p38 MAPK and NF-kB Collaborate to Induce Interleukin-6 Gene Expression and Release EVIDENCE FOR A CYTOPROTECTIVE AUTOCRINE SIGNALING PATHWAY IN A CARDIAC MYOCYTE MODEL SYSTEM*

In cardiac myocytes, the stimulation of p38 MAPK by the MAPKK, MKK6, activates the transcription factor, NF-kB, and protects cells from apoptosis. In the present study in primary neonatal rat cardiac myocytes, constitutively active MKK6, MKK6(Glu), bound to IkB kinase (IKK)-b and stimulated its abilities to phosphorylate IkB and to activate NF-kB. MKK6(Glu) induced NF-kB-dependent interleukin (IL)-6 transcription and IL-6 release in a p38-dependent manner. IL-6 protected myocardial cells against apoptosis. Like IL-6, TNF-a, which activates both NF-kB and p38, also induced p38-dependent IL-6 expression and release and protected myocytes from apoptotis. While TNF-a was relatively ineffective, IL-6 activated myocardial cell STAT3 by about 8-fold, indicating a probable role for this transcription factor in IL-6-mediated protection from apoptosis. TNF-a-mediated IL-6 induction was inhibited by a kinase-inactive form of the MAPKKK, TGF-b activated protein kinase (Tak1), which is known to activate p38 and NF-kB in other cell types. Thus, by stimulating both p38 and NFkB, Tak1-activating cytokines, like TNF-a, can induce IL-6 expression and release. Moreover, the myocyte-derived IL-6 may then function in an autocrine and/or paracrine fashion to augment myocardial cell survival during stresses that activate p38.

Cytokines derived from either infiltrating cells, such as macrophages, or sometimes from cells comprising the tissue itself, play important roles in the wound healing process (1). Since some injuries, such as myocardial infarction, are frequently life-threatening, a better understanding of the roles of cytokines in tissues such as the heart is critical. Following a myocardial infarction there is an accumulation of tumor necrosis factor (TNF␣), 1 interleukin (IL)-1␤, and IL-6 at or near the affected region (2)(3)(4)(5). Previously, production of these cytokines in the heart was attributed to infiltrating macrophages or leukocytes and endothelial cells (e.g. see Ref. 6). However, recent findings have demonstrated the expression of all three of these cytokines in cardiac myocytes following ischemic stress (7)(8)(9). Moreover, certain cytokines are thought to promote cardiac tissue recovery after a brief ischemic insult (4, 9 -11). Accordingly, there is renewed interest in studying the possible beneficial effects of cytokines in the heart as well as elucidating the signal transduction pathways and mechanisms responsible for their induction.
The induction of most cytokine genes requires activation of the transcription factor, nuclear factor (NF)-B (12)(13)(14)(15). NF-B is activated and cytokine expression is increased in rat hearts submitted to ischemic stress; interestingly, these events appear to require the mitogen-activated protein kinase (MAPK), p38 (16). Additionally, in cultured cardiac myocytes, NF-B-dependent reporter gene expression is activated following the selective stimulation of p38 by its upstream activator, MKK6 (17). Moreover, this p38-dependent NF-B activation contributes to protecting cultured myocardial cells from undergoing apoptosis (17). However, while p38 and NF-B are both important for the maintenance of heart function following stress, neither the mechanism by which p38 activates NF-B nor the role of p38-activated NF-B in myocardial cell survival are well understood. Accordingly, the present study was undertaken to begin addressing these questions.
Many signaling pathways can interact with each other through biochemical cross-talk; however, such interactions between p38 and NF-B are not well understood. Like the other MAPK family members, p38 is part of a cascade of kinases. One of the best studied activators of p38 is the MAPKK, MKK6, which lies directly upstream of p38; among ERK, JNK, and p38, MKK6 activates only p38 (18,20). Although MKK6 itself can be activated by several upstream kinases, a particularly interesting finding is that in some cells, MKK6 can be activated by the MAPKKK, TGF-␤-activated protein kinase (Tak1) (18,20). Tak1 was originally named for its ability to be activated by TGF-␤. However, Tak1 can also be activated by cytokines, such as interleukin-1 or TNF-␣ (20 -22) (see Fig. 7 for reference).
In comparison with the p38 pathway, the NF-B pathway is activated through a series of events that are also mediated by a cascade of kinases, many of which are believed to be unique to that pathway. NF-B, which is comprised of two subunits (p65 and p50), is retained in the cytoplasm by virtue of its interaction with the inhibitor of B (IB). IB is phosphorylated in response to NF-B-activating signals; this phosphorylation leads to the ubiquitination and subsequent degradation of IB. This then allows NF-B to translocate to the nucleus, where it binds to critical elements in cytokine genes and increases their transcription (23,24). The kinases responsible for the phosphorylation of IB belong to the IB kinase family, or the IKKs (25)(26)(27)(28)(29). NF-B-inducing kinase phosphorylates and activates IKK (26,30,31). Although it preferentially activates IKK, NF-B-inducing kinase, which bears strong sequence homology to, and is now considered a member of, the MAPKKK family, can also activate the JNK MAPK pathway (32). Interestingly, in addition to being activated by NF-B-inducing kinase, IKK can also be activated by MEKK1 (27), a MAPKKK that is also known to activate JNK in some cell types (33) and by Tak1 (refer to Fig. 5). Thus, at least at the MAPKKK level, there exists the potential for cross-talk between the NF-B and MAPK pathways.
In the present study, we examined the nature of the crosstalk between the NF-B and p38 pathways. Because of the clear importance of these pathways in promoting the recovery, or maintenance of, heart tissue function following stress or injury, we used cardiac myocytes as the model system. To this end, we explored whether NF-B and p38 collaboratively activate the IL-6 gene in cardiac myocytes and, further, whether IL-6 itself has roles that are consistent with functional recovery of the tissue following stress. We have found the following: 1) agonists, such as TNF-␣, can activate IL-6 transcription in cardiac myocytes in a manner that requires both NF-B and p38; 2) the coordinate activation of the NF-B and p38 pathways by TNF-␣ takes place largely through the MAPKKK, Tak1; 3) the p38 pathway can influence NF-B activation, at least partly, through the physical association of MKK6 and IKK␤; and 4) IL-6 can protect cardiac myocytes from undergoing apoptosis induced by sphingosine, a signaling lipid known to increase in the heart following ischemic stress.

