p38α Mitogen-Activated Protein Kinase Plays a Critical Role in Cardiomyocyte Survival but Not in Cardiac Hypertrophic Growth in Response to Pressure Overload

Abstract

The molecular mechanism for the transition from cardiac hypertrophy, an adaptive response to biomechanical stress, to heart failure is poorly understood. The mitogen-activated protein kinase p38α is a key component of stress response pathways in various types of cells. In this study, we attempted to explore the in vivo physiological functions of p38α in hearts. First, we generated mice with floxed p38α alleles and crossbred them with mice expressing the Cre recombinase under the control of the α-myosin heavy-chain promoter to obtain cardiac-specific p38α knockout mice. These cardiac-specific p38α knockout mice were born normally, developed to adulthood, were fertile, exhibited a normal life span, and displayed normal global cardiac structure and function. In response to pressure overload to the left ventricle, they developed significant levels of cardiac hypertrophy, as seen in controls, but also developed cardiac dysfunction and heart dilatation. This abnormal response to pressure overload was accompanied by massive cardiac fibrosis and the appearance of apoptotic cardiomyocytes. These results demonstrate that p38α plays a critical role in the cardiomyocyte survival pathway in response to pressure overload, while cardiac hypertrophic growth is unaffected despite its dramatic down-regulation.

In response to increased hemodynamic stress, the heart initiates a compensatory response in the form of cardiac hypertrophy. The resulting increase in cardiac mass reduces wall stress and thus leads to improvement in cardiac performance. However, sustained excessive workloads may lead to heart failure by activating an intracellular signaling cascade leading to cardiomyocyte dysfunction and death with replacement fibrosis. The signal transduction mechanisms responsible for mediating the transition to heart failure are still far from being conclusively identified.

Mitogen-activated protein (MAP) kinase cascades are highly conserved signal transduction pathways which couple various extracellular signals to a range of intracellular responses that allow the organism to adapt, survive, and maintain homeostasis. The MAP kinase family consists of extracellular signal-regulated protein kinase (ERK), c-Jun NH₂-terminal protein kinase (JNK), and p38 MAP kinase. The p38 MAP kinase is activated in response to proinflammatory cytokines as well as hormones, to ligands for G protein-coupled receptors, and to stresses such as osmotic shock and heat shock (13). p38 has four subfamilies, α, β, γ, and δ, of which p38α is expressed widely and has an important function in cytokine production and the response to many types of stress. Loss of p38α has been established to cause embryonic death at midgestation (1, 30). Experiments with p38α^−/− mice indicated that p38α is required for placental organogenesis and developmental and stress-induced erythropoiesis through regulation of erythropoietin expression.

Previous studies have demonstrated that p38 is activated in cultured neonatal cardiomyocytes by hypertrophic stimulation (7, 23) and in mouse hearts in response to pressure overload (33). In neonatal myocytes, some reports implicate p38 in the myocyte growth response (23, 33, 37), while other studies have suggested that p38 inhibition is not sufficient to attenuate all features of cardiomyocyte hypertrophy (6, 7). Recent studies with transgenic mice do not support the hypothesis that p38 promotes cardiomyocyte growth (3, 19, 38). A report by Braz et al. suggests an antihypertrophic function of p38α (3). In addition, the previous findings suggest that p38α has a protective and/or promoting function in the regulation of cell death in various cells (12, 24, 34) as well as in cardiomyocytes (8, 11, 14, 20, 21, 33, 37).

The involvement of p38α in cardiac hypertrophic growth and cell survival is still far from clear. In the previous reported studies, the p38 inhibitors, such as SB203580 or SB202190 (18), or overexpression of constitutively active or dominant negative mutants of MAP kinases has been used to elucidate a role of p38. The p38 inhibitors are potent inhibitors of both the α and β isoforms of p38 (16). Overexpression of a protein may induce an artificial effect on cardiac structure and function. In the study reported here, we attempted to identify the in vivo role of p38α in cardiac myocytes by using cardiac-specific p38α knockout mice. These mice were viable and displayed normal global cardiac structure and function. Pressure overload induced normal hypertrophic responses in these mice but also cardiac dysfunction and heart dilatation, indicating that p38α plays a critical role in the cardiomyocyte survival pathway in response to pressure overload, while cardiac hypertrophic growth is unaffected despite its dramatic down-regulation.

MATERIALS AND METHODS

Generation of cardiac-specific p38α knockout (p38α CKO) mice.

