Constitutively active MEK1 rescues cardiac dysfunction caused by overexpressed GSK-3α during aging and hemodynamic pressure overload

Yasuhiro Maejima, Jonathan Galeotti, Jeffery D. Molkentin, Junichi Sadoshima, Peiyong Zhai

Abstract

Expression of GSK-3α is increased in aging hearts and those subjected to hemodynamic overload. Overexpressed GSK-3α inhibits ERK and enhances pressure overload (PO)-induced cardiac dysfunction. We studied whether suppression of the MEK1/ERK pathway contributes to cardiac responses induced by overexpressed GSK-3α using constitutively active MEK1 (CA-MEK1)/GSK-3α bigenic mice (bigenic mice), which were obtained by crossing cardiac-specific GSK-3α transgenic mice (Tg-GSK) and cardiac-specific CA-MEK1 transgenic mice (Tg-MEK1). The suppression of ERK phosphorylation observed in Tg-GSK was eliminated in bigenic mice. At 12 mo, left ventricular (LV) weight/tibia length, LV weight/body weight, and cardiac myocyte size were significantly smaller in Tg-GSK than in nontransgenic mice (NTg), but were not significantly different between Tg-MEK1 and bigenic mice. The LV ejection fraction (LVEF), fractional shortening (FS), and change in pressure over time were significantly lower in Tg-GSK than in NTg, but were not significantly different between bigenic mice and Tg-MEK1. The increase in apoptosis in Tg-GSK was abolished in bigenic mice, although the increase in fibrosis was not. After PO, the decrease in cardiac hypertrophy and the enhancement of apoptosis seen in Tg-GSK were abrogated in bigenic mice. After PO, the LVEF and FS were significantly reduced in Tg-GSK compared with its sham, but not in NTg, Tg-MEK1, or bigenic mice compared with their respective shams. There was no significant difference in LVEF and FS between bigenic mice and Tg-MEK1 after PO. In conclusion, inhibition of the MEK1/ERK pathway mediates the hypertrophy suppression and cardiac dysfunction caused by GSK-3α overexpression in cardiac myocytes.

  • cardiac hypertrophy
  • signal transduction
  • apoptosis

gsk-3 is a serine/threonine kinase that has versatile biological functions, including cellular growth, death, and metabolism (6, 7, 17, 28). GSK-3 is unique in that it is active under unstimulated conditions. Upstream stimulating signaling, such as activation of protein kinase B/Akt, phosphorylates GSK-3 and inhibits its function. Many targets of GSK-3 are phosphorylated and suppressed by it. Thus inhibition of GSK-3 activates its targets. GSK-3 has two isoforms, GSK-3α and GSK-3β, which have many different functions despite some similarities. Both GSK-3α and GSK-3β exist in the heart. They are ubiquitously expressed and presumably present in both cardiac myocytes and other cells in the heart. Most previous works primarily focused on the function of GSK-3β, which phosphorylates a variety of mediators of cardiac hypertrophy (2, 3, 8, 9, 21) and therefore plays an important role in negatively regulating cardiac hypertrophy (1, 20, 24). Recently, accumulating lines of evidence have shown that GSK-3α plays a critical role in regulating cardiac physiological growth and cardiac responses to pressure overload (PO) (18, 32, 34). The underlying mechanisms of the phenotypes appear to be diverse, varying from study to study, depending on the focus of each. One study showed that GSK-3α positively regulates β-adrenergic responsiveness (34). We have shown that wild-type GSK-3α, when overexpressed in cardiac myocytes, inhibits ERK activation (32) and that a constitutively active mutant of GSK-3α, systemically knocked-in, suppresses proliferation in response to PO (18). We have also reported that MEK/ERK inhibition abolishes the effects of GSK-3α knockdown in cultured myocytes in vitro (32). However, it is unknown whether MEK/ERK activation can counteract the adverse effects of GSK-3α overexpression in the mouse heart in vivo. In the present study, we used bigenic mice with cardiac-specific overexpression of both GSK-3α and constitutively active MEK1 (CA-MEK1) to further test our hypothesis that ERK inhibition by GSK-3α plays an important role in regulating cardiac hypertrophy and function.

MATERIALS AND METHODS

Mice.