Cell Culture
Primary ventricular myocytes were prepared from 1-4 day old Harlan Sprague-Dawley rats as described (34,35). Briefly, hearts were dissected in DMEM/air; the apical two-thirds of the ventricles were dissected away from the atria. After mincing and washing the ventricles twice with air-compatible DMEM, cells were isolated by multiple rounds of 10-min-long tissue dissociation with 0.001% trypsin. After each incubation with trypsin, the supernatant was added to an equal volume of DMEM/F-12 (1:1) containing 20% fetal bovine serum, and all of the supernatants were combined. Plastic wells were treated for at least 1 h with 5 ng of fibronectin/ml of DMEM/F-12 (1:1). Myocytes were plated in DMEM/F-12 (1:1) containing 10% fetal bovine serum for approximately 16 h. After washing with DMEM/F-12 (1:1), the cultures were incubated in DMEM/F-12 (1:1) with or without any test agents for the indicated times.

Transfection by Electroporation
Immediately following the final dissociation step, myocardial cells were resuspended in serum-free DMEM/F-12 (1:1). Between 5 and 12 ϫ 10 6 cells were combined with the combinations of plasmids indicated in the figure legends in a total volume of 300 l of DMEM/F-12 (1:1). The total amount of plasmids used for each electroporation was equalized using pCMV6. Dose-response experiments were carried out to determine optimal quantities of plasmids for transfection and to verify that the results obtained were consistent over a range of plasmid levels. Generally, the dose-response experiments led to using the following quantities of each plasmid type in each electroporation: 1 g each of pCMV6 (control) or p38␤ 2 ; 15 g of MKK6(Glu), Tak1, or Tab1; 20 g each of luciferase and ␤-galactosidase reporters; 45 g of IKK␤-M, MKK6-M1, MKK6-M2, and Tak1-M. Cells were electroporated at 500 V, 25 microfarads, and 100 ohms in a 0.2-cm gap electroporation cuvette (Bio-Rad) using a Gene Pulser II (Bio-Rad). Under these conditions, only cardiac myocytes are transfected (34,36). For reporter assays, 1.5 ϫ 10 6 cells were plated per 24-mm well, whereas 4.5 ϫ 10 6 myocytes per 35-mm well were plated to perform kinase assays. Plasmids MKK6(Glu), MKK6-M1, and MKK6-M2-pcDNA3 FLAG-MKK6 (Glu) and pcDNA3 FLAG-MKK6(K82A), the latter of which we call MKK6-M1, code for activated and kinase-inactive human MKK6, respectively (19) and were obtained from R. J. Davis (University of Massachusetts, Worcester, MA). In some cases, instead of MKK6-M1 we used MKK6-M2 (FLAG-MKK6(K82A; S207A; T211A)), which is a kinase-inactive form of MKK6 that also has the sites normally phosphorylated by an upstream MAPKKK, Ser 207 and Thr 211 , mutated to Ala. MKK6-M2 produced similar inhibitory effects on the p38 pathway as MKK6-M1.
NF-B-luc-p2X NF-B, which codes for a luciferase reporter driven by a minimal prolactin promoter with two nearby, upstream NF-B consensus sites, was obtained from M. Karin (39) (University of California at San Diego, La Jolla, CA).

Preparation of Recombinant Adenovirus
AdV-p38 and AdV-MKK6(Glu)-The AdEasy system was used for preparing recombinant adenoviral strains using previously described methods (40). Briefly, human MKK6(Glu) and human p38␤ 2 were polymerase chain reaction-amplified from the parent templates (see above) such as to create restriction sites that would facilitate cloning into pAdTrack-CMV, an adenoviral shuttle vector that harbors CMVdriven green fluorescent protein (GFP), and a CMV-flanked multiple cloning site for the insertion of the gene of interest. Polymerase chain reaction-amplified p38␤ 2 and MKK6(Glu) were cloned into the EcoRI and NotI sites of pGEX-6P-1 (Amersham Pharmacia Biotech), which served as a shuttle cloning vector. pGEX-6P-1/p38␤ 2 and pGEX-6P-1/ MKK6(Glu) were then digested with BamHI and NotI, and the resulting products of interest were cloned into the BglII and NotI sites of pAdTrack-CMV to create pAdTrack-CMV-p38␤ 2 and pAdTrack-CMV-MKK6(Glu). pAdTrack-CMV-p38␤ 2 or pAdTrack-CMV-MKK6(Glu) was linearized and then co-transformed with the adenoviral vector, pAdEasy-1, into Escherichia coli strain BM5183. This strain of E. coli allows for homologous recombination of pAdEasy-1 and the pAdTrack-CMV shuttle vector containing the gene of interest. Recombinants were selected on kanamycin and screened by restriction digestion with PacI. Recombinant plasmids were then retransformed into E. coli DH5␣ for propagation purposes. Recombinant adenoviral plasmids were linearized with PacI and then transfected into 293 human embryonic kidney cells, using LipofectAMINE® (Life Technologies, Inc.). Transfection efficiency was determined by observing GFP fluorescence, as described previously (40). The recombinant viruses were then harvested 7-10 days postinfection. Viral titers were determined by observing GFP fluorescence of primary neonatal cardiac myocytes; the minimum quantity of viral stock that afforded 100% transfection efficiency was selected for the experiments in this study.

Reporter Assays
␤-Galactosidase-Following the appropriate time in culture, each well of myocytes was washed twice with phosphate-buffered saline and then lysed in 500 l of ice-cold lysis buffer (25 mM Gly-Gly, pH 7.8, 15 mM MgSO 4 , 4 mM EDTA, 0.25% Triton X-100) containing 1 mM dithiothreitol. The cell debris was removed by centrifugation. To measure ␤-galactosidase activity, 200 l of the cell extract was added to 400 l of ␤-galactosidase buffer (60 mM Na 2 HPO 4 , 40 mM NaH 2 PO 4 , 10 mM KCl, 1 mM MgSO 4 ) containing 1 mg/ml chlorophenol red-␤-D-galactopyranoside and 50 mM ␤-mercaptoethanol. The reaction was incubated for 2 h at 37°C. After stopping each reaction by adding 100 l of 1 M Na 2 CO 3 , the absorbance was measured at 570 nm.