This study was carried out under the supervision of the Animal Research Committee and in accordance with the Guidelines for Animal Experiments of Osaka University and with the Japanese Animal Protection and Management Law (no. 25). A 13-kb NotI fragment including the exon containing the glycine-rich ATP binding loop from a mouse 129 Svj genomic library was used to make the targeting construct. The targeting construct was made by inserting the loxP site into the BglII site located 350 bp downstream of the exon and the loxP sites along with the phosphoglycerate kinase (PGK)-neomycin resistance (neo) gene into the BglII site located 1.3 kb upstream of the exon. The targeting vector contained 1.3 kb of the homologous DNA upstream of the loxP-PGKneo-loxP cassette site and 7.4 kb of the homologous DNA downstream of third loxP site. PCR, Southern blotting and karyotyping analyses were performed in order to obtain embryonic stem (ES) clones exhibiting the desired homologous recombination and normal karyotype. Highly chimeric mice, generated by aggregating these targeted ES cells into BDF1 blastocysts, were bred with C57B/6J mice. To remove the selection marker gene, PGK-neo, and obtain the type II deletion, F₁ mice with germ line transmission of the loxP-targeted p38α allele were crossbred with EIIa-cre mice (17), resulting in heterozygous p38α-floxed mice without PGK-neo. Mice expressing the Cre recombinase under the control of the α-myosin heavy-chain promoter (α-MHCCre mice) in the C57B/6J background were generated as previously reported (2) and were mated with p38α-floxed mice. To confirm the cardiac-specific expression of the Cre recombinase in the α-MHCCre mice, they were mated with a transgenic mouse line carrying a reporter gene construct which directs expression of the Escherichia coli lacZ gene following Cre-mediated excision of the loxP-flanked chloramphenicol acetyltransferase gene (29). Crossing the α-MHCCre mice with the reporter mice resulted in Cre-mediated recombination in the heart but not in other tissues as determined by X-Gal (5-bromo-4-chloro-3-indolyl-β-d-galactopyranoside) staining (data not shown). Primers used for PCR screening were PGK-1 (5′-TAGTGAGACGTGCTACTTCCATTTGTCACG-3′) and A6 (5′-TCTCCTTCCAGCTAAGCTCTACCACCATAG-3′) (2-kb PCR product), A3 (5′-ATGAGATGCAGTACCCTTGGAGACCAGAAG-3′) and A4 (5′-AGCCAGGGCTATACAGAGAAAAACCCTGTG-3′) (180-bp PCR product from the wild-type allele and 230-bp product from the p38α targeted allele), 246F (5′-CGTCTAAGAAACCATTATTATCATGAC-3′) and 246R (5′-ATGGCCAGTACTAGTGAACCTCTTCGA-3′) (170-bp PCR product), A1 (5′-CCACAGAAGAGATGGAGCTATATGGATCTC-3′) and A4 (420-bp PCR product), and Cre1 (5′-GTTCGCAAGAACCTGATGGACA-3′) and Cre2 (5′-CTAGAGCCTGTTTTGCACGTTC-3′) for Cre gene transmission (340-bp PCR product).

TAC.

Thoracic transverse aortic constriction (TAC) was performed in 10- or 11-week-old male p38α CKO mice as previously described (9). The right and left carotid arteries were cannulated with heat-stretched PE 50 tubing combined with a pressure transducer (TP-300T; Nihon Kohden). The aortic pressure was digitized and processed with a computer system (model PE-1000; Nihon Kohden).

In vivo assessment of cardiac functions.

Male mice were anesthetized with a mixture of ketamine (50 to 100 mg/kg) and xylazine (3 to 6 mg/kg) via intraperitoneal injection. The right carotid artery was isolated and cannulated with a 1.4 French Millar catheter connected to an amplifier (TCP-500; Millar Instrument) (22). After insertion of the catheter into the carotid artery, the catheter was advanced into aorta and then into the left ventricle (LV). The LV pressure was digitized and processed by the computer system used for TAC. Systolic or diastolic pressure was measured by cannulating the right carotid artery with heat-stretched PE 50 tubing combined with a pressure transducer (TP-300T; Nihon Kohden).

Echocardiography.

Male mice at 10 weeks of age were anesthetized with 2.5% avertin (8 μl/g), and echocardiography was performed by ultrasonography (SONOS-5500 instrument equipped with a 15-MHz linear transducer; Philips Medical Systems) (35). The heart was imaged in the two-dimensional parasternal short-axis view, and an M-mode echocardiogram of the midventricle was recorded at the level of the papillary muscles. Heart rate, anterior and posterior wall thickness, and end-diastolic and end-systolic internal dimensions of the LV were obtained from the M-mode image.

Western blots and in vitro kinase assay.

Total protein homogenates (5 to 50 μg/lane) were subjected to Western blot analysis with antibodies against p38α, p38β, p38δ, MKK3/6, ERK1, and JNK1 (Santa Cruz Biotechnology); p38γ (Upstate Biotechnology); phospho-MKK3/6, phospho-p38, phospho-JNK, and cleaved caspase-3 (Cell Signaling Technology); and phospho-ERK (Promega). The mitochondrion-rich and cytosolic fractions were isolated and subjected to Western blot analysis with antibodies against Bax (Santa Cruz Biotechnology) and Bcl-2 and cytochrome c (PharMingen) (15). Western blots were developed with ECL (ECL-plus kit or ECL-advance kit; Amersham). Quantification of signals was performed by densitometry of scanned autoradiographs with the aid of Scion image software. The activity of p38α was measured by an immune complex kinase assay. Immunoprecipitation of endogenous p38α was performed on 500 μg of myocardial extracts with anti-p38α antibody (Santa Cruz Biotechnology), and then immune complex kinase activity was measured with myelin basic protein (MBP) as a substrate.

Histological analyses.

The heart samples were arrested in diastole and immediately fixed with buffered 3.7% formalin, embedded in paraffin, and sectioned into 3-μm thickness. Hematoxylin and eosin or Masson-trichrome staining was performed on serial sections. Myocyte cross-sectional area was measured by tracing the outline of 100 to 200 myocytes in each section. Longitudinal cell length was estimated as the distance between cell edges identified by connexin 43 staining (Sigma-Aldrich). Phospho-JNK staining was performed with the Vectastain Elite ABC kit (Vector Laboratories) according to the manufacturer's instructions.

RNA dot blot analysis.

Total RNA was isolated from the ventricular apexes by using TRIzol reagent (Life Technologies). Quantitative assessment of atrial natriuretic factor (ANF), brain natriuretic peptide (BNP), α-skeletal actin, α-myosin heavy chain, collagen I, and collagen III was performed by RNA dot blot analysis as previously described (22, 32). Radiolabeled RNA dots were quantitated with Scion Image software, and the value of each dot was normalized to the glyceraldehyde-3-phosphatase dehydrogenase (GAPDH) signal.

Evaluation of apoptosis.

