The MEKK-JNK Pathway Is Stimulated by α1-Adrenergic Receptor and Ras Activation and Is Associated with in Vitroand in Vivo Cardiac Hypertrophy*

In neonatal rat ventricular myocytes, stimulation of the α1-adrenergic receptor (α1-AdrR) activates a program of genetic and morphological changes characterized by transcriptional activation of the atrial natriuretic factor (ANF) gene and enlargement (hypertrophy) of the cells. The low molecular weight GTPase Ras has been established as an important regulator of hypertrophy bothin vitro and in vivo. Ras activates a kinase cascade involving Raf, the mitogen-activated protein kinase kinase (MEK), and the extracellular signal-regulated protein kinase (ERK). However, the extent of involvement of this pathway in regulating hypertrophic responses is controversial. We demonstrate here that both α1-AdrR stimulation and Ras can also activate the c-Jun NH2-terminal kinase (JNK) in cardiomyocytes. The α1-AdrR effect on JNK occurs through a pathway requiring Ras and MEK kinase (MEKK). A constitutively activated mutant of MEKK that preferentially activates JNK, stimulates ANF reporter gene expression, while a dominant negative MEKK mutant inhibits ANF expression induced by PE. Furthermore, JNK activity is increased in the ventricles of mice overexpressing oncogenic Ras, whereas ERK activity is not. These results suggest that the α1-AdrR mediates ANF gene expression through a Ras-MEKK-JNK pathway and that activation of this pathway is associated with in vitro and in vivo hypertrophy.

In neonatal rat ventricular myocytes, stimulation of the ␣ 1 -adrenergic receptor (␣ 1 -AdrR) activates a program of genetic and morphological changes characterized by transcriptional activation of the atrial natriuretic factor (ANF) gene and enlargement (hypertrophy) of the cells. The low molecular weight GTPase Ras has been established as an important regulator of hypertrophy both in vitro and in vivo. Ras activates a kinase cascade involving Raf, the mitogen-activated protein kinase kinase (MEK), and the extracellular signal-regulated protein kinase (ERK). However, the extent of involvement of this pathway in regulating hypertrophic responses is controversial. We demonstrate here that both ␣ 1 -AdrR stimulation and Ras can also activate the c-Jun NH 2 -terminal kinase (JNK) in cardiomyocytes. The ␣ 1 -AdrR effect on JNK occurs through a pathway requiring Ras and MEK kinase (MEKK). A constitutively activated mutant of MEKK that preferentially activates JNK, stimulates ANF reporter gene expression, while a dominant negative MEKK mutant inhibits ANF expression induced by PE. Furthermore, JNK activity is increased in the ventricles of mice overexpressing oncogenic Ras, whereas ERK activity is not. These results suggest that the ␣ 1 -AdrR mediates ANF gene expression through a Ras-MEKK-JNK pathway and that activation of this pathway is associated with in vitro and in vivo hypertrophy.
Stimulation of the G protein-linked ␣ 1 -adrenergic receptor (␣ 1 -AdrR) 1 in neonatal rat ventricular myocytes triggers a hypertrophic response characterized by the transcriptional acti-vation of a number of genes, including that for atrial natriuretic factor (ANF). The hypertrophic response is further characterized by increases in cell size and organization of myofilaments into sarcomeric units (reviewed in Ref. 1). Ras, a low molecular weight GTPase known to transduce mitogenic signals from growth factor-activated tyrosine kinase receptors, has been shown to be required for ␣ 1 -AdrR-mediated ANF gene expression, increases in cell size and myofilament organization (2). Evidence for the involvement of Ras comes from studies in which microinjection of a dominant-interfering Ras expression vector blocked ␣ 1 -AdrR-induced increases in ANF protein and cell size (2). Furthermore expression of oncogenic Ras in transgenic mice results in the development of pathological hypertrophy (3,4). Thus it is well-established that Ras-dependent pathways mediate hypertrophy in vivo as well as in response to ␣ 1 -AdrR stimulation in vitro.