Luciferase-To measure luciferase activity, 100 l of buffer (25 mM Gly-Gly, pH 7.8, 15 mM MgSO 4 , 4 mM EGTA, 45 mM KPO 4 , pH 7.8, 1 mM DTT, 0.3 mM D-luciferin, 3 mM ATP) was added to 100 l of cell lysate. Light emission of each sample was measured by a BioOrbit 1251 luminometer for 30 s. The relative luciferase activities were determined by dividing the relative luciferase activity by the relative ␤-galactosidase activity.
Cytokine ELISAs IL-6 -To measure IL-6 secretion, IL-6 ELISAs were carried out using a kit according to the manufacturer's protocol (BIOSOURCE International). After the myocardial cells were preplated as described (17) to eliminate noncardiac myocytes, the myocytes were plated at a density of 1300 cells/mm 2 in 10% fetal calf serum for 24 h. Cells were washed and cultured in media with or without recombinant TNF-␣ (Genzyme), with or without 5 M SB203580, or the cells were infected with AdV-MKK6(Glu), AdV-p38␤ 2 , or AdV-Control. Following 48 h of incubation, media were collected and assayed using the IL-6 ELISA kit.

Apoptosis
TUNEL-Myocardial cells were cultured with or without recombinant IL-6 (1 ng/ml) (BIOSOURCE International) supplemented DMEM/F-12 with or without 5 M SB203580 for 48 h prior to the addition of 10 M sphingosine (Calbiochem) for 6 h. TUNEL analyses of fragmented DNA were performed on cultured cardiac myocytes as described previously (17) and according to the manufacturer's protocol (Roche Molecular Biochemicals). Cells were scored for TUNEL-positive nuclei in a researcher-blinded manner.
DNA Laddering-Assessment of apoptosis was performed by DNA laddering essentially as described previously (17). Approximately 10 6 cardiac myocytes were lysed in a digestion buffer containing Proteinase K, scraped, pipetted several times, and incubated at 55°C for 5-14 h. Samples were then extracted with phenol/chloroform/isoamyl alcohol, and then DNA was isopropyl alcohol-precipitated and washed several times with ethanol and allowed to air-dry. After dissolving the DNA and incubating with RNase A, DNA fragments were fractionated on an agarose gel.
Immunocytofluorescence p65 Immunocytofluorescence-To study the cellular localization of the p65 subunit of NF-B, myocardial cells were co-transfected with test expression constructs (1 g each of pCMV6, p38␤ 2 or 15 g of MKK6(Glu)) and 5 g of a plasmid encoding GFP. Following 48 h of culture in minimal media, cells were fixed as described (35), and immunocytofluorescence was carried out using a polyclonal antibody raised against p65/NF-B (Santa Cruz Biotechnology, Inc., Santa Cruz, CA) followed by Texas Red-conjugated anti-rabbit IgG (Molecular Probes, Inc., Eugene, OR). Transfected cardiac myocytes were identified as GFP-positive cells, and p65 localization was visualized in only the transfected cells.
IKK␤ Assay-IKK␤ kinase activity was assessed in myocardial cells by co-transfecting C-FLAG-IKK␤ with or without test expression constructs with or without SB203580 (5 M). After the appropriate times, cultures were extracted in a buffer containing 20 mM Tris, pH 7.6, 20 mM ␤-glycerophosphate, 250 mM NaCl, 3 mM EGTA, 3 mM EDTA, 0.5% Nonidet P-40, 0.1 mM sodium orthovanadate, 10 g/ml aprotinin, 2 mM DTT, 1 mM PMSF, 1 mM p-nitrophenyl phosphate, and 10 g/ml of leupeptin. After brief centrifugation, extracts were incubated for 2 h at 4°C with anti-IKK␤ antibody (Santa Cruz Biotechnology), followed by protein G-Sepharose (Amersham Pharmacia Biotech) precipitation. Immunocomplex kinase assays were carried out using 1 g of recombinant IB-␣-(1-317) (Santa Cruz Biotechnology) per sample and 10 M [␥-32 P]ATP (5000 Ci/mmol) in a final volume of 30 l of kinase buffer (30 mM HEPES, pH 7.4, 10 mM MgCl 2, 1 mM DTT) at 25°C for 15 min. The reactions were terminated by the addition of Laemmli sample buffer, and the phosphorylation level of IB-␣ was evaluated by SDS-PAGE followed by autoradiography and PhosphorImager analyses. The quantity of IKK␤ in each sample was determined by Western analysis; the observed levels were used to normalize relative IB-␣ phosphorylation levels found in PhosphorImager analyses.
STAT3 and P-STAT3-Myocytes were cultured as described above and then treated with or without various test agents for 15 min. The cells from 3-35-mm culture wells (4.5 ϫ 10 6 myocytes) were collected by scraping into 450 l of lysis buffer, which consisted of 20 mM Tris-HCl (pH 7.6), 20 M sodium glycerophosphate, 250 mM NaCl, 3 mM EGTA, 3 mM EDTA, 0.5% Nonidet P-40, 0.25 mM PMSF, 1 g/ml aprotinin, 1 g/ml leupeptin, 1 g/ml pepstatin, 50 mM NaF, and 2 mM Na 3 VO 4 . Extracts were centrifuged for 15 s at 14,000 rpm in a microcentrifuge, and the supernatants were precleared by incubating with protein A-Sepharose beads for 30 -60 min at 4°C. Extracts were then incubated for 12-18 h at 4°C with anti-STAT3 antibody (New England Biolabs; kit no. 9130) at 1:100. Immunoprecipitates were collected on protein A-Sepharose beads, eluted with Laemmli buffer, and analyzed by SDS-PAGE followed by blotting on nitrocellular paper, as described above. Blots were then probed with either anti-STAT3 antibody or with anti-Y-705-STAT3 antibody and washed according to the manufacturer's protocol (New England Biolabs, Inc., Beverly, MA). Upon visualizing using chemiluminescence, blots were digitized using a PhosphorImager, and the bands representing Y-705-STAT3 and total STAT3 were quantitated using ImageQuant software (Molecular Dynamics).