The terminal deoxynucleotidyl transferase-mediated dUTP-biotin nick end labeling (TUNEL) assay was used for paraffin-embedded heart sections, using an in situ apoptosis detection kit (Takara) according to the manufacturer's instructions. The number of TUNEL-positive nuclei was counted by examining the entire section with a ×40 objective. Triple staining with propidium iodide (Vector Laboratories), TUNEL, and anti-α-sarcomeric actin antibody (Sigma-Aldrich) was performed.

Isolation of ventricular myocytes and survival assay.

Mouse neonatal cardiomyocytes isolated from p38α CKO mice (35) were treated with isoproterenol for 48 h. By using a 3-(4,5-dimethlthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay-based kit (Cell Counting Kit-8; Dojindo), numbers of surviving cells were determined in triplicate.

Osmotic minipump infusion.

Osmotic minipumps (model 1007D; Alzet) were used to administer isoproterenol at a dose of 7.5 mg/kg/day for 2 days. The pumps were implanted into 10-week-old male mice subcutaneously. Pumps were removed 24 h before echocardiography and histological examination.

Statistical analysis.

Results are shown as means ± standard errors of the means. Paired data were evaluated by Student's t test. A one-way analysis of variance with Bonferroni's post hoc test or repeated-measures analysis of variance was used for multiple comparisons. A P value of <0.05 was considered statistically significant.

RESULTS

Generation of p38α CKO mice.

To obtain p38α CKO mice, conditional inactivation of the p38α gene was achieved by the inserting loxP sites cloned 5′ and 3′ of the exon encoding amino acids 40 to 83 of p38α, which is a part of the catalytic domain that includes the ATP binding loop (Fig. 1A). Selection cassettes comprising a neomycin resistance gene (neo) for positive selection and a diphtheria toxin gene for negative selection were positioned between two loxP sites, upstream of the floxed exon and at the 3′ end of the targeting vector, respectively (Fig. 1A). Homogenous recombinants were identified by PCR and Southern blotting (Fig. 1B). Mice which harbored the p38α-floxed allele and the selection marker gene, that is, the PGK-neo cassette, were crossbred with EIIa-cre mice (17) to obtain heterozygous p38α-floxed mice without the PGK-neo cassette as well as conventional heterozygous p38α null mice (p38α^+/−). The homozygous p38α-floxed mice (p38α^flox/flox) appeared normal and were externally indistinguishable from littermates of other genotypes, and no difference was observed in the expression of p38α between the p38α^flox/flox mice and control wild-type mice (data not shown). The p38α^−/− mice generated by type I deletion were lethal at embryonic day 10.5, as also previously reported (1, 30) (data not shown). To selectively disrupt the p38α gene in the myocardium, the p38α^flox/flox mice were crossed with mice expressing the Cre recombinase under the control of the α-myosin heavy-chain promoter (α-MHCCre mice). In this manner, we bred p38α^flox/flox:α-MHCCre(−) mice with p38α^flox/+:α-MHCCre(+) mice, resulting in p38α^flox/flox:α-MHCCre(+) mice, which feature p38α CKO mice. Immunoblotting demonstrated an approximately 75% reduction of p38α protein in the hearts of p38α CKO mice (Fig. 1C), with no detectable reduction in other tissues such as liver, kidney, and spleen (data not shown). In addition, there was almost no difference in cardiac p38α levels between wild-type and p38α^flox/flox:α-MHCCre(−) mice (data not shown), so we used p38α^flox/flox:α-MHCCre(−) littermates as controls (CTL). When adult cardiomyocytes were isolated for a partially purified single cell preparation, the p38α protein level in p38α CKO cardiomyocytes was approximately 10% of that in CTL (Fig. 1C). An in vitro kinase assay was performed 30 min after acute injection of phenylephrine (10 mg/kg), resulting in a 1.63-fold activation of p38 kinase activity in the hearts of CTL mice (Fig. 1D). In contrast, p38α CKO mice showed a significant reduction in p38α kinase activity, both at baseline and after phenylephrine stimulation, compared with CTL mice. There was no compensatory up-regulation of other p38 isoforms such as p38β, p38γ, and p38δ (Fig. 1C), nor were there any differences between p38α CKO and CTL hearts in the basal protein levels of MKK3/6, ERK, and JNK (Fig. 1C).

FIG. 1.

Targeted modification of the p38α gene. (A) Schematic structures of genomic p38α sequences, the targeting construct, and targeted allele. The closed boxes and triangles represent the floxed p38α exon and loxP site, respectively. The targeting construct includes the PGK-neomycin resistance gene cassette (Neo) flanked by loxP sites and the diphtheria toxin (DTA). The small arrows correspond to the primer sequences for PCR, and the open boxes (5′ probe) correspond to the sequence used for Southern blotting analysis in panel B. Restriction sites: N, NcoI; X, XhoI; H, HindIII; Bg, BglII; Sm, SmaI; Ba, BamHI; Sc, SacII; Sl, SalI. (B) Genomic analysis of ES cells. Genomic DNA was isolated from five ES cell clones identified to be homologous recombinants, digested with XhoI, and analyzed by Southern blotting with the 5′ probe. (C) Protein expression of the MAP kinase family in p38α CKO heart. Total cell lysates from p38α CKO and CTL hearts were examined for p38 isoforms, MKK3/6, ERK, and JNK with immunoblotting. The expression of p38α protein in isolated cardiomyocytes (CM) was also examined. (D) p38α activity in vitro kinase assay from phosphate-buffered saline- and phenylephrine (PE)-injected p38α CKO or CTL mice. Upper panel, phosphorylation of MBP was monitored by in vitro kinase assay with p38α-specific antibody. Lower panel, quantification of phospho-MBP protein levels by densitometric analysis. Open and closed bars represent phosphate-buffered saline- and PE-injected mice, respectively (n = 3). *, P < 0.05 versus corresponding phosphate-buffered saline-injected mice. #, P < 0.05 versus PE-injected CTL. Error bars indicate standard errors of the means.

p38α CKO mice exhibit normal cardiac structure and function.