The precise signaling events that ensue subsequent to Ras activation and are important in the development of the hypertrophic phenotype are not clearly understood. Ras activates the Raf-MEK-ERK pathway in a number of cell systems. This pathway has been linked to cell proliferation and growth regulation. Studies from several laboratories have implicated Raf (5), MEK (6), and the ERKs (7)(8)(9) in the regulation of ␣ 1 -AdrRinduced hypertrophic responses. On the other hand, ERK1 and ERK2 are activated by agonists (e.g. carbachol and ATP) that fail to induce ANF expression or myofilament organization (10). In addition, expression of activated MEK, the kinase that activates the ERKs, failed to induce ANF expression and paradoxically inhibited it (11). Thus, it appears that activation of the Raf-MEK-ERK pathway alone is insufficient to transduce the ␣ 1 -AdrR-induced hypertrophic responses.
Ras can also couple to and activate MEKK (12,13). This kinase was originally described as an activator of the MEK-ERK cascade (14) but has more recently been shown to activate JNKK, a dual specificity kinase that phosphorylates and activates JNK, another member of the MAP kinase superfamily (15)(16)(17). JNK, in turn, phosphorylates c-Jun which dimerizes with c-Fos to activate AP-1-dependent gene transcription (18). The role of the MEKK-JNKK-JNK pathway in regulating cardiomyocyte hypertrophy has not been extensively studied. Recent studies in myocytes demonstrate that cellular stressors can activate JNK (19,20). Additionally, transfection of myocytes with activated MEKK and JNKK has been shown to increase cell size and ANF transcriptional activation (21).
We have investigated the effects of ␣ 1 -AdrR stimulation on activation of the MEKK-JNK signaling cascade. Our data demonstrate that PE causes a sustained activation of JNK and that the resultant increase in c-Jun transcriptional activity is dependent on Ras and MEKK. We show that MEKK, rather than Raf, is a mediator of PE-induced ANF gene expression. In-creased JNK activity is also detected in the ventricles of transgenic Ras mice, whereas ERK activities are not elevated above those in control mice. We propose that the maintenance of elevated JNK activity is important in the development of the hypertrophic state in vitro as well as in vivo.

EXPERIMENTAL PROCEDURES
Cell Culture-Neonatal ventricular myocytes were cultured from 1to 3-day-old Harlan Sprague Dawley rats as described previously (22). Trisected hearts were dissociated by treatment with collagenase II (Worthington) and pancreatin (Life Technologies, Inc.), and myocytes were purified by centrifugation through a discontinuous Percoll gradient. Cells were plated at a density of 4 ϫ 10 4 /cm 2 on gelatin-coated tissue culture dishes and maintained overnight in 4:1 Dulbecco's modified Eagle's medium/medium 199 (Life Technologies, Inc.) containing 10% horse serum, 5% fetal calf serum, and antibiotics (100 units/ml penicillin and 100 g/ml streptomycin).
Ras Activity Assay-Myocytes plated on 100-mm dishes were serumstarved for 24 h, washed twice with phosphate-free Dulbecco's modified Eagle's medium, and incubated with phosphate-free Dulbecco's modified Eagle's medium containing 0.5 mCi/ml [ 32 P]orthophosphate for 3 h. Cells were stimulated with 100 M phenylephrine plus 2 M propranolol for various times and lysed in a 0.5% Nonidet P-40 buffer containing 50 mM Tris-HCl (pH 7.5), 150 mM NaCl, 20 mM MgCl 2 , 20 g/ml aprotinin, and 1 mM Na 3 VO 4 . Lysates were cleared of unincorporated [ 32 P]orthophosphate using activated charcoal. Ras was immunoprecipitated with a rat anti-Ras antibody (Y13-259, Santa Cruz Biotechnology, Inc.) in conjunction with a rabbit anti-rat secondary antibody and protein A-Sepharose (Pharmacia Biotech Inc.). Labeled GDP and GTP were eluted from the immunocomplexes with a buffer containing 20 mM Tris-HCl (pH 7.5), 20 mM EDTA, 2% SDS, 0.5 mM GDP, and 0.5 mM GTP and separated by thin layer chromatography (Bakerflex PEI-F cellulose TLC plates). The data were visualized by autoradiography, and the intensity of the spots on the film was analyzed by densitometry.