IKK␤/MKK6 Association-Myocytes were cultured for 72 h in DMEM/F-12 containing 1% fetal bovine serum. 9 ϫ 10 6 myocytes were lysed in 500 ml of extraction buffer containing 20 mM Tris, pH 7.6, 20 mM ␤-glycerophosphate, 250 mM NaCl, 3 mM EGTA, 3 mM EDTA, 0.5% Nonidet P-40, 0.1 mM sodium orthovanadate, 10 g/ml aprotinin, 2 mM DTT, 1 mM PMSF, 1 mM p-nitrophenyl phosphate, and 10 g/ml leupeptin. The cell debris was removed by brief centrifugation. Lysates were incubated at 4°C with anti-IKK␤ polyclonal antibody (Santa Cruz Biotechnology) or with anti-MEK-6 polyclonal antibody (Stressgen) at a dilution of 1:250 for 2 h. Thirty l of a protein G-Sepharose slurry was then added to each sample and incubated for 30 min at 4°C. Following centrifugation, Laemmli buffer was added to the pellet, which was then boiled for 5 min. Proteins were resolved by a 12% SDS-PAGE and transferred to a nitrocellulose membrane in methanol transfer buffer at 100 V for 2 h. Blots were blocked for 30 min in 5% nonfat milk dissolved in TBS-Tween (0.01%) at room temperature and then probed with anti-FLAG M2 monoclonal antibody (Sigma) at a dilution of 1:300 in 5% nonfat milk for 2 h at room temperature. After washing for 1 h in TBS-Tween (0.01%), the membrane was incubated in 1:2000 dilution of goat anti-mouse IgG horseradish peroxidase (Jackson ImmunoResearch Laboratories). Visualization of immune complex was carried out by an enhanced chemiluminescence (ECL) method using ECL Western blotting detection reagents (NEN Life Science Products) according to the manufacturer's instructions.

MKK6 and p38 Induce NF-B Translocation and Transcrip-
tion-Previous studies have demonstrated that constitutively active MKK6 (MKK6(Glu)) stimulates myocardial cell p38 but not ERK or JNK MAPK (35,41). MKK6(Glu) was also shown to protect cardiac myocytes from apoptosis (17). Since NF-B is known to protect several other cell types from apoptosis (e.g. see Ref. 12), we explored whether p38 could activate NF-B in primary cardiac myocytes. Accordingly, myocardial cell cultures were transfected with constructs that code for the expression of wild type p38␤ 2 2 and/or MKK6(Glu). NF-B activation was estimated qualitatively by visualizing the extent of nuclear translocation of one of the NF-B subunits, p65/NF-B, by immunocytofluorescence and quantitatively by examining the activation of NF-B-dependent transcription of a co-transfected reporter gene. In cells transfected with a control plasmid, p65/ NF-B cross-reactivity was distributed in a punctate pattern 2 We found that when co-transfected with MKK6(Glu), p38␣ induced apoptosis in cultured cardiac myocytes, while p38␤ 2 (same as p38-2) did not. This is consistent with a previous study indicating that p38␣ is proapoptotic in cardiac myocytes (41). Accordingly, all experiments in this study that required the overexpression of p38 employed only constructs encoding p38␤ 2 .

FIG. 1. Effects of p38 and MKK6(Glu) on NF-B translocation and NF-B-mediated transcription.
A-D, myocardial cells were transfected with 1 g of plasmids encoding either no protein (pCMV6 (Control; 1 g)) 3 or p38␤ 2 (1 g) and/or MKK6(Glu) (15 g), as shown. Cultures were plated onto glass slides and incubated for 48 h in serumfree media, after which they were fixed and stained for p65/NF-B and visualized using a Texas Red-conjugated second antibody. Shown are single cells that are representative of the population of cells observed following the treatments indicated. E, myocardial cells were co-transfected with p2X NF-B (20 g) 3 and pCH110 (␤-gal; 20 g) with or without plasmids encoding p38␤ 2 (1 g) and with MKK6(Glu) (15 g) and treated with or without SB203580 (5 M), as indicated. After 48 h in serum-free medium, culture extracts were assayed for luciferase and ␤-galactosidase activities. Relative luciferase (Rel Luc, luciferase/␤-gal) was normalized to the maximum relative luciferase value in each experiment and then displayed as the percentage of that value. Each bar represents the mean of the results obtained from three identically treated cultures with or without S.E. Each experiment was replicated at least three times; the results shown are from one representative experiment. mostly in the cytoplasm of cardiac myocytes (Fig. 1A). 3 In cells transfected with a plasmid encoding wild type p38␤ 2 only, there was a small increase in cross-reactive p65/NF-B in the nucleus (Fig. 1B), compared with control cells. However, in cells transfected with a plasmid encoding MKK6(Glu), there were clear increases in nuclear p65/NF-B (Fig. 1C), which was even more pronounced upon co-transfection with plasmids encoding p38␤ 2 and MKK6(Glu) (Fig. 1D). Consistent with the nuclear accumulation of p65/NF-B were the abilities of overexpressed p38␤ 2 and/or MKK6(Glu) to increase luciferase expression from an NF-B-dependent luciferase reporter plasmid (Fig. 1E). Overexpression of either p38␤ 2 or MKK6(Glu) resulted in approximately 2-and 3-fold induction of NF-B/luciferase, respectively. However, co-expression of p38␤ 2 and MKK6(Glu) conferred a robust, 16-fold induction of reporter expression. The p38-specific inhibitor, 4 SB203580, diminished this induction by about 70%. Thus, upon overexpression of wild type p38␤ 2 and/or MKK6(Glu), the extent of NFB-dependent luciferase induction correlated well with the nuclear accumulation of p65/NF-B. These results are consistent with the hypothesis that in cardiac myocytes, activation of the p38 pathway can lead to the nuclear translocation of NF-B and to the 3 Dose-response experiments were carried out with all plasmids to verify that the effects shown are representative of those obtained using a variety of plasmid levels and to determine optimal plasmid doses. 4 At the levels used in this study (5 M), SB203580 has been shown to be a selective inhibitor of p38. The effectiveness of 5 M SB203580 as a p38 inhibitor is also demonstrated in Fig. 3C. Duplicate myocardial cultures were transfected with IKK␤ and the combinations of plasmids encoding p38␤ 2 and/or MKK6(Glu) with or without SB203580, as shown. After 72 h, extracts were submitted to immunoprecipitation using a IKK␤-specific antibody. IKK␤ kinase assays were then carried out using IB-␣ as the substrate (see "Materials and Methods"). The phosphorylation of IB-␣ was assessed by SDS-PAGE followed by autoradiography. The relative levels of IB-␣ phosphorylation following each treatment were assessed using Molecular Dynamics ImageQuant software. Shown above each autoradiogram is the -fold stimulation normalized to total IKK␤ protein levels as determined by subsequent Western analysis of each kinase assay. B, IKK␤ MKK6 association. Myocardial cells were transfected with plasmids encoding IKK␤ and/or MKK6(Glu). After 72 h, culture extracts were submitted to immunoprecipitation using IKK␤-or MKK6-specific antibody. Immunoprecipitates were then submitted to SDS-PAGE and electroelution onto a nitrocellulose membrane. Western blotting was then carried out using a FLAG-specific antibody. This experiment was repeated at least three times; the results shown are from one representative experiment.