Expected Mendelian ratios of p38α^flox/flox:α-MHCCre(+), p38α^flox/+:α-MHCCre(+), p38α^flox/flox:α-MHCCre(−), and p38α^flox/+:α-MHCCre(−) mice were observed for the offspring of p38α^flox/flox:α-MHCCre(−) and p38α^flox/+:α-MHCCre(+) mice, indicating no significant embryonic lethality. The p38α CKO mice were born normally and appeared externally indistinguishable from CTL littermates. They developed to adulthood, were fertile, and had a normal life span. There were no differences in body weight, heart weight, and LV or right ventricle weight between p38α CKO and CTL mice (Table 1). The p38α CKO hearts showed no evidence of any of the cardiac morphological defects observed in conventional p38α^−/− embryos (Fig. 2A), nor did histological examination of the hearts demonstrate any myofibrillar disarray, necrosis, or ventricular fibrosis (Fig. 3A). Finally, there was no significant difference in the myocyte cross-sectional areas of p38α CKO and CTL mice (156.6 ± 3.8 μm² for p38α CKO mice and 150.5 ± 4.4 μm² for CTL mice; n = 3).

TABLE 1.

Physiological parameters in p38α CKO mice

Parameter (unit)a	Value (mean ± SEM) for:
	p38α CKO mice (n = 9)	CTL mice (n = 10)
Body wt (g)	27.5 ± 0.8	28.9 ± 0.6
Heart wt (mg)	135.1 ± 7.3	136.0 ± 4.6
Left ventricle wt (mg)	97.1 ± 4.4	97.1 ± 2.7
Right ventricle wt (mg)	23.4 ± 1.5	24.8 ± 1.3
Tibia length (mm)	18.0 ± 0.2	18.2 ± 0.9
Systolic blood pressure (mm Hg)	104.6 ± 7.6	105.8 ± 5.0
Diastolic blood pressure (mm Hg)	82.0 ± 6.3	85.9 ± 3.5
LV end-diastolic pressure (mm Hg)	2.76 ± 0.64	1.90 ± 0.53
LV dP/dtmax (mm Hg/s)	6328.6 ± 753.6	6312.5 ± 400.6
LV dP/dtmin (mm Hg/s)	−4571.4 ± 530.4	−4825.0 ± 378.8
Heart rate (beats/min)	373.6 ± 22.8	375.0 ± 19.8

^aLV dP/dt_max and dP/dt_min are the maximum rates of pressure development during contraction and relaxation, respectively.

FIG. 2.

Morphological and functional consequences of pressure overload in p38α CKO heart. (A) Macroscopic hematoxylin-eosin-stained histological sections of hearts from CTL and p38α CKO mice before and 7 days after TAC or sham operation. (B) Transthoracic M-mode echocardiographic tracings from a p38α CKO mouse and a CTL mouse before and 1 week after TAC or sham operation. (C) Changes in the echocardiographic parameters end-diastolic (LVDd) and end-systolic (LVDs) LV diameters and FS by TAC. Echocardiography was sequentially performed on mice 2 days before operation (Pre) and 7 days after operation (Post). Closed boxes, closed circles, open boxes, and open circles represent TAC-operated p38α CKO (n = 9), TAC-operated CTL (n = 10), sham-operated p38α CKO (n = 5) and sham-operated CTL (n = 6) mice, respectively. *, P < 0.05 versus corresponding preoperation. Error bars indicate standard errors of the means.

FIG. 3.

Histological analysis of p38α CKO mice after TAC. (A) Microscopic Masson-trichrome-stained histological sections (magnification, ×200) of hearts from CTL and p38α CKO mice before and 1 week after TAC or sham operation. (B) The fibrotic lesions in LV myocardium were measured by image analysis software. Open and closed bars represent sham-operated and TAC-operated mice, respectively (n = 3). *, P < 0.05 versus all other groups, including corresponding sham-operated mice. Error bars indicate standard errors of the means. (C) mRNA expression of collagen I or III was evaluated 1 week after TAC by dot blot analysis.

To determine whether cardiac-specific p38α knockout would affect the cardiac function, cardiac performance was evaluated by means of echocardiography and cardiac catheterization in 10-week-old mice. Echocardiographic studies showed that there were no significant differences in LV end-diastolic and end-systolic dimensions, septal wall thickness, posterior wall thickness, or fractional shortening (FS) between p38α CKO and CTL mice (Fig. 2B and C). Furthermore, hemodynamic data did not indicate any differences between the maximum first derivative of LV pressure (LV dP/dt_max) and the minimum first derivative of LV pressure (LV dP/dt_min) of p38α CKO and CTL mice (Table 1). These findings demonstrated that p38α CKO mice had normal global cardiac structure and functioning.

Need for p38α in the adaptive compensatory response to biomechanical stress caused by pressure overload.