Generation of transgenic mice with cardiac-specific overexpression of GSK-3α (Tg-GSK) (32) and transgenic mice with cardiac-specific overexpression of CA-MEK1 (Tg-MEK1) (4) has been previously reported. Bigenic mice with cardiac-specific overexpression of both GSK-3α and CA-MEK1 (Bigenic) were obtained by crossing Tg-GSK and Tg-MEK1. All experiments involving animals were approved by the Institutional Animal Care and Use Committee of New Jersey Medical School. All investigations conformed to the National Institutes of Health's Guide for the Care and Use of Laboratory Animals.

Transverse aortic constriction.

Transverse aortic constriction (TAC) was performed according to the method published previously (10) to induce PO in the left ventricle (LV). Briefly, mice were anesthetized with pentobarbital (60 mg/kg ip), intubated, and ventilated with a tidal volume of 0.2–0.3 ml and a respiratory rate of 110 breaths/min using a Harvard rodent ventilator (model 683; Harvard Apparatus, Holliston, MA). The chest cavity was entered through the second intercostal space, and the aorta was isolated. A 7-0 prolene suture was placed around the aorta between the innominate artery and left carotid artery. A 27-gauge needle was tied onto the aorta and later removed. The chest was then closed in layers.

Echocardiography.

Mice were anesthetized using 12 μl/g body weight of 2.5% avertin (Sigma), the chest was shaved, the animal was placed on a warm pad, and echocardiography was performed using ultrasonography (Acuson Sequoia C256; Siemens Medical Solutions, Malvern, PA) as previously described (33). A 13-MHz linear ultrasound transducer was used. Electrode needles were connected to each limb, and the electrocardiogram was simultaneously recorded. Two-dimension (short-axis)-guided M-mode measurements (at the level of the papillary muscles) of LV internal diameter were taken from 3 or more beats and averaged. Left ventricular (LV) end-diastolic dimension (LVEDD) was measured at the time of the apparent maximal LV diastolic dimension, whereas LV end-systolic dimension (LVESD) was measured at the time of the most anterior systolic excursion of the posterior wall. Systolic function was estimated from LV dimensions by the cubed method as LV ejection fraction (LVEF) as follows: LVEF = (LVEDD3 − LVESD3)/LVEDD3. LV short-axis fractional shortening (LVFS) was determined as follows: LVFS (in %) = (LVEDD − LVESD)/LVEDD × 100.

Hemodynamic analysis.

Mice were anesthetized with 2.5% avertin (12 μl/g ip), and the right carotid artery was cannulated with a high-fidelity microtip pressure transducer catheter (1.4 Fr, Model SPR-839; Millar Instruments, Houston, TX). The catheter was advanced into the aorta to measure aortic pressure and into the LV to measure LV function. LV pressure and +dP/dt and −dP/dt (change in pressure over time) were recorded using a chart recorder (Model MT95K2; Astro-Med, West Wanwick, RI).

Histological analysis and evaluation of apoptosis in tissue sections.

Histological analyses of the heart sections were conducted as described previously (33). Heart specimens were fixed with 10% neutral buffered formalin, embedded in paraffin, and sectioned at 6-μm thickness. Sections were deparaffinized and used for staining for fibrosis, cell size, and terminal deoxynucleotidyl transferase dUTP nick-end labeling (TUNEL). Images were obtained using a Nikon Digital Camera (DXM 1200) attached to a Nikon microscope (Nikon Eclipse E800).

Interstitial fibrosis was evaluated by picric acid Sirius red (PASR) staining. Images were taken using a 10× objective lens and analyzed using the ImagePro Plus 5.0 software system (Media Cybernetics). The positively stained (red) fibrotic area was expressed as a percentage of total area.

Cardiac myocyte cross-sectional area was determined from wheat germ agglutinin-Rhodamine (Vector Laboratories)-stained cardiac tissue sections. Images of cross-sectioned cardiac myocytes were taken using a 40× objective lens and analyzed using the ImagePro Plus 5.0 software system (Media Cybernetics). The mean cardiac myocyte cross-sectional area was calculated for each animal, and the group mean was then calculated for each group.

