Genetic Inhibition or Activation of JNK1/2 Protects the Myocardium from Ischemia-Reperfusion-induced Cell Death in Vivo*

The c-Jun NH2-terminal kinase (JNK) branch of the mitogen-activated protein kinase signaling cascade has been implicated in the regulation of apoptosis in a variety of mammalian cell types. In the heart, disagreement persists concerning the role that JNKs may play in regulating apoptosis, since both pro- and antiapoptotic regulatory functions have been reported in cultured cardiomyocytes. Here we report the first analysis of cardiomyocyte cell death due to JNK inhibition or activation in vivo using genetically modified mice. Three separate mouse models with selective JNK inhibition were assessed for ventricular damage and apoptosis levels following ischemia-reperfusion injury. jnk1–/–, jnk2–/–, and transgenic mice expressing dominant negative JNK1/2 within the heart were each shown to have less JNK activity in the heart and less injury and cellular apoptosis in vivo following ischemia-reperfusion injury. To potentially address the reciprocal gain-of-function phenotype associated with sustained JNK activation, transgenic mice were generated that express MKK7 in the heart. These transgenic mice displayed elevated cardiac c-Jun kinase activity but, ironically, were also significantly protected from ischemia-reperfusion. Mechanistically, JNK-inhibited mice showed increased phosphorylation of the proapoptotic factor Bad at position 112, whereas MKK7 transgenic mice showed decreased phosphorylation of this site. Collectively, these results underscore the complexity associated with JNK signaling in regulating apoptosis, such that sustained inhibition or activation both elicit cellular protection in vivo, although probably through different mechanisms.

The c-Jun NH 2 -terminal kinase (JNK) branch of the mitogen-activated protein kinase signaling cascade has been implicated in the regulation of apoptosis in a variety of mammalian cell types. In the heart, disagreement persists concerning the role that JNKs may play in regulating apoptosis, since both pro-and antiapoptotic regulatory functions have been reported in cultured cardiomyocytes. Here we report the first analysis of cardiomyocyte cell death due to JNK inhibition or activation in vivo using genetically modified mice. Three separate mouse models with selective JNK inhibition were assessed for ventricular damage and apoptosis levels following ischemia-reperfusion injury. jnk1؊/؊, jnk2؊/؊, and transgenic mice expressing dominant negative JNK1/2 within the heart were each shown to have less JNK activity in the heart and less injury and cellular apoptosis in vivo following ischemia-reperfusion injury. To potentially address the reciprocal gain-of-function phenotype associated with sustained JNK activation, transgenic mice were generated that express MKK7 in the heart. These transgenic mice displayed elevated cardiac c-Jun kinase activity but, ironically, were also significantly protected from ischemia-reperfusion. Mechanistically, JNK-inhibited mice showed increased phosphorylation of the proapoptotic factor Bad at position 112, whereas MKK7 transgenic mice showed decreased phosphorylation of this site. Collectively, these results underscore the complexity associated with JNK signaling in regulating apoptosis, such that sustained inhibition or activation both elicit cellular protection in vivo, although probably through different mechanisms.
The mitogen-activated protein kinases (MAPKs) 3 are composed of a series of successively acting kinases that amplify and transduce signals following growth and/or stress stimulation (1). In its broadest sense, the MAPK signaling cascade is composed of at least three distinct phospho-rylation cascades that culminate in the activation of the extracellular signal-regulated kinases (ERKs), the c-Jun NH 2 -terminal kinases (JNKs), and the p38 MAP kinases (1). JNKs are activated in response to environmental stress or membrane-bound receptor signaling through GTPases of the Rho family through the MAPK kinase kinases (2). These MAPK kinase kinases then promote activation of the MAPK kinases such as MKK4 and MKK7, which function as dual specificity protein kinases to directly phosphorylate JNK1, JNK2, and JNK3 (2). Gene targeting in mice has revealed a critical role for JNK signaling in regulating cellular viability, induction of apoptosis, and cellular proliferation through phosphorylation of AP-1, p53, c-Myc, nuclear factor of activated T cells, Sap-1, and Bcl-2 family members (2)(3)(4)(5).