It has been reported that the p38 pathway is activated in mouse hearts that have been exposed to pressure overload following TAC (33). In the TAC model, banding of the transverse aorta in mice leads to hyperfunctional hypertrophy after 1 week without any signs of heart failure (27). Activation of p38 activities during the development of hypertrophy suggested that p38 might have a function in the pathway mediating such cardiac hypertrophy (33). To determine whether p38α is indeed involved in the biomechanical stress response in hearts, we used echocardiography to evaluate LV functioning in p38α CKO mice 1 week after pressure overload by TAC. One week after TAC, 30% of p38α CKO mice had died (n = 60), whereas 92% of CTL mice were still alive at this time (n = 48). The mechanical stress produced during TAC was estimated by measuring in vivo transstenotic pressure gradients 7 days after TAC. Although the TAC procedure caused a significant increase in the pressure gradients between the two carotid arteries, there was no significant difference in pressure gradients between p38α CKO and CTL mice (40.3 ± 4.1 mm Hg for p38α CKO mice [n = 9] and 44.0 ± 2.1 mm Hg for CTL mice [n = 10]). Sham-operated p38α CKO and CTL mice showed no differences in cardiac structural and functional characteristics compared to their nonoperated control counterparts (Fig. 2 and 3). Although wall thickness was similar to that of controls, the LV end-diastolic and end-systolic diameters in p38α CKO mice showed a comparatively significant increase (Fig. 2B and C). Cardiac contractility as assessed by FS was significantly reduced in p38α CKO mice (Fig. 2C). One week after TAC, enlargement of p38α CKO hearts compared to CTL hearts had become evident (Fig. 2A). In addition, heart weight, LV weight, and the average ratio of LV weight to tibial length or body weight of p38α CKO mice after TAC were significantly greater than those of their sham-operated control counterparts (Table 2 and Fig. 4A). There was no significant difference between p38α CKO and CTL mice in the LV weight-to-body weight ratio or in the LV weight-to-tibia length ratio after TAC. Furthermore, there was no significant difference between the myocyte cross-sectional areas of p38α CKO and CTL mice (252.3 ± 6.9 μm² for p38α CKO mice and 246.9 ± 4.9 μm² for CTL mice; n = 3) (Fig. 4B). The mean cell area of the myocytes isolated from p38α CKO mice was not significantly different from that of the myocytes isolated from CTL mice 1 week after either sham or TAC operation (2,504.7 ± 98.0 μm² for sham-operated CTL mice, 2,623.9 ± 77.3 μm² for sham-operated p38α CKO mice, 3,125.3 ± 55.6 μm² for TAC-operated CTL mice, and 3,333.2 ± 79.2 μm² for TAC-operated p38α CKO mice; n = 3) (Fig. 4C). After TAC, the levels of ANF, BNP, and α-skeletal actin mRNAs had significantly increased in p38α CKO and CTL mice (Fig. 4D), while the levels of α-myosin heavy-chain mRNA had decreased. These findings suggest that biochemical hypertrophic responses were not impaired in p38α CKO mice and that cardiac dysfunction resulted in enhanced BNP induction in p38α CKO mice compared with CTL mice. Histological analysis demonstrated that intermuscular as well as perivascular fibrosis was observed in p38α CKO hearts, whereas there was only slight perivascular fibrosis in CTL hearts (Fig. 3A and B). These histological findings were supported by a significant increase in the expression levels of collagen I and III mRNAs in p38α CKO hearts after TAC compared with the levels in CTL mice (Fig. 3C).

TABLE 2.

Physiological parameters in p38α CKO mice following pressure overload

Parameter (unit)a	Value (mean ± SEM) after the indicated operation for:
	p38α CKO mice		CTL mice
	Sham (n = 5)	TACb (n = 9)	Sham (n = 6)	TAC (n = 10)
BW (g)	27.3 ± 1.0	26.4 ± 0.6	28.4 ± 0.7	26.0 ± 0.6
HW (mg)	129.4 ± 5.5	191.4 ± 7.9*	127.6 ± 3.6	168.7 ± 8.0*
LVW (mg)	94.6 ± 3.5	134.5 ± 3.0*	91.4 ± 2.2	126.8 ± 4.9*
RVW (mg)	22.2 ± 1.6	27.2 ± 2.3	23.3 ± 1.1	21.9 ± 1.2
TL (mm)	17.9 ± 0.2	18.1 ± 0.1	18.0 ± 0.2	18.0 ± 0.1
HW/BW (mg/g)	4.7 ± 0.1	7.3 ± 0.4*	4.5 ± 0.1	6.5 ± 0.4*
LVW/BW (mg/g)	3.5 ± 0.1	5.1 ± 0.2*	3.2 ± 0.1	4.9 ± 0.2*
HW/TL (mg/mm)	7.2 ± 0.3	10.6 ± 0.4*	7.1 ± 0.3	9.4 ± 0.5*
LVW/TL (mg/mm)	5.3 ± 0.2	7.4 ± 0.1*	5.1 ± 0.1	7.1 ± 0.3*

^aHW, heart weight; BW, body weight; LVW, left ventricle weight; RVW, right ventricle weight; TL, tibia length.

^b*, P < 0.05 compared to corresponding sham-operated mice.

FIG. 4.

Cardiac hypertrophy by pressure overload in p38α CKO mice. TAC or sham operation was applied in either CTL or p38α CKO mice. Open and closed bars represent sham-operated and TAC-operated mice, respectively. *, P < 0.05 versus corresponding sham-operated mice. #, P < 0.05 versus TAC-operated CTL. Error bars indicate standard errors of the means. (A) The LV weight (LVW) (milligrams)/body weight (BW) (grams) ratio (left panel) or the LVW (milligrams)/tibia length (TL) (millimeters) ratio (right panel) was obtained 1 week after TAC. TAC-operated p38α CKO. n = 9; TAC-operated CTL, n = 10; sham-operated p38α CKO, n = 5; and sham-operated CTL, n = 6. (B) Cardiomyocyte cross-sectional areas. Myocyte cross-sectional area was measured by tracing the outline of 100 to 200 myocytes in each section (n = 3). Longitudinal cell length was estimated as the distance between cell edges identified by connexin 43 staining. (C) Cell surface area (left panel) and cell length (right panel). Cardiomyocytes were isolated from mouse hearts (n = 3) 1 week after TAC. Myocyte surface area was measured by tracing the outline of 100 to 200 myocytes. (D) mRNA expression of ANF, BNP, α-skeletal actin (αSkA), or α-myosin heavy chain (αMHC) 1 week after TAC was evaluated by dot blot analysis (n = 4). Densitometric analyses were performed. The mean of the ANF/GAPDH, BNP/GAPDH, αSkA/GAPDH, or αMHC/GAPDH ratio in CTL mice subjected to sham operation was expressed as 1.