DNA fragmentation was detected in situ using TUNEL as described (33). Briefly, deparaffinized sections were incubated with proteinase K, and DNA fragments were labeled with fluorescein-conjugated dUTP using TdT (Roche Diagnostics). Nuclear density was determined by manual counting of DAPI (4′,6-diamidino-2-phenylindole)-stained nuclei in 20 fields from each animal using the 40× objective and of TUNEL-positive nuclei in the same fields using the same power objective. By limiting the counting of total nuclei and TUNEL-positive nuclei to areas with a true cross section of myocytes, it became possible to selectively count only those nuclei that clearly were within myocytes.

Estimation of total cardiac myocyte number was made using hematoxylin-eosin-stained tissue sections as previously described (32).

RT-real-time PCR.

Neonatal rat cardiac myocytes were infected with adenovirus harboring LacZ (Ad-LacZ) or GSK-3α (Ad-GSK-3α). Total RNA was isolated from myocytes using TRIzol reagent (Invitrogen) after 48–72 h of transduction, and 1 μg RNA was used for cDNA synthesis (Thermoscriptase; Ambion). Real-time PCR (quantitative PCR) was carried out using primers (Table 1) specific for transforming growth factor-β (TGF-β), FGF-1 and -2 (FGF-2), and GAPDH.

View this table:
Table 1.

Primer list for RT-PCR

Immunohistochemistry.

Deparaffinized tissue sections were antigen-unmasked using citrate buffer and blocked in 5% bovine serum albumin in PBST (0.3% Triton X-100 in PBS) for 1 h. The following antibodies were used as primary antibodies: TGF-β (Abcam, 1:250) and troponin T (NeoMarkers, 1:500).

Immunoblotting analysis.

Cardiac tissue homogenates were made in a radioimmunoprecipitation assay (RIPA) buffer containing 150 mM NaCl, 1% Triton-X 100, 0.5% sodium deoxycholate, 0.1% sodium dodecyl sulphate, and 50 mM Tris, pH 8.0, and supplemented with a protease inhibitor cocktail (Sigma), 5 mM NaF, and 1 mM sodium orthovanadate. We used antibodies against phospho-ERK, ERK, α-tubulin, GAPDH, sarcoplasmic reticulum Ca2+ ATPase type 2a (SERCA2a), and TGF-β for immunoblotting.

Statistical analysis.

The data were presented as mean ± SE and first analyzed using ANOVA. If significant differences were observed, the post hoc analysis was performed using the Scheffe's test. A P value less than 0.05 was considered significant.

RESULTS

Inhibition of ERK by GSK-3α is reversed by CA-MEK1.

In agreement with our previous report (32), Tg-GSK had a lower phospho-ERK level than NTg (Fig. 1, A and B). Both Tg-MEK1 and Bigenic had higher phospho-ERK levels than either NTg or Tg-GSK (Fig. 1, A and B). Importantly, the phospho-ERK level in Bigenic was similar to that in Tg-MEK1 (Fig. 1, A and B). After 4 wk of PO, the phospho-ERK level in Tg-GSK was still lower than that in NTg (Fig. 1, C and D). The phospho-ERK level in Bigenic was significantly higher than that in Tg-GSK, but was not significantly different from that in Tg-MEK1 (Fig. 1, C and D). These data indicate that inhibition of ERK by overexpressed GSK-3α is reversed by CA-MEK1.

Fig. 1.

Phospho-ERK (pERK) and ERK in aging (A and B) and pressure overloaded (PO; C and D) mouse hearts. A: representative immunoblotting results at baseline. B: baseline pERK-to-GAPDH ratio. C: representative immunoblotting results after 4 wk of PO. D: pERK-to-GAPDH ratio after 4 wk of PO; N = 4 in each group. NTg, nontransgenic mice; GSK-3α, GSK-3α transgenic mice; MEK1, constitutively active MEK1 (CA-MEK1) transgenic mice; Bigenic, bigenic mice expressing both GSK-3α and CA-MEK1; NS, not significant.

Inhibition of ERK mediates GSK-3α-induced suppression of cardiac hypertrophy.