An emerging area of investigation has demonstrated a critical, yet often paradoxical role for JNKs and their upstream activators in regulating cellular apoptosis. For example, jnk null fibroblasts were shown to be deficient in mitochondria-driven apoptosis, suggesting that JNK plays an important role in this process (6). Similarly, jnk3 null mice showed significant reductions in excitotoxicity-induced hippocampal neuron apoptosis (7). Moreover, jnk2 null mice also showed apoptotic defects in immature thymocytes (8), whereas jnk1-jnk2 double null mice showed region-specific alterations in apoptosis in the developing mouse brain (9). Upstream of JNK1/2/3, apoptosis signal-regulating kinase (ask) null cells showed significant impairment in H 2 O 2 -and TNF-␣induced cell death (10). Consistent with this observation, JNK1 inhibition was reported to protect cardiac myocytes from ischemia-induced apoptosis, whereas JNK2 inhibition had no effect (11). Similarly, inhibition of JNK in H9c2 myocytes blocked apoptosis induced by stress stimulation (12), whereas infection of adult cardiac myocytes with a dominant negative JNK-expressing adenovirus antagonized H 2 O 2 -induced apoptosis (13). Moreover, ␤-adrenergic-stimulated apoptosis of cultured adult cardiomyocytes was inhibited by expression of dominantnegative JNK through a mitochondria-dependent mechanism (14,15). Paradoxically, mekk1 null embryonic stem cells (MEKK1 is the direct upstream activator of MKK4 and MKK7) differentiated into cardiac myocytes showed enhanced oxidative stress-induced apoptosis (16), consistent with the observation that mekk1Ϫ/Ϫ mice have more cardiac apoptosis following pressure overload stimulation (17). Additional discordance was suggested by the observation that dominant negative JNK1 or dominant negative MKK4-expressing cultured cardiomyocytes had increased nitric oxide-induced apoptosis (18). Moreover, ischemiareperfusion (I-R)-induced apoptosis in cultured neonatal cardiomyocytes was increased by expression of JNK inhibitory mutants (19). Thus, it is readily apparent that JNK signaling has both pro-and antiapoptotic ramifications, although the overall dominance of either effect has not been extended to the heart in vivo.

MATERIALS AND METHODS
Terminal Deoxynucleotidyltransferase-mediated dUTP Nick End Labeling (TUNEL)-TUNEL from histological heart sections was performed with the In Situ cell death detection kit (Roche Applied Science) as per the manufacturer's instructions with slight modification (20). Hearts were collected 24 h after ischemia-reperfusion injury for assessment of TUNEL within the infarction zone proper. At least 16 sections were analyzed throughout the entire longitudinal axis of the hearts, from which ϳ5,000 myocytes were assessed each (n ϭ 2-3 hearts/group).
Transgenic and Gene-targeted Mice-Transgenic mice containing concatomers of dominant negative JNK1 and JNK2 (FVB/N background) as well as jnk1Ϫ/Ϫ and jnk2Ϫ/Ϫ gene-targeted mice (C57Bl/6 background) were previously described (8,22). MKK7 transgenic mice (FVB/N background) were generated by cloning the mouse cDNA for wild-type MKK7 into the heart-specific ␣-myosin heavy chain promoter, as described for the dominant negative JNK1/2 transgenic mice (22). Six MKK7 transgenic founders were initially identified at the time of genotyping (8 days of age), only one of which was viable long term due to lower levels of expression (the other founders died between 8 days and 5 weeks of age).
Ischemia-Reperfusion in Mice-Cardiac ischemia-reperfusion injury was performed in 6 -8-week-old mice as described previously (23). The thoracotomy was closed, and the mice were revived for a 45-min (C57Bl/6 background) or a 60-min (FVB/N background) ischemic period, after which the knot was released and the heart was allowed to reperfuse for 24 h. Upon the completion of the reperfusion period, mice were sacrificed by CO 2 asphyxiation, and the hearts were quickly removed for analysis of infarction injury with 2% triphenyl tetrazolium chloride as described previously (23). Images were quantified for area not at risk, area at risk (AAR), and infarcted area (IA).