induction of NF-B-sensitive genes.
MKK6 and p38 Induce NF-B-dependent IL-6 Transcription-Since p38 is activated during the stress response in many cells types and since NF-B is known to facilitate cytokine gene induction during cell stress, we explored the hypothesis that, together, p38 and NF-B could mediate cytokine induction in cardiac myocytes. Accordingly, myocardial cells were transfected with a construct that possesses 1186 nt of the IL-6 gene-regulatory region and promoter-driving luciferase, wild type IL-6 (IL-6 (wt)) ( Fig. 2A). Overexpressing wild type p38␤ 2 or MKK6(Glu) resulted in approximately 2-or 3-fold reporter induction, respectively (Fig. 2B). However, co-expression of p38␤ 2 and MKK6(Glu) together conferred an approximately 7-fold induction of reporter expression (Fig. 2B). These results indicate that the IL-6 promoter is transcriptionally active in cardiac myocytes 5 and that myocardial cell IL-6 transcription can be stimulated by p38␤ 2 and MKK6. IL-6 transcriptional induction in response to p38␤ 2 and/or MKK6(Glu) was decreased by 70 -100% upon the addition of SB203580 (Fig. 2B). Notably, IL-6 transcriptional induction was completely dependent upon NF-B, since a point mutation that abolishes the only NF-B binding site in the IL-6 gene (IL-6-M in Fig. 2A; Ref. 15) resulted in a nearly complete loss of p38␤ 2 -and/or MKK6(Glu)inducible IL-6 reporter activity. Interestingly, the abilities of p38 and/or MKK6(Glu) to induce the IL-6 promoter were also inhibited by overexpression of a kinase-inactive, dominantinterfering form of IKK␤, IKK␤-M (Fig. 2B). Taken together, these results indicate that IL-6 transcriptional induction in cardiac myocytes depends on the activation of NF-B and is increased upon stimulation of the p38 pathway in a manner that is dependent upon IKK␤.
TNF-␣ stimulates p38 in other cell types (42,43). Consistent with those results, we found that in cardiac myocytes, among the MAPKs, TNF-␣ preferentially stimulated p38 by about 6-fold, compared with approximately 2-fold for either JNK or ERK (Fig. 3, A and B). Accordingly, we evaluated whether TNF-␣ could augment NF-B-dependent transcription and, if so, whether this enhancement was sensitive to the inhibition of p38. For these experiments, we employed a reporter gene composed of the minimal IL-6 promoter flanked only by NF-B binding sites (NF-B/IL-6 in Fig. 2A). Indeed, TNF-␣ conferred a robust, 10-fold induction of NF-B-dependent IL-6 transcription in cardiac myocytes (Fig. 2C), which was diminished by about 30 -40% by SB203580 or by a dominant-negative, kinase-inactive MKK6 (MKK6-M1), indicating a partial require- 5 The electroporation paramaters used in this study have been shown to allow for the transfection of plasmid DNA into cardiac myocytes only and not any other cell types that may be present in the cultures (34,36). Accordingly, any reporter expression observed in culture extracts is from cardiac myocytes only.  High). C, effects of adenoviral MKK6(Glu) on IL-6 secretion. Myocardial cells were infected with AdV-(Con), AdV-p38␤ 2 , and/or AdV-MKK6(Glu) such that 100% of the cells were transfected, and they were treated with or without SB203580, as shown and as described for B and under "Materials and Methods." After 48 h, medium samples were assayed for the presence of IL-6 by ELISA. Each value represents the mean of three identically cultures Ϯ S.E. ment for p38. TNF-␣-mediated induction of IL-6 transcription was not inhibited by treating cells with 5 M of the MEK inhibitor, PD098059, or by co-transfecting cells with a dominant-negative form of MEK (not shown), indicating that ERK is probably not involved. Transfection with IKK␤-M resulted in an approximately 40 -50% reduction in TNF-␣-mediated induction of IL-6 transcription, and SB203580 and IKK␤-M together completely blocked the effects of TNF-␣. The effectiveness of SB203580 as an inhibitor of p38 was assessed by assaying MAPKAP-K2, a kinase that lies downstream of p38. When cultures were treated with 5 M, SB203580, there was a complete blockade of the ability of TNF-␣ to stimulated MAPKAP-K2, supporting the contention that at this concentration, SB203580 is an effective inhibitor of myocardial cell p38 (Fig.  3C). These results are consistent with a signaling process whereby TNF-␣-activated p38 and NF-B constitute the major pathways responsible for TNF-␣-mediated NF-B-dependent IL-6 transcriptional induction.