The left ventricular weight (LVW, mg)/body weight (BW, g) and LVW (mg)/tibia length (TL, mm) in Tg-GSK were significantly lower than those in NTg (2.40 ± 0.09 vs. 2.93 ± 0.15, P < 0.01; and 4.34 ± 0.36 vs. 5.08 ± 0.31, P < 0.05; Fig. 2, A and B). However, the LVW/BW and LVW/TL in Tg-MEK1 (3.76 ± 0.32, P < 0.05; and 6.96 ± 0.48, P < 0.05) were significantly higher than those in NTg (Fig. 2, A and B). The LVW/BW and LVW/TL in the Bigenic (3.83 ± 0.24 and 7.19 ± 0.62) were significantly higher than those in either NTg (P < 0.05) or Tg-GSK (P < 0.01) (Fig. 2, A and B). Importantly, the LVW/BW and LVW/TL in Bigenic were not significantly different from those in Tg-MEK1. We also measured cardiac myocyte cross-sectional area in wheat germ agglutinin-stained tissue sections (Fig. 2C). The cardiac myocyte cross-sectional area in Tg-GSK was significantly smaller than that in NTg (263 ± 0.9 μm2 vs. 310 ± 1.6 μm2; P < 0.01; Fig. 2, C and D). The cardiac myocyte cross-sectional area in Tg-MEK1 (452 ± 28 μm2) was significantly bigger than that in either NTg (P < 0.001) or Tg-GSK (P < 0.001) (Fig. 2, C and D). The cardiac myocyte cross-sectional area in Bigenic (452 ± 6 μm2) was also significantly bigger than that in either NTg (P < 0.001) or Tg-GSK (P < 0.001), and, importantly, was not significantly different from that in Tg-MEK1 (Fig. 2, C and D). Echocardiographic measurements showed that the diastolic LV posterior wall thickness (LV DPWT) and end-diastolic dimension (LVEDD) in Tg-GSK were slightly but significantly smaller than those in NTg (Fig. 2, E and F). The LV DPWT in Bigenic was significantly greater than that in either NTg or Tg-GSK, but was not significantly different from that in Tg-MEK1 (Fig. 2E). The LVEDD in Bigenic was slightly but significantly smaller than that in NTg or Tg-GSK, but was not significantly different from that in Tg-MEK1 (Fig. 2F). These results collectively suggest that CA-MEK1 rescues cardiac myocyte growth inhibited by overexpressed GSK-3α in cardiac myocytes. Because basal cell loss is a part of aging process, we estimated total cardiac myocyte number. Aging NTg had a slight decrease in cardiac myocyte number than young NTg mice (Fig. 3), whereas aging Tg-GSK had a marked decrease in cardiac myocyte number, which was also significantly less than that in aging NTg (Fig. 3). Aging Tg-MEK1 and Bigenic did not show significant cardiac myocyte loss (Fig. 3). These data suggest that the marked loss of cardiac myocyte in aging Tg-GSK was reversed in CA-MEK1 in aging Bigenic.

Fig. 2.

Left ventricular weight (LVW) and cardiac myocyte size. A: LVW-to-body weight (BW) ratio. B: LVW-to-tibia length (TL) ratio. In both A and B, n = 9. For GSK-3α, n = 9. For MEK1, n = 8. For Bigenic, n = 8. C: wheat germ agglutinin-rhodamine staining of cardiac sections. Representative images are shown. D: cardiac myocyte cross-sectional area. E: left ventricular diastolic posterior wall thickness (LV DPWT). F: left ventricular end-diastolic dimension (LVEDD). For B and C, N = 5 in both NTg and GSK-3α; N = 4 in both MEK1 and Bigenic. For E and F, n = 6 in NTg; n = 5 in GSK-3α; n = 4 in both MEK1 and Bigenic.

Fig. 3.

Estimated total cardiac myoycte number. N = 3 in each group. Young = 3 to 4 mo old; old = 12 mo old.

Inhibition of ERK mediates the GSK-3α-induced increase in apoptosis.

We measured apoptosis by calculating the percentage of TUNEL-positive nuclei (Fig. 4). At 12 mo, Tg-GSK still had a significantly higher percentage of TUNEL-positive nuclei than NTg (0.18 ± 0.02% vs. 0.07 ± 0.01%; P < 0.05). The percentage of TUNEL-positive nuclei in Tg-MEK1 (0.06 ± 0.03%) was significantly lower than that in Tg-GSK (P < 0.05), but not significantly different from that in NTg. The percentage of TUNEL-positive nuclei in Bigenic (0.09 ± 0.02%) was also significantly lower than that in Tg-GSK (P < 0.05), but importantly, was not significantly different from that in Tg-MEK1. These data suggest that inhibition of ERK mediates the GSK-3α-induced increase in apoptosis.

Fig. 4.