Echocardiography and Working Heart Analyses-For echocardiography, mice were anesthetized with 2% isoflurane, and hearts were visualized using a Hewlett Packard Sonos 5500 instrument and a 15-MHZ transducer. Cardiac ventricular dimensions were measured on M-mode three times for the number of animals indicated in the table. The isolated ejecting mouse heart preparation has been described in detail previously (24).
Kinase Assays-Cardiac ERK, JNK, and p38 kinase activities were determined using a published method (25) with minor modifications (22). Soluble protein extracts (1 mg) were incubated with agarose-conjugated antibodies to ERK1 and ERK2 or p38 for 3 h. Complexes were washed two times in lysis buffer and two times in kinase buffer. MBP was used as substrate for ERK and p38, whereas GST-c-Jun (amino acids 1-79) was used to isolate JNK kinases and as a substrate. The reaction was started by the addition of 100 M ATP and 10 Ci of [␥-32 P]ATP and incubated at 30°C for 20 min followed by SDS-polyacrylamide gel electrophoresis.
Statistical Analysis-Data are expressed as means Ϯ S.E. Differences between experimental groups were evaluated for statistical significance using Student's t test for unpaired data or one-way analysis of variance followed by Bonferroni's post-test. p values of Ͻ0.050 were considered to be statistically significant. All statistical analyses were performed by using Instat 3.0 software (GraphPad).

Characterization of Cardiac JNK Activity in JNK-inhibited Mice-
Previous reports attempting to establish a causal association between JNK signaling and apoptosis in cardiomyocytes have been largely equivocal, possibly due to a primary reliance on culture-based models. Here we evaluated three distinct animal models with inhibited JNK signaling: mice lacking the jnk1 gene, mice lacking the jnk2 gene, and mice containing a heart-specific transgene that simultaneously expresses dominant negative mutants of JNK1 and JNK2. We have previously shown that all three genetic models have a significant reduction in total JNK activity within the heart at base line and following acute agonist stimulation in vivo (22). These mice did not show alterations in base line or inducible ERK or p38 activity within the heart, demonstrating specificity for JNK signaling itself (22). With respect to I-R injury to the heart, wild-type mice in either the FVB/N or C57Bl/6 strain background showed significant elevations in total cardiac JNK activity (p Ͻ 0.05) (Fig. 1, A and B). However, JNK1/2 dominant negative transgenic mice, jnk1Ϫ/Ϫ mice, and jnk2Ϫ/Ϫ mice each showed significant attenuation in JNK activity in the heart following I-R injury (p Ͻ 0.05) (Fig. 1, A and  B). No alterations were observed in ERK or p38 activity in response to the dnJNK1/2 transgene or the jnk1 or jnk2 null alleles at base line or FIGURE 1. Cardiac c-Jun NH 2 -kinase activity in wild-type or JNK-inhibited mice. A, JNK activity in the heart from wild-type (Wt) mice or dnJNK1/2 transgenic mice in the FVBN background at base line or following I-R injury. A graph is shown below that summarizes the results from two independent experiments performed in triplicate. *, p Ͻ 0.05 versus sham. B, JNK activity in the heart from wild-type mice or jnk1Ϫ/Ϫ or jnk2Ϫ/Ϫ mice in the C57Bl/6 background at base line or following I-R injury. A graph is shown below that summarizes the results from two independent experiments performed in triplicate. *, p Ͻ 0.05 versus sham.
after I-R injury (data not shown). Collectively, these results indicate that I-R injury is a potent inducer of cardiac JNK activity and that all three genetic models used here show a significant and specific reduction in cardiac JNK signaling.
JNK-inhibited Mice Show Less Cardiac Injury following I-R-To investigate the functional correlations between reduced JNK activity and cardiac cellular damage following ischemic injury, each of the three genetic models was subjected to a surgical I-R procedure. The data demonstrate significant cardiac injury in wild-type control mice as assessed by measurement of viable myocardium at risk (AAR) versus the nonviable IA (p Ͻ 0.05) ( Fig. 2A). However, dnJNK1/2 transgenic mice, jnk1Ϫ/Ϫ mice, and jnk2Ϫ/Ϫ mice each showed a significant protection from I-R-induced myocardial damage when compared with their strainmatched controls (p Ͻ 0.05) ( Fig. 2A). No difference in the AAR versus the left ventricular (LV) area was observed in any of the groups, suggesting that the difference in infarction injury was not a secondary consequence of variable perfusion area. Collectively, these results indicate that JNK-inhibited mice have reduced susceptibility to cardiac injury following I-R injury in vivo.