Mechanical stress leads to apoptosis in p38α CKO hearts.

TUNEL was used to examine whether the reduction in LV function after TAC was related to apoptosis (Fig. 5A). TUNEL-positive cells were identified as cardiac myocytes by anti-α-sarcomeric actin staining (Fig. 5A) and showed the condensed chromatin and fragmented nuclei which are morphological characteristics of apoptosis (Fig. 5B). The number of TUNEL-positive cells in p38α CKO hearts was 3.7 times that in CTL hearts (Fig. 5C). We also examined the activation of caspase-3 by using anti-cleaved-caspase-3 antibody (Fig. 5D). Although the 17-kDa band contains nonspecific protein, we detected an increase in the amount of 19-kDa cleaved caspase-3 in p38α CKO hearts after TAC.

FIG. 5.

Apoptosis in p38α CKO heart after TAC. (A) Confocal analysis of p38α CKO ventricular myocardium 1 week after TAC. Triple staining (propidium iodide, TUNEL, and anti-α-sarcomeric actin antibody) was performed. Staining for propidium iodide and anti-α-sarcomeric actin antibody is shown in red, and that for TUNEL is in green. In the overlay image, a nucleus stained by both TUNEL and propidium iodide is shown in yellow. (B) Morphology of nuclei in TUNEL-positive cells. (C) Number of TUNEL-positive cells in p38α CKO hearts compared to that in CTL 7 days after TAC. Open and closed bars represent sham (n = 3)- and TAC (n = 6)-operated mice, respectively. *, P < 0.05 versus all other groups. Error bars indicate standard errors of the means. (D) Analysis of cleaved caspase-3 protein levels in cardiac tissue 7 days after TAC or sham operation. Immunoblotting was performed with anti-cleaved caspase-3 antibody. (E) Analysis of cytochrome (Cyt) c protein levels in the cytosolic fraction of LV 7 days after TAC or sham operation. Immunoblotting was performed with anti-Cyt c antibody. Densitometric analyses were performed (right panel). TAC-operated p38α CKO, n = 6; TAC-operated CTL, n = 6; sham-operated p38α CKO, n = 4; and sham-operated CTL, n = 4. *, P < 0.05 versus all other groups. (F) Immunoblot analysis of apoptosis-related proteins in the mitochondrial fraction from LV myocardium after TAC or sham operation. Densitometric analyses of TAC-operated mice were performed, and the ratio of Bax to Bcl-2 is shown (right panel). TAC-operated p38α CKO, n = 4; TAC-operated CTL, n = 3. *, P < 0.05 versus TAC-operated CTL.

To determine whether mitochondrion-mediated apoptosis is involved in aortic-banded p38α CKO, we examined the release of cytochrome c into cytosol from the mitochondria by using an anti-cytochrome c antibody (Fig. 5E). Immunoblot analysis indicated 1.6- and 3.8-fold increases in TAC-operated CTL and p38α CKO mice compared to sham-operated CTL mice, respectively (Fig. 5E). We next examined the expression of apoptosis-related mitochondrion proteins that are known to promote or inhibit apoptosis. Immunoblotting showed increased levels of Bax and Bcl-2 in TAC-operated p38α CKO or CTL mice compared with those in sham-operated controls (Fig. 5F). The ratio of Bax to Bcl-2 protein (Bax/Bcl-2 ratio), which determines survival or death after an apoptotic stimulus (24), was 1.74 times higher in TAC-operated p38α CKO mice than that in TAC-operated CTL mice (Fig. 5F). These findings suggest the involvement of a mitochondrial death mechanism in apoptosis in p38α CKO mice.

Activation of mitogen-activated protein kinase family in response to pressure overload.

We evaluated activation of JNK, ERK, and MKK3/6 in CTL and p38α CKO hearts subjected to sham operation or TAC for 7 days (Fig. 6A and B). Immunoblotting indicated increases in phosphorylation levels of JNK, ERK, and MKK3/6 in TAC-operated CTL mice compared with those in sham-operated CTL. In p38α CKO, the phosphorylation levels of JNK, ERK, and MKK3/6 increased in response to pressure overload. There was no difference in the phosphorylation level of ERK between TAC-operated p38α CKO and CTL mice. However, the levels of phospho-JNK and phospho-MKK3/6 in p38α CKO mice were significantly higher than those in CTL mice following TAC. We then examined localization of JNK activation in hearts by using immunofluorescence staining. We detected phospho-JNK staining only in noncardiomyocytes and not in cardiomyocytes in p38α CKO hearts after TAC. We detected no difference in the levels of the staining in cardiomyocytes and noncardiomyocytes in CTL hearts after TAC (Fig. 6C).

FIG. 6.

Western blot analysis of JNK, ERK, or MKK3/6 phosphorylation and phospho-JNK staining in the hearts of p38α CKO or CTL mice. Ventricular protein lysates were obtained 7 days after TAC or sham operation. (A) For each group, the phosphorylation level of the protein was estimated by using antiphospho (phos) antibody. Total protein levels were also examined with their specific antibodies. n.s., nonspecific bands. (B) Densitometric analyses were performed. Open and closed bars represent sham-operated and TAC-operated mice, respectively; n = 4 for each group. *, P < 0.05 versus corresponding sham-operated mice. #, P < 0.05 versus TAC-operated CTL. Error bars indicate standard errors of the means. (C) Phospho-JNK staining of p38α CKO or CTL hearts 7 days after TAC.

p38α-deficient cardiomyocytes are more susceptible to β-adrenergic stress.