Apoptosis. A: representative images of terminal deoxynucleotidyl transferase dUTP-mediated nick-end labeling (TUNEL) staining are shown. B: percentage of TUNEL-positive nuclei. For NTg, n = 5. For GSK-3α and MEK1, n = 4. For Bigenic, n = 6. DAPI, 4′,6-diamidino-2-phenylindole.

Inhibition of ERK does not mediate the GSK-3α-induced increase in fibrosis.

We measured fibrosis in picric acid Sirius red-stained cardiac tissue sections (Fig. 5). At 12 mo, Tg-GSK had a significantly higher percentage of fibrosis than NTg (8.82 ± 1.21% vs. 2.56 ± 0.46%; P < 0.01). Surprisingly, the percentage of fibrosis in Tg-MEK1 (11.25 ± 1.66%) was also significantly higher than that in NTg (P < 0.01) and not significantly different from that in Tg-GSK. The percentage of fibrosis in Bigenic (9.43 ± 1.85%) was also significantly higher than that in NTg (P < 0.01) and was not significantly different from that in either Tg-GSK or Tg-MEK1. These results indicate that it is unlikely that the fibrosis in Tg-GSK is mediated through inhibition of ERK.

Fig. 5.

Interstitial fibrosis. A: picric acid sirius red staining of heart sections. Representative images are shown. B: percentage of interstitial fibrosis. N = 5 in both NTg and GSK-3α; N = 4 in both MEK1 and Bigenic. C: mRNA expression of transforming growth factor-β (TGF-β) and FGF-1 and -2 (FGF2) in neonatal rat cardiac myocytes. *P < 0.05 vs. LacZ. D: immunostaining of TGF-β in cardiac tissue sections.

The overexpression of GSK-3α in cardiac myocytes caused increases in interstitial fibrosis, suggesting that some profibrotic factor produced and secreted by cardiac myocytes must be affected by GSK-3α. To find out which factor is influenced by GSK-3α overexpression, neonatal rat cardiac myocytes were transduced with Ad-LacZ or Ad-GSK-3α. RT-quantitative PCR results showed that the expression of TGF-β, but not FGF1 or FGF2, was significantly increased by GSK-3α (Fig. 5C). We then examined TGF-β expression by immunohistochemistry in the four groups of mice. TGF-β protein level was significantly increased in Tg-GSK (Fig. 5D). Interestingly, the protein level of TGF-β was also increased in Tg-MEK1 and Bigenic (Fig. 5D).

Inhibition of ERK mediates the GSK-3α-induced decrease in cardiac function.

Echocardiographic measurements revealed that LVEF and LVFS were significantly lower in Tg-GSK than in NTg (0.57 ± 0.05 vs. 0.68 ± 0.01, P < 0.05; and 24.6 ± 2.8% vs. 31.6 ± 0.9%, P < 0.05, respectively; Fig. 6, AC) at 12 mo of age. The LVEF and LVFS in Tg-MEK1 (0.776 ± 0.053 and 40.3 ± 4.5%, respectively) were significantly higher than those in either NTg (P < 0.05) or Tg-GSK (P < 0.01; Fig. 6, AC). The LVEF and LVFS in Bigenic (0.811 ± 0.022 and 42.9 ± 2.2%, respectively) were also significantly higher than those in either NTg (P < 0.05) or Tg-GSK (P < 0.01; Fig. 6, AC). Importantly, the LVEF and LVFS in Bigenic were not significantly different from those in Tg-MEK1 (Fig. 6, AC). Hemodynamic measurements showed that, at 12 mo, the +LV dP/dt (in mmHg/s) in Tg-GSK was significantly lower than that in NTg (5,025 ± 172 vs. 7,100 ± 485, P < 0.05). The +LV dP/dt (in mmHg/s) in Bigenic (7,513 ± 403) was significantly higher than that in Tg-GSK (P < 0.05), but was not significantly different from that in Tg-MEK1 (7,076 ± 698). The −LV dP/dt (in mmHg/s) in Tg-GSK was significantly lower than that in NTg (4,213 ± 291 vs. 6,100 ± 560; P < 0.05). The −LV dP/dt (in mmHg/s) in Bigenic (5,988 ± 385) was significantly higher than that in Tg-GSK (P < 0.05), but was not significantly different from that in Tg-MEK1 (6,000 ± 575). Tg-GSK had a significantly higher LV end-diastolic pressure (LVEDP; in mmHg) than NTg (11.8 ± 2.1 vs. 6.8 ± 1.7, P < 0.05). The LVEDP in Bigenic (7.0 ± 1.3) was significantly lower than that in Tg-GSK (P < 0.05), but was not significantly different from that in Tg-MEK1 (8.7 ± 1.2). Both the echocardiographic data and the hemodynamic data support a role for inhibition of ERK in mediating the GSK-3α-induced decrease in LV function during aging.