Transient Transfection Assay-Ventricular myocytes, plated on 60-mm dishes or 6-well plates, were cultured overnight and cotransfected for approximately 18 h with a total of 2.2-3.4 g/ml plasmid DNA using a modified calcium-phosphate technique as described previously (26,27). Myocytes were washed extensively and incubated for an additional 48 h in serum-free medium alone or with PE. Luciferase activity was measured, normalized to protein concentration as determined by Bradford analysis, and used as an index of gene expression.
Kinase Assays-For assaying endogenous kinase activity, cells were plated on 60-mm dishes, serum-starved for 24 h, and then mock-stimulated or stimulated with PE. To examine the effect of activated Ras on JNK activity, myocytes plated on 100-mm dishes were cotransfected with Leu-61 Ras and SR␣-HA-JNK1 and then washed extensively and cultured in maintenance media for 48 h. Cell lysates were then prepared and kinase activities assayed as described previously (28). Briefly, cells were harvested in lysis buffer containing 0.5% Nonidet P-40, and kinases were immunoprecipitated using either mouse monoclonal 12CA5 anti-HA antibodies (Boehringer Mannheim) or rabbit polyclonal anti-JNK1, ERK1, or ERK2 antibodies (Santa Cruz Biotechnology, Inc.) conjugated to protein A-Sepharose. The kinase assays were then carried out at 30°C for 15 min (ERK) or 20 min (JNK) using [␥-32 P]ATP and substrate (myelin basic protein (Sigma) for ERK or GST-c-Jun-(1-79) for JNK). Phosphorylated substrates were analyzed by SDS-polyacrylamide gel electrophoresis and autoradiography. 32 P incorporation was quantitated by radioanalytic scanning (AMBIS).
Western Blot Analysis-Cell lysates (50 g of protein) from above were boiled in Laemmli buffer and resolved on 12% SDS-polyacrylamide gels. The gels were transferred onto a polyvinylidene difluoride Immobilon-P membrane (Millipore) and probed with a rabbit polyclonal anti-JNK1 antibody which is more specific for JNK1 but also recognizes the JNK2 isoform. Membranes were then incubated with horseradish peroxidase-conjugated anti-rabbit secondary antibody (Sigma) and visualized by chemiluminescence detection (Amersham Corp.).

RESULTS
␣ 1 -AdrR Stimulation Activates Ras-It is well-established that Ras mediates PE-induced ANF expression (2). To directly establish that stimulation of the ␣ 1 -AdrR causes activation of Ras, myocytes were metabolically labeled with [ 32 P]orthophosphate and treated with or without PE for short times. Ras was then isolated and the proportion of GTP-bound Ras determined. The amount of Ras in its GTP-bound (activated) state increased by 35% after 1 min of agonist stimulation (Fig. 1). This increase, although small, was reproducible and was sustained for at least 10 min.
␣ 1 -AdrR Stimulation and Oncogenic Ras Activate JNK-Multiple signaling pathways can be activated downstream of Ras. These include the Raf-MEK-ERK cascade and the more recently characterized MEKK-JNKK-JNK cascade. We have previously shown that PE stimulates ERK activity in myocytes (10). To determine if ␣ 1 -AdrR stimulation also activates JNK, we examined the effects of PE on JNK activity. Myocytes were treated in the absence or presence of PE for various times, and JNK was immunoprecipitated and assayed for its ability to phosphorylate its substrate GST-c-Jun. Induction of JNK activity was observed following 20 min of PE stimulation, and its activation level was sustained through 48 h (Fig. 2, A and B). Immunoblotting using an anti-JNK antibody demonstrated that the observed increase in JNK activity was not associated with an increase in JNK protein levels even at long times of PE stimulation (Fig. 2C). These data indicate that PE activates existing JNK rather than increasing its expression. To determine if oncogenic Ras also activates JNK, constitutively acti- vated (Leu-61) Ras was cotransfected into myocytes along with HA-tagged JNK1. As shown in Fig. 3, Leu-61 Ras caused a significant activation of HA-JNK1, increasing its activity 8-fold above basal. Leu-61 Ras was also an effective activator of HA-ERK2, stimulating its activity 20-fold above control (data not shown). These data clearly demonstrate that JNK can be activated by both PE and Ras and prompted our further investigation concerning the pathway for JNK activation.