Recently, it has been shown that in HeLa cells, TNF-␣ stimulates TGF-␤-activated protein kinase 1 (Tak1), a newly discovered, multifunctional MAPKKK capable of binding to and activating IKK␤ and MKK6 ( Fig. 7; Ref. 22). To determine whether this MAPKKK can mediate NF-B-and p38-dependent IL-6 induction in myocardial cells, Tak1 and its required activator protein, Tak1-binding protein-1 (Tab1) (44), were both expressed in cardiac myocytes, along with the NF-B/IL-6 luciferase reporter. This combination of Tak1 and Tab1 resulted in an approximate 5-fold activation of NF-B-dependent IL-6 reporter gene expression, 6 which was inhibited by about 50 -60% by SB203580, by kinase-inactive MKK6 (MKK6-M2), or by kinase-inactive IKK␤ (IKK␤-M) (Fig. 2D). Together, SB203580 and IKK␤-M completely blocked Tak1-stimulated NF-B-dependent IL-6 reporter gene expression. These results are consistent with the capacity of Tak1 to stimulate both the IKK␤/NF-B and the MKK6/p38 pathways in cardiac myocytes, the latter of which we have observed. 7 Moreover, in conjunction with results shown in Fig. 2C, these findings indicate that the p38 pathway is required for optimal TNF-␣-and Tak1-mediated NF-B activation in cardiac myocytes.
To evaluate whether TNF-␣-mediated IL-6 induction in-  2-4), with TNF-␣ (1 ng/ml) (lanes 6 and 7), or with IL-6 (1 ng/ml) (lanes 8 and 9). Sphingosine (10 M) was then added to some cultures (lanes 4 -9), and after 4 h they were assessed for apoptosis by DNA ladder analysis, as described under "Materials and Methods." Standard DNA ladders consisting of fragments composed of 100-base pair increments are shown in lanes 1 and 10. The 400-base pair standard is shown. C, ladder quantitation. Densitometric analysis of the 360-base pair bands from E was carried out, and the average Ϯ S.D. of the results of each treatment are shown relative to the control (no cytokine, no sphingosine), which was set at 1.0. D, STAT3 phosphorylation. Duplicate cultures were treated for 15 min with or without 10 ng/ml of each test compound shown, and then extracts were assessed for the phosphorylation of STAT3 on Y-705 (P-STAT3) or for total STAT3 levels as described under "Materials and Methods." The controls are extracts of A431 cells that have been treated with (ϩ) or without (Ϫ) epidermal growth factor, which is known to induce STAT3 phosphorylation on Y-705. The -fold increase in STAT3 phosphorylation was determined by densitometry followed by normalizing each P-STAT3 value to the total STAT3 in that sample. Values shown are mean Ϯ S.D.; n ϭ 2 cultures/treatment. This blot is representative of three similar experiments.
volves the endogenous myocardial cell Tak1, which is present in neonatal rat heart (45), a kinase-inactive Tak-1 mutant (Tak1-M) was expressed. Tak1-M strongly inhibited TNF-␣induced IL-6 transcription by about 70% (Fig. 2E). As expected from the results in Fig. 2C, the p38 inhibitor, either SB203580 or MKK6-M1, blocked TNF-␣-inducible IL-6 transcription by about 50%. Further, SB203580 and the Tak1-M together nearly completely inhibited the effects of TNF-␣ (Fig. 2E). These results indicate that in cardiac myocytes, TNF-␣ signals through Tak1 to activate IL-6 transcription in a partially p38-dependent manner. This is consistent with the ability of Tak1 to activate MKK6 and p38.
MKK6 and p38 MAPK Activate IKK␤-Since MKK6 and p38 can activate NF-B and, thus, lead to the induction of IL-6, a cytokine with potentially important functions in the heart, we carried out experiments to begin elucidating the mechanism of cross-talk between the p38 and NF-B pathways. Initially, the effects of p38␤ 2 and/or MKK6(Glu) on the activity of IKK␤ in cultured myocardial cells were assessed. While expression of wild type p38␤ 2 alone had little effect on the kinase activity of IKK␤, expression of MKK6(Glu) or MKK6(Glu) together with p38␤ 2 increased IKK␤ kinase activity by about 2-and 5-fold, respectively (Fig. 4A). The IKK␤ activation mediated by p38␤ 2 and MKK6(Glu) was decreased by about 50% upon the addition of SB203580, indicating a requirement for p38.
MKK6 Associates with IKK␤-To pursue possible mechanisms by which MKK6(Glu) and p38␤ 2 can influence IKK␤ activity, experiments were carried out to evaluate whether p38␤ 2 and/or MKK6 can interact directly with IKK␤. Myocardial cells were transfected with constructs encoding FLAG-IKK␤ and FLAG-MKK6(Glu) and/or FLAG-p38 and the abilities of these proteins to interact with each other were assessed using immunoprecipitation followed by Western blotting. While transfection of FLAG-IKK␤ together with FLAG-p38␤ 2 did not result in any apparent complex formation (not shown), co-expression of FLAG-IKK␤ and FLAG-MKK6 did result in complex formation. Immunoprecipitating with either IKK␤-or MKK6-specific antisera resulted in the pull-down of both IKK␤ and MKK6 (Fig. 4B), consistent with the abilities of the two proteins to interact in vivo. Taken together, the results of these kinase and association experiments indicate that MKK6 can directly interact with IKK␤ and that this interaction may position p38 in the NF-B pathway, resulting in an enhancement in the activity of IKK␤.
MKK6 and p38 MAPK Induce IL-6 Release from Cardiac Myocytes-To evaluate whether IL-6 promoter induction correlates with increased expression of IL-6 protein itself, cardiac myocytes were treated with various doses of TNF-␣, and the quantity of IL-6 released into the culture medium was assessed by ELISA. TNF-␣ increased the quantity of IL-6 cross-reactive material in the medium in a dose-dependent manner by up to about 3-fold (Fig. 5A). The quantity of IL-6 in the medium was reduced significantly when SB203580 was added, consistent with a role for p38 in TNF-␣-simulated IL-6 release from myocardial cells.
