroblasts and the epithelial cell-derived cancer cells. In rodent fibroblasts, expression of an activated Ras oncogene has been reported to constitutively activate the Raf-dependent MAPK pathway. Consequently, inhibition of ERK activation by PD98059 reverses the Ras-transformed phenotype including morphological reversion (38), suggesting a dominant role for the Raf/ERK pathway in mediating Ras function. However, Ras effector mutants with an impaired ability to bind Raf were still capable of causing transformation in fibroblasts (23). When coexpressed with activated Raf-1, these mutants transform cells synergistically, suggesting that both Raf-dependent and Raf-independent signals are required for Ras transformation. There is increasing evidence that effector-mediated signaling pathways that contribute to oncogenic Ras function are regulated in a cell type-specific manner. For example, although oncogenic Ras and constitutively active Raf mutants each induced transformation of NIH 3T3 fibroblasts, activation of the Raf/MEK/ERK pathway alone was unable to cause potent morphological and growth transformation of RIE-1 rat intestinal epithelial cells (52). Recent works have demonstrated that transformation of RIE-1 cells is Raf independent and appears to be mediated by an autocrine-dependent mechanism (53). This indicates that Ras transformation associated with a certain MAPK module can be cell type specific, and cell type-specific factors, such as developmental background and cell cycle regulation, may be important determinants of the biological outcome of Ras signaling. However, it is worth noting that some common features in Ras transformation, such as up-regulation of AP-1 activity (see below),do exist in both fibroblasts and epithelial cells, suggesting that different MAPK module signals can converge.

Induction of c-Jun expression in Ras-transformed NCI-H82 cells seems to be mediated primarily by the JNK pathway. The mechanisms by which Ras-mediated JNK activation lead to c-Jun induction may include the transcriptional regulation of c-Jun gene expression as well as the regulation of c-Jun protein stability (19). JNK-mediated c-Jun phosphorylation at serines 63 and 73 within its NH₂-terminal transactivation domain can stimulate c-Jun transcriptional activity and increase the half-life of c-Jun protein by inhibiting ubiquitination, thereby leading to c-Jun induction. Our results demonstrate that the enhanced c-Jun transcriptional activity in Ras-expressing NCI-H82 cells is dependent on the phosphorylation status of c-Jun at serines 63/73. Furthermore,Ras-expressing cells show an increase in the steady-state level of c-Jun mRNA and an elevated c-Jun transcription rate, indicating that c-Jun induction is in part attributable to JNK-mediated transcriptional regulation of c-Jun gene expression. The elevated c-Jun gene expression may result partially from positive autoregulation of a TRE/AP-1 site in its promoter that is recognized by c-Jun-ATF2 heterodimers (39, 54). Similar to c-Jun, the transcriptional activity of ATF2 is stimulated by phosphorylation of sites in its transactivation domain by JNK (17). A direct role for JNK in c-Jun induction is further supported by the fact that expression of JNK1(APF) decreases c-Jun protein expression. However, it is unknown whether Ras activation of JNK also increases the half-life of c-Jun protein (55).

The correlation between JNK-mediated c-Jun/AP-1 activation and Ras transforming activity suggests that one mechanism underlying JNK-mediated Ras transformation is up-regulation of AP-1 activity. The constitutive activation of AP-1 transcription factors is thought to be a critical event in Ras-mediated transformation. Inhibition of AP-1 activity by dominant negative c-Jun mutants reverts the transformed phenotype of Ras-overexpressing NIH 3T3 fibroblasts (56, 57). Furthermore, c-Jun^−/− cells cannot be transformed by activated Ras protein and are markedly impaired in their AP-1 transcriptional response (42),suggesting that AP-1 complexes containing c-Jun are essential downstream effectors of Ras. Other studies show that c-Jun activity/expression is necessary for the initiation and/or maintenance of the Ras-transformed state (42, 55, 58). The data presented here are consistent with these reports; Ras transforming efficiency correlates with the steady-state level of c-Jun protein, and stable expression of JNK1(APF) inhibits the effects of oncogenic Ras and results in decreased c-Jun/AP-1 activity, but the MEK inhibitor PD98059 does not. These results suggest that JNK activation is the major Ras-dependent signaling pathway leading to increased c-Jun/AP-1 activity.

The up-regulation of AP-1 activity appears to be a common feature in the in vitro transition from SCLC to NSCLC phenotype. A recent work from Risse-Hackl et al. (59) shows that the H-Ras/c-Myc-mediated transition of SCLC to NSCLC phenotype is accompanied by a strong induction of AP-1-binding activity. Interestingly, it was found that AP-1 is abundantly present in NSCLC cells but not in SCLC cells. Moreover, the induction of AP-1-binding activity in phenotypically converted SCLC cells is intimately linked to the up-regulation of AP-1 target genes, such as CD44, which is preferentially expressed in NSCLC-type tumor and cell lines (60). These data suggest that AP-1 may be an important mediator during lung cancer development/progression. Therefore, by modulating AP-1 activity, and probably that of other transcription factors, the JNK pathway may mediate the long-term effects of Ras in oncogenic transformation and tumorigenesis. It should be noted that in contradiction to our results, a recent report (61) suggests that elevated JNK or AP-1 activities are not required to maintain the Ras-induced transformed state in NIH 3T3 fibroblasts. This result suggests that the dependence of Ras-induced transformation on the JNK/c-Jun/AP-1 pathway is cell specific.

In summary, the data presented here provide direct evidence that JNK plays an essential role in Ras transformation of NCI-H82 human lung carcinoma cells. Our results suggest that persistent JNK activation may promote proliferation and transformation of cancer cells. With respect to the cell type-specific responses to JNK activation (19), it is likely that the effect of JNK signaling is cell context dependent and may be modified by the activation state of other cellular signaling pathways. Future studies in a range of carcinoma cells harboring Ras-activating mutations are required to further define the roles of the JNK and ERK pathways in mediating Ras actions.

We are grateful to Dr. R. J. Davis for kindly providing expression constructs for JNK1 and JNK1(APF) and to Dr. M. Karin for providing expression constructs for GAL4-c-Jun and 5×GLA4-LUC. We also thank Drs. W. S. May, M. P. Carroll, and B. Davis for critical comments on the manuscript and Greg Tyler for technical expertise and administrative assistance.