Fig. 6.

Cardiac function determined by echocardiography in 12-mo-old mice. A: representative M-mode echo images are shown. B: left ventricular ejection fraction (LVEF). C: left ventricular fractional shortening (LVFS). For NTg, n = 6. For GSK-3α, n = 5. N = 4 in both MEK1 and Bigenic mice. D: representative results of immunoblotting of sarcoplasmic reticulum Ca2+-ATPase type 2a (SERCA2a; left) and normalized band density (right; n = 4 in each group).

Because reduction in SERCA2a protein expression and activity has been involved in heart failure (19, 31), we measured SERCA2a in these mice. The expression of SERCA2a was significantly decreased in Tg-GSK mice, but was significantly increased in Tg-MEK1 and Bigenic mice (Fig. 6D). This result indicates that alterations intrinsic properties of calcium handling through MEK/ERK-dependent mechanisms in cardiac myocyte might be involved in cardiac dysfunction caused by overexpression of GSK-3α.

Inhibition of ERK mediates the GSK-3α-induced decrease in cardiac function during PO.

We next examined cardiac hypertrophy, apoptosis, and function during PO in young mice (3 to 4 mo old). In sham-operated mice, the LVW/TL and cardiac myocyte cross-sectional area were significantly lower in Tg-GSK than in NTg (Fig. 7, A and B). The LVW/TL in Tg-MEK1 and in Bigenic were not significantly different from each other, but were significantly higher than that in NTg (Fig. 7A). The cardiac myocyte cross-sectional area in those mice exhibited the same trend as the LVW/TL (Fig. 7B). Four weeks after TAC, LVW/TL and cardiac myocyte cross-sectional area in NTg were increased 74.6% and 40.4%, respectively, while that in Tg-GSK increased 52.0% and 21.6%, respectively, significantly less than those in NTg. Similarly, echocardiographic measurements showed that the diastolic LV posterior wall thickness (LV DPWT) in Tg-GSK increased significantly less than that in NTg (43.5% vs. 59.3%; Fig. 7C). Taken together, these data suggest that GSK-3α significantly attenuates PO-induced cardiac hypertrophy. However, GSK-3α failed to reduce the extent of PO-induced cardiac hypertrophy in the presence of CA-MEK1 [LVW/TL: Tg-MEK1 = 6.3 ± 1.6; Bigenic = 6.7 ± 0.5, P > 0.05; cardiac myocyte cross-sectional area (μm2): Tg-MEK1 = 298.2 ± 1.7, Bigenic = 291.6 ± 9.9, P > 0.05; LV DPWT (mm): Tg-MEK1 = 1.39 ± 0.04, Bigenic = 1.35 ± 0.06, P > 0.05; Fig. 7, AC]. These data suggest that inhibition of MEK/ERK signaling mediates GSK-3α-induced suppression of cardiac hypertrophy in response to PO.

Fig. 7.

Cardiac hypertrophy, apoptosis, and function in young mice 4 wk after transverse aortic constriction (TAC). A: LVW-to-TL ratio. B: cardiac myocyte cross-sectional area. C: LV DPWT. D: percentage of TUNEL-positive nuclei. E: LVEF. F: LVFS. In AD, N = number of animals.

In sham-operated mice, the percentage of TUNEL-positive nuclei was significantly higher in Tg-GSK than in NTg, Tg-MEK1, or Bigenic (Fig. 7D). However, GSK-3α failed to increase the percentage of TUNEL-positive nuclei in the presence of CA-MEK1 in Bigenic (Fig. 7D). After PO, there was a significantly greater increase in the percentage of TUNEL-positive nuclei in Tg-GSK than in NTg (2.34-fold vs. 1.72-fold, P < 0.05; Fig. 7D), suggesting that GSK-3α more robustly enhances apoptosis during PO. In contrast, GSK-3α failed to further increase the percentage of TUNEL-positive nuclei in the presence of CA-MEK1 in Bigenic (1.40-fold increase in Bigenic vs. 1.44-fold increase in Tg-MEK1, P > 0.05; Fig. 7D). These results indicate that inhibition of ERK is involved in the increase in apoptosis caused by overexpressed GSK-3α during PO.