JNK-inhibited Mice Show Reduced Indexes of Apoptosis following I-R Injury-The observed reduction in left ventricular injury following I-R in each of the JNK-inhibited mice suggested an alteration in cellular apoptosis rates in the heart. To address this possibility, left ventricular tissue was harvested after I-R injury and subjected to Western blotting for cleaved caspase-3 (activated). The data demonstrate significant caspase-3 cleavage in FVB/N and C57Bl/6 wild-type hearts following I-R injury but significantly less cleavage in dnJNK1/2, jnk1Ϫ/Ϫ, and jnk2Ϫ/Ϫ hearts (p Ͻ 0.05) (Fig. 3A). On average, each of the three JNK-inhibited models showed an approximate 50% reduction in total caspase-3 cleavage across two independent experiments, each containing triplicate samples (six hearts each). dnJNK1/2 transgenic mice also showed significant reductions in DNA laddering after I-R injury compared with wild-type controls (data not shown). Finally, total cell apoptosis rates were also evaluated by TUNEL within the LV infarcted area itself following I-R injury (after 24 h of reperfusion). The data demonstrate that I-R injury dramatically enhances TUNEL within the IA of the LV of wild-type C57Bl/6 mice but that the jnk1/2 gene-targeted mice displayed a significant decrease in this index of cell death (Fig. 3B). A consistent trend toward less poly(ADP-ribose) polymerase cleavage was also observed in the JNK-inhibited mice (data not shown). Collectively, these data indicate that JNK-inhibited mice are less susceptible to cardiomyocyte cell death following I-R injury.
Sustained JNK Activation Also Protects the Heart from Death in Vivo-The data discussed above demonstrate that genetic inhibition of JNK in the heart provides protection from I-R-induced cell death. Thus, we initially hypothesized that activation of JNK within the heart would produce the antithetic effect of increased cell death. To validate this hypothesis, we generated transgenic mice expressing the JNK-specific dual specificity kinase MKK7 using the heart-specific ␣-myosin heavy chain promoter (Fig. 4A). Whereas six lines were initially identified at the time of genotyping (8 days of age), five died before breeding age. One line died at 5 weeks of age with severe dilated cardiomyopathy upon histological examination (Fig. 4B, presumed higher expressor). The  . Assessment of apoptosis markers in wild-type and JNK-inhibited mice. A, Western blot analysis of cleaved caspase-3 (Csp3) or glyceraldehyde-3-phosphate dehydrogenase (GAPDH) (control) from heart protein extracts generated from the indicated cohorts of mice at base line or following I-R injury as described under "Materials and Methods." A graph is shown below that summarizes the results from two independent experiments performed in triplicate. Wt, wild-type. *, p Ͻ 0.05 versus sham. B, quantitation of TUNEL within the AAR following I-R injury of the indicated cohorts. At least three mice were analyzed in each group using multiple histological sections that essentially cover the entire AAR. Non-I-R-injured hearts showed negligible rates of TUNEL in each of the cohorts. Statistical comparison utilized Student's t test.
lethality observed in the majority of MKK7 transgenic mice is consistent with previous descriptions of severe cardiomyopathy in independently generated MKK7 transgenic mice with robust JNK activation (26,27). These results are also in contrast to our previous experience in generating MEK1 heart-specific transgenic mice in which all nine original founders were viable, even with exceedingly high expression levels (28).