We next examined the effect of p38α deficiency on isoproterenol-induced cell death. p38α CKO mice were infused via osmotic pumps for 2 days with saline or isoproterenol. Infusion of isoproterenol into p38α CKO mice led to a significant reduction in heart function in conjunction with a reduction in FS (14.1% ± 2.6% [n = 3] for p38α CKO mice and 48.1% ± 2.6% [n = 3] for CTL mice) (Fig. 7A and B). The isoproterenol-treated p38α CKO hearts had become larger than those of CTL, and the number of TUNEL-positive cells was significantly higher in p38α CKO hearts than in CTL hearts (Fig. 7C). Cardiomyocytes isolated from the hearts of neonatal p38α CKO or CTL mice were then exposed to isoproterenol (0 to 40 μM) for 2 days under identical culture conditions (Fig. 7D). p38α CKO cardiomyocytes were more susceptible to isoproterenol than CTL cells as determined by an MTT assay.

FIG. 7.

Isoproterenol-induced cardiomyocyte death. (A) Representative transthoracic M-mode echocardiographic tracings from isoproterenol-treated p38α CKO (p38α CKO-ISO) and CTL (CTL-ISO) mice. Mice were treated with isoproterenol for 2 days. (B) Changes in the echocardiographic parameters LVDd and FS after isoproterenol treatment. Open and closed bars represent saline-treated and isoproterenol-treated mice, respectively. Saline-treated p38α CKO, n = 4; isoproterenol-treated p38α CKO, n = 3; saline-treated CTL, n = 3; isoproterenol-treated CTL, n = 3. *, P < 0.05 versus all other groups, including saline-treated CTL. Error bars indicate standard errors of the means. (C) Relative number of TUNEL-positive cells in p38α CKO hearts compared to that in CTL hearts. Open and closed bars represent saline-treated and isoproterenol-treated mice, respectively. Saline-treated p38α CKO, n = 3; isoproterenol-treated p38α CKO, n = 5; saline-treated CTL, n = 3; isoproterenol-treated CTL, n = 3. *, P < 0.05 versus all other groups, including saline-treated CTL. (D) Cell viability was assessed by using a Cell Counting Kit-8 (n = 3). Viability of cells is expressed as the percentage of viability of cells in the absence of isoproterenol. Mouse neonatal cardiomyocytes isolated from p38α CKO (closed bars) or CTL (open bars) mice were incubated with various concentrations of isoproterenol for 48 h. *, P < 0.05 versus corresponding CTL.

DISCUSSION

p38α plays an essential role in cardiomyocyte survival but not in cardiac hypertrophic growth in response to pressure overload.

A major finding of our study is that cardiac hypertrophy by pressure overload is unaffected despite dramatic down-regulation of p38α in p38α CKO hearts. This suggests that (i) endogenous p38α does not play a part in the regulation of cardiomyocyte hypertrophic growth, (ii) the small amount of retained p38α might still play some role in cardiac hypertrophy, or (iii) parallel signaling mechanisms are sufficient to permit cardiac hypertrophy to develop normally in response to pressure overload.

The p38α^+/+:α-MHCCre(+) and p38α^flox/+:α-MHCCre(+) mice did not show evidence of cardiac dysfunction or heart dilatation following pressure overload as observed in p38α CKO mice (data not shown). This eliminates the possibility that overexpression of Cre recombinase is a cause of pressure overload-induced cardiac dysfunction and heart dilatation. To ensure that the resultant anatomical and functional alterations were not in part due to positional effects, that is, the position where the Cre recombinase transgene was integrated into α-MHCCre mice, we crossed p38α^flox/flox mice with mice expressing the Cre recombinase under the control of the myosin light-chain 2v (MLC2v) promoter (4). The resultant p38α^flox/flox:MLC2v-Cre^+/− mice exhibited pressure overload-induced heart dilatation and dysfunction as also seen in p38α CKO mice, and no compensatory activation of ERK or JNK in p38α CKO cardiomyocytes in response to pressure overload was observed. Thus, we can conclude that p38α is part of an essential stress-activated myocyte survival pathway.

The study presented here showed a correlation between an increase in apoptosis and the development of cardiac dysfunction and dilatation in p38α CKO mice in response to pressure overload or isoproterenol. Pressure overload and/or the associated neurohumoral changes such as those reflected in catecholamine, angiotensin II, cytokines, and reactive oxygen species lead to apoptosis in TAC-operated p38α CKO mice. Our study suggests that pressure overload-induced apoptosis might be a critical event in the transition to heart failure in p38α CKO mice. It is possible, however, that apoptosis is not a direct cause of the development of the heart failure but a secondary phenomenon. In the decompensated stage, cardiomyocyte remodeling in the myocardium is to be expected, as a result of lengthening of cardiomyocytes together with side-to-side slippage. However, there was no difference in the longitudinal length of cardiomyocytes in ventricular cellular organization between p38α CKO and CTL mice following TAC. This indicates that side-to-side slippage rather than lengthening of cardiomyocytes contributes to ventricular remodeling in p38α CKO mice.