To further determine the role of the p38 pathway in myocardial cell IL-6 induction and release, cardiac myocytes were treated with recombinant adenoviral strains harboring the genes for p38␤ 2 or MKK6(Glu). The levels of virus were adjusted so that all of the myocytes expressed the transgenes. Shown in Fig. 5B is an example using AdV-MKK6(Glu), a strain of adenovirus that possesses sequences encoding both MKK6(Glu) and GFP under the control of separate CMV promoters. Cultured myocytes were infected with virus and then photographed under phase and fluorescence microscopy 48 h later (see "Materials and Methods"). Upon inspection of low power micrographs by phase and fluorescence microscopy, it was apparent that essentially all of the cells visible in the phase field expressed GFP (compare Phase-Low with GFP-Low in Fig. 5B). Upon inspection of higher power micrographs, it was apparent that AdV-MKK6(Glu)-infected myocytes displayed the visual characteristics typical of hypertrophic cells, such as enlarged cell area and extensive sarcomeric organization (Fig. 5B, GFP-High and Phalloidin-High). This is consistent with previous studies of the effects of MKK6(Glu) on cardiac myocyte morphology (35,41). Cells infected with AdV-MKK6(Glu) and AdV-p38␤ 2 display similar morphology (not shown).
Infecting myocytes with AdV-p38␤ 2 alone did not significantly increase medium levels of IL-6 compared with control (Fig. 5C), consistent with the relatively weak activation of IL-6 transcription afforded by p38 alone (see Fig. 2A). However, infecting myocytes with AdV-MKK6(Glu) alone or with both AdV-p38␤ 2 and AdV-MKK6(Glu) together increased the release of IL-6 by about 2-and 8-fold, respectively (Fig. 5C), levels that are very similar to their abilities to induce IL-6 transcription (see Fig. 2A). Moreover, MKK6(Glu)-and p38/MKK6(Glu)dependent IL-6 release were both effectively inhibited by SB203580 (Fig. 5C), indicating that the selective activation of the p38 pathway confers significant IL-6 induction and release from cultured cardiac myocytes.
IL-6 Protects Cardiac Myocytes from Apoptosis-To explore whether IL-6 might exert functions consistent with myocardial recovery following stress, the effect of IL-6 on myocardial cell apoptosis was determined. When cardiac myocytes were treated for a limited time with sphingosine, a lipid second messenger known to mediate apoptosis in many cell types (46 -48), there was an approximately 3-fold increase in the number of apoptotic cells, as determined using TUNEL analysis (17). Strikingly, IL-6 exerted potent, dose-dependent, antiapoptotic effects, affording complete protection from sphingosine-induced apoptosis at doses above 0.1 ng/ml (Fig. 6A). TNF-␣ also conferred protection against sphingosine-induced apoptosis, again with nearly maximal effects being observed at 0.1 ng/ml cytokine. These results were confirmed using DNA laddering, where the fragmentation of DNA to create discrete increments of approximately 180 base pairs is diagnostic of apoptosis. Treating cells with spingosine resulted in an approximate 6-fold increase in the intensity of the DNA bands observed, and incubation with either IL-6 or TNF-␣ at 1 ng/ml resulted in complete protection from apoptosis, as determined by DNA laddering (Fig. 6, B and C). The mechanisms by which TNF-␣ and Il-6 confer protection from apoptosis remain to be determined. However, several recent studies have suggested that certain cytokines that signal through the gp130 receptor can activate the Janus kinase/STAT and protect against apoptosis in cardiac myocytes (49,50). Moreover, one recent study has shown that, although it does not couple through gp130, TNF-␣ can activate STAT3 in 3T3-L1 adipocytes (51). Accordingly, we tested the abilities of TNF-␣ and IL-6 to activate myocardial cell STAT3 by evaluating the level of Janus kinasemediated phosphorylation of Y-705 on immunoprecipitated STAT3. Treatment with TNF-␣ resulted in a small, 2-fold increase in STAT3 activation, while treatment with IL-6 induced a robust 8.2-fold increase in STAT3 activation (Fig. 6D). Thus, while it is conceivable that TNF-␣ could induce STAT3-sensitive genes involved in protection from apoptosis, it seems more likely that TNF-␣ exerts its cardioprotective effects in this cultured cell model via the activation of NF-B and the subsequent induction of cytokines, such as IL-6, which may signal to cardiac myocytes in an autocrine manner.

DISCUSSION
The way the heart responds to stress is critical for maintaining proper cardiac function. One of the stress responses that has attracted recent attention is the release of cardiac-derived cytokines. Although it is widely believed that these cytokines serve important functions in the heart, the precise nature of those functions remains unclear. Our findings indicate that the myocardial IL-6 system serves an important autocrine signaling role that contributes to myocyte survival following certain cardiac stresses. Moreover, it appears that the optimal induction of IL-6 in cardiac myocytes involves both the NF-B and p38 pathways. Our results indicate further that a potential point of collaboration between these signaling pathways resides in the ability of MKK6 to augment the activity of IKK␤ in a p38-dependent manner (Fig. 7).
The possibility that p38 plays a role in NF-B activation has been explored in cardiac myocytes (8,17) and in other cell types (e.g. see Ref. 42); however, the mechanism of interaction between the two pathways has not been fully resolved and has been addressed in only a few studies. To address whether p38/NF-B cross-talk could occur at the transcriptional level, Vanden Berghe et al. (15) used a Gal4-based one-hybrid assay and presented data indicating that p38 increased the transcriptional activation potential p65/NF-B. In contrast to that study, in preliminary experiments, we were unable to demonstrate such trans-activation of p65/NF-B in cardiac myocytes. 8 Moreover, while phosphorylation of p65/NF-B augments its ability to bind to DNA (52) and is apparently required for its transcriptional activity (53), recent data indicate that p38 itself does not phosphorylate p65/NF-B (54). To our knowledge, the results in the present study constitute the first demonstration that p38/NF-B cross-talk may exist at the level of MKK6 and IKK␤.
Although the mechanism by which MKK6 augments IKK␤ kinase activity is yet to be determined, our results showing that MKK6 and IKK␤ can physically interact provide some clues. For example, it is formally possible that MKK6, which has never been shown to phosphorylate any protein other than p38 (19,38), might phosphorylate IKK␤ and, in so doing, stimulate its kinase activity. However, we were unable to observe any change in the phosphorylation status of IKK␤ upon incubation with MKK6(Glu) in vitro (not shown). Alternatively, in addition to activating p38, MKK6 might serve to position p38 near IKK␤ and other members of the NF-B signaling pathway. This positioning might then lead to the observed p38-dependent increases in NF-B activity. Interestingly, we have been unable to observe the binding of p38 itself to IKK␤, which further supports an anchoring role for MKK6. IKK␤ and MKK6 are both members of the MAPKK family, and IKK␤ is known to associate with other MAPKKs in the NF-B signalsome (28,29). Thus, while it has not previously been shown that MKK6 can bind to other MAPKKs, it seems probable that IKK␤ could form dimers with MKK6 via the same binding domains involved in the formation of IKK homodimers. Future studies directed toward mapping such association domains in the IKKs, MKK6, and other MAPKKs will be required to address this provocative yet very feasible hypothesis.