Despite the lethality associated with MKK7 transgene, one low copy line (based on Southern blotting) was identified that appeared relatively normal upon histological examination at 7 weeks of age (Fig. 4B, presumed low expressor relative to the other lethal lines). Quantitative Western blotting demonstrated 3.8-fold more MKK7 protein in the heart compared with wild-type mice (Fig. 4C). Whereas "low" MKK7 transgenic mice were viable and relatively normal, they did show a mild 10 -15% increase in normalized heart weight, suggesting some degree of cardiac hypertrophy (TABLE ONE). Despite this observation, MKK7 transgenic mice demonstrated no signs of functional decompensation at 6 and 9 months of age as assessed by echocardiography and working heart preparations (TABLES ONE and TWO, and data not shown). In fact, MKK7 transgenic mice were slightly hyperfunctional in systole (ϩdP/dt), probably due to the 10 -15% compensated hypertrophy response that characterizes these mice (TABLE TWO). Low expressing MKK7 transgenic mice also did not show any signs of cardiac pathology upon histological examination (data not shown).
Hearts harvested from adult MKK7 transgenic mice consistently showed 4 -5-fold greater c-Jun kinase activity (Fig. 4, D and E). MKK7 transgenic hearts did not show any alterations in p38 phosphorylation levels at base line or following I-R injury. However, MKK7 transgenic hearts did show a mild increase in ERK1/2 phosphorylation at base line (Fig. 4F). Remarkably, adult MKK7 transgenic mice showed significantly smaller infarctions following I-R injury compared with nontransgenic littermates (Fig. 4G). Consistent with this observation, MKK7 transgenic mice also demonstrated lower levels of TUNEL in the infracted area of the LV compared with wild-type controls (p Ͻ 0.05) (Fig. 4H). Thus, MKK7 expression in the heart, which is associated with a mild increase in JNK1/2 activity, also promoted cellular protection following I-R injury in vivo.
JNK-inhibited Mice Demonstrate Increased Bad Phosphorylation at Amino Acid 112-A series of Western blots was performed to assay expression levels and/or phosphorylation status of key apoptosis-regulatory factors in an attempt to identify a potential mechanism that might underlie, in part, the relative cardiac protection observed in the JNKinhibited mice. Western blots were performed for most Bcl-2 family members, only one of which showed a significant correlation with JNK inhibition in vivo. For example, no change was observed in Bcl-2, Bcl-xl, Bax, or Bad in the hearts of JNK-inhibited mice at base line or following I-R injury; nor were changes observed in Bad phosphorylation at serines 128 and 136 (data not shown). In contrast, a consistent increase in Bad phosphorylation at serine 112 was observed at base line and after I-R injury in the hearts of jnk1Ϫ/Ϫ and dnJNK1/2 transgenic mice, and a trend toward an increase was also observed in jnk2Ϫ/Ϫ hearts (Fig. 5, A-C). No alteration in total Bad protein levels was observed (Fig. 5, A-C). Consistent with these results, MKK7 transgenic mice showed a reproducible decrease in Bad serine 112 phosphorylation in the heart (Fig. 5D). Thus, alteration in Bad phosphorylation at serine 112 may represent one potential effector mechanism downstream of JNK signaling in the heart that participates in the apoptotic response. However, phosphorylation of serine 112 inactivates the proapoptotic functions of Bad; thus, the decrease in phosphorylation observed in MKK7 transgenic hearts would not be consistent with protection from I-R injury (see "Discussion").

DISCUSSION
A number of equally robust studies conducted in cultured cardiomyocytes have reported both prosurvival and prodeath regulatory effects associated with altered JNK signaling (29). A similar dichotomous paradigm has emerged in neurobiology, such that JNK is both protective and apoptotic, depending on the region of the brain analyzed. For example, jnk1/2 double null mice die as embryos and display reduced levels of cell death in regions of the hindbrain prior to neural tube closure but increased levels of apoptosis in the forebrain and other regions of the hindbrain at a later time point (9,30). Moreover, embryonic stem cells or fibroblasts derived from mekk1 null and mkk4/7 double null mice, which were each defective in JNK signaling, were still capable of undergoing apoptosis to multiple stimuli (31,32). In direct contrast to these results, jnk1/2 double null fibroblasts were largely resistant to cell death induced by similar stimuli (6). A similar paradigm is also observed with the apoptotic propensity of T-and B-lymphocytes. For example, jnk1 or jnk2 null T cells were resistant to anti-CD3-induced apoptosis, suggesting a proapoptotic regulatory role (33). However, jnk1 null pre-B cells showed diminished survival, suggesting an antiapoptotic role for JNK in these cells (34). Thus, numerous examples exist to support the dichotomous regulatory functions attributed to JNK signaling in controlling cellular survival or death.