PE-induced Stimulation of c-Jun Transcriptional Activity Is Dependent on Ras and MEKK-Since PE and Leu-61 Ras both
increased JNK activity, we sought to establish whether PE activated JNK through its effect on Ras. To address this question, we employed an assay that uses c-Jun transcriptional activity as a readout for JNK activation in transiently transfected cells. c-Jun transcriptional activity is stimulated following its phosphorylation by JNK at Ser-63 and Ser-73 (29,30). Thus the ␣ 1 -AdrR-induced increases in JNK activity should result in induction of c-Jun transcriptional activity. Phenylephrine stimulated the expression of a GAL4-luciferase reporter gene that was responsive to a GAL4-c-Jun-(1-223) fusion protein containing the c-Jun transactivation domain. This stimulation was completely inhibited by coexpression of dominant negative (Asn-17) Ras (Table I) (Table I). In parallel experiments, we found that coexpression of dominant negative Raf paradoxically elevated the basal GAL4-luciferase activity; nonetheless, there was little decrease in the fold stimulation of GAL4-luciferase by PE (data not shown). These data indicate that PE stimulates JNK activity, leading to the phosphorylation and transcriptional activation of c-Jun, and that these events are dependent on Ras and MEKK but not Raf.
Both Activated MEKK and Activated Raf Stimulate ANF Reporter Gene Expression-MEKK was originally identified as an alternative to Raf in activating MEK, leading to the activation of the ERKs (14). However, it is now clear that Raf is a more specific activator of the ERK pathway and MEKK is a more specific activator of the JNK pathway (16,17,25). To determine the effects of activated MEKK and activated Raf on ANF expression in cardiomyocytes, we coexpressed activated MEKK or activated Raf with the ANF-luciferase reporter gene. As shown in Table II  interfering Raf-1 or dominant interfering MEKK along with the ANF reporter gene and subsequently stimulated with PE. Dominant interfering Raf-1 failed to block PE-mediated ANFluciferase expression, although it was able to inhibit the response to activated (Val-12) Ras (Fig. 4, A and B). In contrast, dominant negative MEKK significantly inhibited PE-and Rasinduced ANF reporter gene expression (Fig. 4, A and B). These data suggest that PE mediates its effects on ANF gene expression, in addition to its effects on JNK (Table I), through MEKK rather than Raf.

JNK but Not ERK Is Activated in Left Ventricles of Ras Transgenic
Mice-It has recently been shown that ventricular expression of Val-12 Ras in mice induces cardiac hypertrophy characterized by increased ratios of left ventricular weight to body weight and increased ventricular ANF expression (3). The left ventricles of 6-week-old transgenic Ras and wild-type (control) mice were rapidly isolated, frozen, homogenized, and subjected to kinase assays to determine JNK, ERK1, and ERK2 activities. Tissue from six Ras animals and four control animals were analyzed. JNK activity was consistently higher in the tissue from Ras animals; however, no difference was observed in either ERK1 or ERK2 activity (Fig. 5). These data demonstrate that constitutive activation of Ras in vivo correlates with increased JNK, but not ERK, activity.