To the best of our knowledge, the present study is also the first to demonstrate that a functional consequence of p38mediated NF-B activation by MKK6(Glu) is the induction of the IL-6 gene and release of IL-6 from cardiac myocytes. Consistent with roles for both pathways in IL-6 induction were our findings that either TNF-␣ or Tak1, both of which activate p38 and NF-B in other cell types (22) and in cardiac myocytes, 6 induced the IL-6 promoter. In agreement with results showing that the activation of p38 and NF-B by MKK6(Glu) leads to protection from apoptosis in cardiac myocytes (17) was our finding that IL-6 is antiapoptotic.
Although many studies have been carried out using IL-6related cytokines, clear physiological functions for IL-6 in the stressed heart remain elusive. For example, the expression of IL-6 in the heart has been associated with acute ischemia and cardiac failure, findings that have been interpreted as indicative of either a productive or a deleterious role for the cytokine (2)(3)(4)(5). Interestingly, in addition to the findings reported in this study, there is considerable evidence, mostly in noncardiac myocytes, that supports a cytoprotective and even growthpromoting role for IL-6. For example, IL-6 has been shown to protect hepatocytes and myeloma cells from apoptosis (55,56). Additionally, ET-1, which is a well known myocardial cell growth factor (57), is believed to mediate some of its cardiac growth-promoting effects through IL-6 (58). Interestingly, IL-6 was found to improve survival in an animal model of viral myocarditis, in part by reducing myocardial cell death (59). Moreover, the induction of IL-6 upon myocardial ischemia (6,8) and the finding that p38 can activate NF-B during cardiac ischemia (16) suggest that upon such stresses the heart responds by producing cytokines that may serve important roles as barriers against apoptosis.
Consistent with possible cytoprotective roles for cytokines in the heart was our finding that IL-6 afforded significant protection from apoptosis. The mechanism by which IL-6 might confer such protection remains to be elucidated. However, IL-6 belongs to the superfamily of trophic factors that includes cardiotrophin (CT-1), ciliary neurotrophic factor, leukemia inhibitory factor, and oncostatin M (60), most of which have been shown to activate ERK, STAT3, and Akt (61,62) and foster the growth and survival of cardiac myocytes, neurons, and other cells types via a ubiquitously expressed gp130 signal transduction mechanism. Thus, it is reasonable to speculate that it is through a related gp130 signaling mechanism that IL-6 confers the protection from apoptosis observed in the present study. Supporting this view are recent findings demonstrating defects in cardiac growth and development in gp130 knock-out mice (63). It is conceivable, therefore, that in some circumstances 8 R. Craig and C. C. Glembotski, unpublished observation. Shown is a simplified diagram depicting the signaling pathways under study in the present report. Tak1 is a MAP-KKK that requires a co-activator, Tab, for optimal activity. IKK␤ and MKK6 are MAPKK family members that can be activated by Tak1. The shaded arrows indicate possible cross-talk between the p38 and the NF-B pathways that is supported by the present study. the induction of IL-6 in the heart following ischemia or perhaps some other stresses may not always represent a cytotoxic response but may confer a cytoprotective response to the injury (4). Indeed, cardiac myocytes in the region surrounding a myocardial infarct express IL-6, which may have an autocrine cytoprotective effect on the myocytes that survive the initial ischemic insult (9). While the precise mechanism by which cytokines, like IL-6, might confer protection from apoptosis is unknown, it is of interest to note that both TNF-␣ and IL-1␤ can induce several cytoprotective proteins in the heart, the free-radical scavenger, manganese superoxide dismutase (62), and several of the heat shock proteins (64).
In summary, the results of the present study provide evidence for a signaling pathway whereby stimulators of p38, such as TNF-␣, can augment the activity of the NF-B pathway via cross-talk between MKK6 and IKK␤ (Fig. 7). Much remains to be understood about this interesting convergence of the NF-B and p38 MAPK pathways. For example, what is the mechanism by which MKK6 and p38 collaborate to enhance the activity of IKK␤? Do MKK6 and other known activators of p38, such as MKK3, occupy positions in the growing list of proteins comprising the NF-B signalsome? Do multiple isoforms of p38, such as p38␣, -␤, -␦, and -␥, all participate in this form of cross-talk? Perhaps most interesting are recent findings indicating that bone morphogenic proteins, which are required for directing the cardiac lineage during embryogenesis, operate through Tak1 to enhance cardiac-specific gene induction and growth (66). Since we found that in cardiac myocytes Tak1 is a potent activator of p38 and NF-B, and since both p38 and NF-B serve to enhance the transcription and release of IL-6 and probably other cytokines, it is tempting to speculate that in some ways the myocardial stress response recapitulates the early cardiac development program. For example, perhaps during early cardiac embryogenesis, BMP-mediated activation of p38 and NF-B augments the expression of genes sensitive to both pathways. Indeed, we and others have found that p38 can enhance the expression of a variety of cardiac genes that are normally up-regulated during early development and during myocardial hypertrophy. In fact, in one recent study, it was shown that Tak1 is expressed in high levels in the developing mouse heart, and it is activated during overload-induced hypertrophy in the adult mouse heart; additionally, overexpression of a constitutively active Tak1 was shown to induce myocardial hypertrophy in a transgenic mouse model (67). Thus, it is probable that, in collaboration with Tak1, p38 and NF-B together play important roles in the developing and hypertrophic myocardium, serving to induce NF-B-sensitive genes, such as IL-6, which could contribute to myocardial cell growth and the associated protection from apoptosis. Addressing these fascinating ideas will undoubtedly reveal more detail about the interesting connection between the developing and the stressed myocardium.