Under sham-operated conditions, LVEF and LVFS in Tg-GSK were not significantly different from those in NTg (Fig. 7, E and F), and LVEF and LVFS in Bigenic were not significantly different from those in Tg-MEK1 (Fig. 7, E and F). Four weeks after PO induced by TAC, LVEF and LVFS were significantly lower in Tg-GSK than in their sham-operated controls (Fig. 7, E and F), whereas LVEF and LVFS in NTg were not significantly different from those in their sham-operated controls, demonstrating that GSK-3α promotes cardiac dysfunction in response to PO. In contrast, the LVEF and LVFS in Bigenic were not significantly different from those in Tg-MEK1 (Fig. 7, E and F), suggesting that GSK-3α failed to cause cardiac dysfunction in the presence of CA-MEK1 under PO. These results indicate that inhibition of ERK mediates GSK-3α overexpression-induced cardiac dysfunction in response to PO.

DISCUSSION

The role that GSK-3α plays during cardiac growth has recently been explored by several studies using different genetic manipulation methods (18, 32, 34). Using a transgenic approach, we have shown that GSK-3α, when overexpressed in cardiac myocytes, inhibits postnatal cardiac growth (32), and this effect of GSK-3α in the heart has been confirmed by others using a loss-of-function method (34). Our previous data have indicated that GSK-3α, when overexpressed in the heart, inhibited the phosphorylation of ERK (32), an effect that has also been confirmed by others using the loss-of-function approach (34). GSK-3 has also appeared to inhibit phosphorylation of ERK in human colon cancer cell lines (27) and mouse bone marrow-derived dendritic cells (22), lending further support to our findings. Our previous in vitro results further suggested that inhibition of ERK phosphorylation by GSK-3α occurred through a MEK1-dependent mechanism and played an important role in mediating the inhibitory effect of GSK-3α on cardiac myocyte growth in culture (32). In our current study, CA-MEK1, which constitutively activates ERK, rescued the cardiac growth inhibited by GSK-3α overexpressed in the heart, up to 12 mo of age, providing in vivo evidence supporting the idea that GSK-3α inhibits cardiac growth though MEK/ERK-dependent mechanisms.

The role that GSK-3α plays in the development of hypertrophy and heart failure in response to PO has also been explored in several studies (18, 32, 34), yet with seemingly conflicting conclusions. The different genetic manipulation methods used may play some role in this discrepancy. GSK-3α, when overexpressed in cardiac myocytes in the mouse heart, decreased cardiac hypertrophy but exacerbated cardiac dysfunction in response to PO (32). Mice with systemic knock-in of GSK-3αS21A, which is constitutively active, developed more severe PO-induced cardiac hypertrophy and heart failure (18). Knockout of GSK-3α in all mouse tissues resulted in enhanced cardiac hypertrophy and contractile dysfunction in response to PO (34). The latter two studies used a systemic approach, in which the potential contribution of other cell types in the heart to the phenotypes observed could not be excluded. For example, recent studies have shown that cardiac fibroblasts play key roles in myocardial hypertrophy and function (5, 12, 25). Cardiac fibroblasts in the postnatal heart promote baseline cardiac myocyte hypertrophy (12). During PO, Ras-associated domain family 1 isoform A in cardiac fibroblasts inhibits fibrosis and cardiac hypertrophy and preserves cardiac function (5). In response to PO, cardiac fibroblast-specific deletion of Krüppel-like factor 5 suppresses cardiac fibrosis and hypertrophy, whereas cardiomyocyte-specific deletion does not (25). These studies strongly indicate that altering a signaling molecule in cardiac fibroblasts has profound effects on cardiac responses to stress. In our current study, GSK-3α is only overexpressed in cardiac myocytes, whereas its expression in other cell types is not manipulated, allowing us to focus on the effects of GSK-3α in cardiac myocytes.