One potential explanation for the seemingly opposite results dis- . Activation of JNK1/2 in the heart also protects from I-R injury. A, schematic of the construct used to generate MKK7 transgenic mice driven from the ␣-myosin heavy chain promoter. B, hematoxylin/eosin-stained histological sections of hearts from a wildtype mouse (Wt), a presumed high expressing MKK7 transgenic founder that died (5 weeks of age), and the viable low expressing MKK7 transgenic line (7 weeks of age). C, Western blot for MKK7 protein from the hearts of wild-type and low expressing MKK7 transgenic mice (n ϭ 3 each). D, SDS-PAGE of a c-Jun kinase activity assay from the hearts of wild-type or MKK7 transgenic mice at 2 months of age (four hearts each). E, a graph of the data from D. *, p Ͻ 0.05 versus wild-type. F, Western blotting for phosphorylated ERK1/2 and p38 as well as total ERK1/2 and p38 from the hearts of the indicated mice after sham or I-R injury. G, quantitation of IA to AAR in adult wild-type (n ϭ 7) and MKK7 transgenic mice (n ϭ 8) subjected to I-R injury. *, p Ͻ 0.05 versus wild-type. H, TUNEL from the LV infarcted area of the hearts shown in G. *, p Ͻ 0.05 versus wild-type I-R. Statistical comparison utilized Student's t test.
cussed above is that JNK signaling intersects with multiple apoptotic effector pathways that might differ with respect to cell type and/or timing of expression and interaction with JNK. The Bcl-2 family appears to serve as prominent downstream effectors of JNK signaling that may partially underlie the prodeath versus prosurvival signaling decisions of JNK. For example, stress-induced cytochrome c release and apoptosis was shown to require JNK activity and effective Bax/Bak function, suggesting a prominent role in regulating the intrinsic apoptotic pathway at the level of the mitochondria (35). Indeed, Izumo and colleagues showed that activated JNK and MKK4 localized to the mitochondria, induced cytochrome c release, and promoted the apoptosis of cardiomyocytes (15). That a fraction of JNK protein localizes to the mitochondria was also observed by Ping and colleagues (37) and is consistent with an earlier observation made in fibroblasts whereby JNK was required for cytochrome c release and mitochondria-driven cell death (6). JNK signaling can also induce apoptosis through a transcription-dependent mechanism that involves c-Jun activation and the induction of proapoptotic genes, which might also be linked to Bcl-2 family member modulation (38). With respect to more direct regulatory mechanisms, JNK was shown to phosphorylate the prosurvival factor Bcl-2, resulting in its inactivation and the induction of apoptosis (39,40). JNK was also shown to directly phosphorylate serine 128 on Bad, enhancing cell death through the intrinsic pathway (41). However, JNK-mediated phosphorylation of Bad at threonine 201 had the opposite effect of enhancing cell survival by inactivating Bad (42). Thus, JNK signaling can have dichotomous effects on cell survival even through the same effector protein by putatively phosphorylating different residues. We did not observe alterations in Bad phosphorylation at serine 128 or 136 in response to reduced or increased JNK signaling in the heart, in contrast to the observed increase and decrease in serine 112 phosphorylation from JNK-inhibited and activated hearts, respectively. Phosphorylation of serine 112 in Bad was shown to inactivate the proapoptotic function of this Bcl-2 family member, providing cellular protection (43)(44)(45)(46). jnk1Ϫ/Ϫ and dnJNK1/2 mice showed a significant increase in Bad serine 112 phosphorylation, a result that might partially underlie the protective effect associated with reduced cardiac JNK activity. However, MKK7 transgenic mice had a significant reduction in serine 112 phosphorylation, a result that is consistent with proximal regulation by JNK but is inconsistent with protection from cell death observed in MKK7 transgenic hearts. Decreased phosphorylation of serine 112 should be more permissive to cell death not protective as observed in MKK7 transgenic hearts. Thus, whereas there is a strong correlation in regulation between JNK signaling and Bad 112 phosphorylation, regulation of Bad does not play a dominant role in determining the survival