Exactly how p38α signaling pathways modify the rate of apoptosis has not been established. p38 has been shown to prevent apoptosis by down-regulating NF-κB activity and Fas expression (12) and by means of phosphorylation and induction of αB-crystallin (11). p38 has also been shown to activate MAP kinase-activated protein kinases (28), which phosphorylate several proteins, including hsp27 (5) and CREB and ATF-1 (31). These MAP kinase-activated protein kinase-activated pathways may mediate the p38-induced protection from apoptosis. It has been reported that the Janus kinase (JAK)/signal transducers and activators of transcription (STAT) signaling pathway plays an important role in the transition to heart failure in response to pressure overload (10, 36). Since we observed no difference between the activation levels of STAT3 in p38α CKO and CTL mice after TAC (data not shown), however, the possibility that the JAK/STAT signaling pathway is involved in p38α-mediated apoptosis is eliminated. The mitochondria play a crucial role in the apoptotic death of mammalian cells by releasing apoptogenic proteins, including cytochrome c. Upon receipt of an apoptotic signal, Bax translocates from the cytosol to the mitochondria, which in turn induces cytochrome c release, which is inhibited by Bcl-2 on the mitochondrial membrane. The Bax-to-Bcl-2 ratio is an important survival-or-death indicator for cells and is influenced by competitive dimerization between the two proteins (25). This suggests that p38α signaling might coordinate expression of these genes in relation to mitochondrion-mediated apoptosis. The downstream mechanisms whereby p38α signaling regulates mitochondrion-mediated apoptosis warrant further investigation. The p38α CKO mice generated by us will undoubtedly prove to be useful for the identification of downstream targets of p38α in order to enhance cardiomyocyte survival.

Agreement and disagreement between our study and previous studies.

In vitro studies using neonatal cardiomyocytes indicate that p38 is an important regulator of cardiac hypertrophy (7, 23, 33, 37). Wang et al. (33) used isoform-specific adenovirus vectors to identify distinct roles of p38α and p38β in neonatal cardiomyocytes. The in vitro results indicate that activation of p38α appears to induce cell death and possibly suppress hypertrophy and contrast with our findings that p38α plays an essential part in myocyte survival but not in cardiac hypertrophy in response to pressure overload. Transgenic hearts expressing activated MKK3 or MKK6, which were found to induce cardiomyocyte hypertrophy in cultured neonatal cardiomyocytes, did not develop cardiac hypertrophy but exhibited marked interstitial fibrosis without signs of apoptosis (19). The growth response of cultured neonatal cardiomyocytes is not identical to that of the adult heart, since neonatal cardiomyocytes are immature in terms of sarcomere organization and organization of signaling complexes associated with the Z-disk and T-tubule network. These considerations suggest that p38 activation does not induce the hypertrophic growth of adult heart but leads to stimulation of cytoprotective mechanisms, in contrast to its effect on cultured neonatal cardiomyocytes. However, we cannot exclude a possibility that the discrepancy regarding a role of p38α in cardiac hypertrophy could be due to the different hypertrophic stimuli in these studies; p38α may play a role in some hypertophic responses but not in others.

In support of our results, Zhang et al. (38) reported that dnp38α mice with a Black Swiss background showed no hypertrophic phenotypes at baseline. Pressure overload caused a similar level of hypertrophy in dnp38α mice as in controls. In contrast to these findings, Braz et al. (3) demonstrated that dnp38α mice with an FVB/N background developed cardiac hypertrophy at baseline and showed enhanced development of cardiac hypertrophy following aortic banding, suggesting an antihypertrophic function of p38α. Our p38α CKO mice and the dnp38α transgenic mice used by Braz et al. (3) showed essentially no increase in cardiac p38α activity after stimulation. The developmental compensation by increased expression of p38 isoforms other than p38α may explain the apparent discrepancy between our results and those of Braz et al. (3). However, we did not observe any changes in the expression levels of p38 isoforms compared with those in controls. While dnp38α may block signaling to the effectors which p38α, p38β, p38γ, and p38δ have in common (3), our knockout strategy inactivates only p38α signaling. This may be the cause of the discrepancies mentioned earlier. Another possible explanation of the apparent discrepancies is that the secondary changes in protein expression or intracellular signaling caused by overexpression of dnp38α and deficiency of the p38α gene may affect resultant phenotypes. For example, the dnp38α transgenic mice used by Braz et al. (3) exhibited up-regulated Bcl-2 expression at baseline (14), but our p38α CKO mice did not. Alteration of the protein expression profile is a limitation of studies using gene manipulation or even pharmacological agents.

Compelling evidence supports both protective and promoting roles of p38 in the regulation of cell death in various cells (12, 24, 34), including cardiomyocytes (8, 11, 21, 33, 37). The different effects of pharmacological p38 inhibition on apoptosis in various cell types may reflect heterogeneity in the expression and/or activation of p38 isoforms. Another possibility is that p38α has a dual role in terms of cell death. This hypothesis is supported by the previous report by Molkentin's group, in which dnp38α transgenic mice showed a reduction in cardiac injury and cell death following ischemia-reperfusion (14), but those mice showed reduced cardiac contractility at baseline (3). Our results reported here indicate that the function of p38α is cytoprotective in response to stress, but conventional heterozygous p38α knockout mice exhibited resistance to ischemia-reperfusion insult (26). Pressure overload constitutes a sublethal and long-lasting stress, whereas ischemia-reperfusion insult constitutes an acute lethal stress. Although the effect of cardiac-specific p38α deficiency on ischemia-reperfusion injury needs to be examined further, various aspects of stress, such as time, place, quality, and quantity, may determine the precise function of p38α. p38α^flox/+:α-MHCCre(+) mice or p38α^+/− mice after TAC did not exhibit any signs of cardiac dysfunction or heart dilatation, suggesting that the amount of p38 is not related to the apparently opposite roles of p38 in the stress response. Further investigation will be necessary for precise characterization of the molecular mechanism which determines the role of p38α in response to stress.