The expression of GSK-3α in old monkey hearts was much higher than that in young monkey hearts (data not shown), suggesting that GSK-3α might play some role in aging. Our current study shows that GSK-3α overexpression in cardiac myocytes in the mouse heart resulted in cardiac dysfunction and increased apoptosis at 12 mo of age, indicating an accelerated aging phenotype. Importantly, our results suggest that the phenotype can be reversed by constitutively active MEK1. Our data suggest that decreased expression of SERCA2a through MEK/ERK-dependent mechanisms by overexpressed GSK-3α might play some role in the development of cardiac dysfunction. Accumulating lines of evidence suggest that MEK/ERK signaling is critical to the aging process. The level of active phosphorylated ERK was decreased in senescent cells (26). Decreased ERK activity was found to be associated with an age-related decline in the proliferative response of hepatocytes to mitogenic stimulation (11). On the other hand, caloric restriction, an intervention that increases rodent life span and delays the onset of many age-related declines in physiologic function, prevented an age-related loss in ERK activity (13). The beneficial effect of activation of MEK/ERK signaling in aging has recently been demonstrated in type 5 adenylyl cyclase knockout mice (30). Although some controversy exists as to whether the hypertrophy caused by activation of the MEK/ERK pathway is compensatory or pathological (4, 29), activation of MEK/ERK signaling was shown to enhance cardiac function in both reports. Specifically, transgenic mice with cardiac-specific expression of activated MEK1 had an increased fractional shortening (4), and knockin mice expressing a Raf1 mutant leading to MEK/ERK activation exhibited increased fractional shortening, stroke volume, cardiac output, and maximum dP/dt (29). However, the mechanism through which MEK/ERK activation improves cardiac contractile function remains to be elucidated. We speculate that the decrease in apoptosis caused by MEK/ERK activation might play some role in enhancing cardiac contractile function. Cardiac myocyte loss increases with aging (14). The loss of cardiac myocytes is an important etiological part of the development of cardiac dysfunction, and apoptosis is likely involved in cardiac myocyte loss during aging. Evidence obtained from both human and animal models suggests that apoptosis is an important mode of cell death, i.e., loss of cardiac myocytes, during the development of cardiac dysfunction (16). Cardiac apoptosis most likely contributes to the progression of cardiac dysfunction (23). However, the downstream mechanism involved in the protective effect of MEK/ERK activation against apoptosis remains unknown.

It is a surprise that cardiac fibrosis is not corrected by CA-MEK1 in the bigenic mice. GSK-3α in cardiac myocytes may not act through inhibition of ERK to affect cardiac fibrosis, leaving the mechanism through which GSK-3α overexpression in cardiac myocytes causes fibrosis to be elucidated. Cardiac myocytes and fibroblasts communicate with each other through several mechanisms, including paracrine factors, alterations in extracellular matrix homeostasis, and direct cell-cell interactions (reviewed in Ref. 15). GSK-3α overexpression in cardiac myocytes may alter interstitial fibrosis by one or more of these mechanisms. Our results suggest that increased TGF-β production by overexpressed GSK-3α in cardiac myocytes might, at least partially, be responsible for the increased fibrosis in Tg-GSK and Bigenic, through paracrine mechanisms.

In summary, CA-MEK1 rescued cardiac dysfunction and apoptosis caused by GSK-3α overexpression in transgenic mice. Constitutive activation of MEK/ERK signaling also abrogated the inhibitory effect of GSK-3α on cardiac hypertrophy. Our results suggest that GSK-3α overexpression in cardiac myocytes inhibits cardiac hypertrophy and causes cardiac dysfunction, at least in part through inhibition of ERK activation.

GRANTS

This work was supported by American Heart Association Grant 0930179N (to P. Zhai).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author(s).

AUTHOR CONTRIBUTIONS

Author contributions: Y.M., J.G., and P.Z. performed experiments; Y.M. and P.Z. analyzed data; Y.M. and P.Z. prepared figures; J.D.M., J.S., and P.Z. interpreted results of experiments; J.S. edited and revised manuscript; P.Z. conception and design of research; P.Z. drafted manuscript.

ACKNOWLEDGMENTS

We thank Daniela Zablocki and Christopher Brady for critical reading of this manuscript.

REFERENCES

  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
  27. 27.
  28. 28.
  29. 29.
  30. 30.
  31. 31.
  32. 32.
  33. 33.
  34. 34.
View Abstract