Echocardiographic and cardiac gravimetric analysis of MKK7 transgenic mice All measurements are means Ϯ S.E., and each animal was measured three separate times (3-6 mice/group). The echocardiographic data were obtained from 6-month-old mice, whereas the gravimetric data were collected at 7 weeks of age, although similar results were observed at 6 months of age (data not shown). Septal and LV wall thicknesses were assessed in diastole and shown as mm. LVED, left ventricular end-diastolic dimension; LVES, left ventricular end-systolic dimension; FS, fractional shortening; HW/BW, heart weight/body weight (mg/g); HW/TL, heart weight/tibia length (mg/mm); TG, transgenic.  of myocytes in the hearts of MKK7 transgenic mice. Protection from I-R injury in MKK7 transgenic hearts may be partially associated with increases in ERK1/2 phosphorylation (Fig. 4), since MEK1-ERK1/2 signaling was previously shown to protect the mouse heart from I-R injury in vivo (28,47). Alternatively, the mild cardiac hypertrohic phenotype itself, associated with the MKK7 transgene, may provide a degree of protection from cell death through undefined mechanisms. It is also ironic that loss of a kinase (JNK) is associated with greater phosphorylation of Bad at serine 112, whereas increases in JNK activity are associated with less phosphorylation of serine 112. A number of kinases have been shown to directly mediate phosphorylation of serine 112 in Bad, such as protein kinase A, Rsk, p21-activated kinase 1, and PIM-2, any one of which could be indirectly regulated by JNK at the mitochondria (43)(44)(45)(46). JNK activation was also recently shown to directly activate Bid cleavage, independent of caspase-8, thus further facilitating the mitochondrial death pathway (48). Finally, JNK was reported to directly phosphorylate two additional proapoptotic members of the BH3-only subgroup of Bcl-2 proteins, Bim and Bmf, facilitating their translocation to the mitochondria (49). The observed association between JNK inhibition and the modulation of Bcl-2 family members is reminiscent of our previous observations in which p38 MAPK-inhibited transgenic mice showed a dramatic up-regulation in Bcl-2 expression within the heart (23). In fact, genetic inhibition of p38 MAPK signaling in the heart also provided protection from I-R injury similar to that of JNK inhibition presented here (23). However, Glembotski and colleagues generated transgenic mice with constitutive p38 activation in the heart due to expression of MKK6, which was also associated with profound protection from I-R injury similar to the protection observed here due to constitutive JNK activation mediated by MKK7 (50). Thus, p38 MAPK appears to function analogous to JNK in the heart, such that sustained activation or inhibition provides protection from cell death. Mechanistically, we favor the hypothesis that both the pro-and antiapoptotic effects associated with JNK and p38 modulation occur through regulatory relationships with specific Bcl-2 family members.
In addition to differential effects on diverse Bcl-2 family members that vary by cell type or time of expression, JNK signaling might even enhance or attenuate apoptosis within the same cells, depending on the absolute magnitude, duration, or context of the signaling response. For example, robust and sustained JNK activation might augment apoptosis, whereas transient or mild JNK activation might act in a protective manner by "preconditioning" or "desensitizing" a given cell type to prodeath stimuli. Alternatively, the dichotomous apoptotic effects attributed to JNK signaling (and p38 signaling) may also depend on co-regulation or recruitment of other intracellular signaling pathways that are independent of the Bcl-2 family. The data presented here, although seemingly dichotomous, actually solidify the hypothesis that JNK signaling specifically and simultaneously modulates pro-and antiapoptotic effector mechanisms within cardiomyocytes. However, with respect to a potential therapeutic application of this pathway in providing cellular protection, we favor the hypothesis that acute inhibition of JNK will potentially benefit the myocardium most (as opposed to chronic activation of JNK signaling), consistent with a recent pharmacological study (36).