DISCUSSION
Hypertrophic agents such as the ␣ 1 -AdrR agonist PE activate a program of genetic and morphological changes in terminally differentiated cardiac myocytes. The PE-induced hypertrophic responses have been shown to be Ras-dependent (2). Consistent with this, we find that PE increases the amount of GTP-bound Ras and hence the activity of Ras. These data confirm findings reported by Thorburn's laboratory (31) and raised the question as to the effector(s) of Ras involved in regulating hypertrophic responses.
Stimulation of cardiomyocytes with PE activates the ERKs (32,33), and several studies have implicated the Raf-MEK-ERK pathway in mediating PE-induced gene expression (5)(6)(7)(8)(9). However, the extent of involvement of this signaling pathway remains controversial (10,11). Our current data demonstrate that while constitutively activated Raf can induce ANF expression, a dominant negative mutant displays no inhibitory effect on PE-induced ANF expression. As a result, activation of an additional signaling pathway would appear to be necessary for the transduction of PE-evoked signals leading to hypertrophy.
The JNK family of MAP kinases can be activated via heterotrimeric G proteins (34 -38), as well as low molecular weight GTPases including Ras (24,25,39,40). While JNK activation has previously been associated with stress responses such as apoptosis (41)(42)(43), it is now evident that it does not simply regulate these pathological states. In cardiac myocytes, we demonstrate that ␣ 1 -AdrR stimulation increases JNK activity within 20 min of PE treatment and that JNK activity remains elevated for at least 48 h. This is in contrast to the ERKs, which have been shown to be maximally activated at 5 min, declining to basal levels thereafter (8,10). That the kinetics of JNK  activation is slower than that of the ERKs is in agreement with other published findings (19,34). The differences in activation kinetics may be explained by recent data that demonstrate that JNK activation induces MAP kinase phosphatase-1 activity, correlating with down-regulation of ERK activity (44). The maintenance of JNK activation may be important in the progression of cardiomyocyte hypertrophy, since the accompanying genetic and morphological changes are not triggered immediately but rather take hours to develop.
Our findings indicate that MEKK is a critical intermediate in PE-induced JNK activation and ANF expression. We have shown that activation of the GAL4-c-Jun hybrid transcription factor by PE, presumably via JNK-induced phosphorylation, is dependent on both Ras and MEKK. In examining the role of MEKK in cardiac gene expression, we found that constitutively activated MEKK increased ANF expression, and a dominant negative mutant of MEKK significantly inhibited PE-induced ANF reporter expression. Our results indicate a role for the MEKK-JNK signaling pathway in regulating ␣ 1 -AdrR-induced ANF expression and in serving as a downstream mediator of signals generated through Ras in cardiomyocytes.
In vivo studies using transgenic Ras mice support this conclusion. These data show that JNK, but not ERK, activity is elevated, relative to control, in the hypertrophic left ventricles of these mice. Although kinase activity was measured in homogenates of whole left ventricles, the myosin light chain-2v promoter driving expression of the Val-12 ras transgene is specific for ventricular myocytes; thus, we believe that the observed kinase activity changes reflect those that occur in myocytes. Since the kinetics of ERK and JNK activation differ in the in vitro myocyte culture system, their activation kinetics might also differ in vivo; thus we cannot rule out that ERK activity is increased transiently in the course of development of hypertrophy, whereas JNK activity remains elevated. Nonetheless, our data indicate that constitutive activation of Ras in the mouse heart ventricle leads either directly or indirectly to stimulation of cardiomyocyte JNK activity.
That JNK activation and hypertrophy are related events is further supported by studies using transverse aortic constriction to induce cardiac pressure overload in mice. In this mouse model, left ventricular hypertrophy is evident at 7 days postbanding, and JNK activity is also observed to be significantly increased. 2 The extent to which JNK activation is the cause or effect of the associated hypertrophy remains to be determined. If elevation of JNK activity is sufficient to induce hypertrophy, then we expect that transgenic mice expressing activated MEKK would also exhibit a hypertrophic phenotype and that hypertrophy induced by pressure overload will be blunted in JNK knockout mice. We are currently investigating these possibilities.