Mammalian enabled (Mena) is a critical regulator of cardiac function

Abstract

Mammalian enabled (Mena) of the Drosophila enabled/vasodilator-stimulated phosphoprotein gene family is a cytoskeletal protein implicated in actin regulation and cell motility. Cardiac Mena expression is enriched in intercalated discs (ICD), the critical intercellular communication nexus between adjacent muscle cells. We previously identified Mena gene expression to be a key predictor of human and murine heart failure (HF). To determine the in vivo function of Mena in the heart, we assessed Mena protein expression in multiple HF models and characterized the effects of genetic Mena deletion on cardiac structure and function. Immunoblot analysis revealed significant upregulation of Mena protein expression in left ventricle tissue from patients with end-stage HF, calsequestrin-overexpressing mice, and isoproterenol-infused mice. Characterization of the baseline cardiac function of adult Mena knockout mice (Mena^−/−) via echocardiography demonstrated persistent cardiac dysfunction, including a significant reduction in percent fractional shortening compared with wild-type littermates. Electrocardiogram PR and QRS intervals were significantly prolonged in Mena^−/− mice, manifested by slowed conduction on optical mapping studies. Ultrastructural analysis of Mena^−/− hearts revealed disrupted organization and widening of ICD structures, mislocalization of the gap junction protein connexin 43 (Cx43) to the lateral borders of cardiomyoycytes, and increased Cx43 expression. Furthermore, the expression of vinculin (an adherens junction protein) was significantly reduced in Mena^−/− mice. We report for the first time that genetic ablation of Mena results in cardiac dysfunction, highlighted by diminished contractile performance, disrupted ICD structure, and slowed electrical conduction.

heart failure (HF) is a progressive disease with poor prognosis and continues to be one of the leading causes of morbidity and mortality in the world. In the United States alone, the prevalence of HF is roughly 5.7 million cases, with 670,000 new cases per year and estimated costs of $37.2 billion for 2009 (24, 34). To investigate the pathology of HF, numerous genetically engineered mouse models of HF have been created. In particular, ablation of the muscle LIM protein (MLP), a muscle-restricted cytoarchitectural protein (3), or myocardial-targeted transgenic overexpression of calsequestrin (CSQ), a high-capacity sarcoplasmic reticulum Ca²⁺-binding protein (11, 26), resulted in dilated cardiomyopathy with significant left ventricular (LV) dysfunction. These transgenic mice exhibit several hallmarks of the failing human heart, as well as biochemical abnormalities in the β-adrenergic receptor (β-AR) signaling cascade that lead to a reduction or loss of β-AR inotropic reserve (43). Such HF models have provided powerful tools to investigate the development of the HF phenotype at the gene expression level. Using oligonucleotide microarray analysis of LV mRNA, we have reported a cardiac gene expression profile that could blindly predict HF (8). This profile included significantly altered expression of several genes already implicated in HF, such as atrial natriuretic factor (ANF) and brain natriuretic peptide (10, 42), as well as regulated expression of several novel and known genes not previously associated with HF.

Mammalian enabled (Mena), the mammalian homolog of Drosophila enabled (Ena), was identified as one of the top-ranked genes dramatically upregulated during HF (8). Mena, in addition to vasodilator-stimulated phosphoprotein (VASP) and Ena/VASP-like protein (Evl), is one of three established members of the Ena/VASP family of actin regulatory proteins (20). In the adult mouse, Mena is predominantly expressed in brain, testes, ovaries, adipose tissue, and to a lesser extent in the heart (33). Interestingly, cardiac Mena is principally localized at the intercalated disc (ICD), an intercellular junctional complex important for maintaining structural integrity and synchronized cardiac contraction (33). Cardiac expression of Mena is progressively downregulated from neonate to detectable but low levels in the adult heart (19). Upregulation of Mena during end-stage HF (8) mirrors the characteristic reversion to a “fetal gene program” (12), similar to ANF and β-myosin heavy chain, which were concomitantly identified as significant predictors of HF development (8).

Little is known about the specific role of Mena in the heart and the importance of its localization in the ICD. However, Ena/VASP family members have been described as critical modulators of actin assembly and cell motility (20, 35). Furthermore, Mena is implicated in other actin-dependent processes such as axon guidance, neural tube closure, and cell-cell adhesion (6, 15, 33, 49, 57). To investigate the functional role of Mena in the heart, we investigated possible variations in Mena protein expression and localization in several HF models. Importantly, we have also characterized the cardiac phenotype of Mena knockout (Mena^−/−) mice. Our findings demonstrate that, analogous to its “fetal gene” pattern of expression, Mena protein levels correlate with HF phenotype. Moreover, genetic deletion of Mena results in deteriorating cardiac performance and structural alterations, particularly at the ICD. We propose that Mena plays a critical role in the maintenance of normal cardiac contraction, and the disruption thereof may contribute to the pathophysiology of HF.

MATERIALS AND METHODS

Antibodies.

Antibodies used were directed against Mena and phosphorylated Mena (mouse, 1:1,000; Gertler laboratory), connexin 43 (Cx43) (rabbit, 1:1,000, Sigma), Na_v1.5 (rabbit, 1:500; Sigma), vinculin and GAPDH (mouse, 1:1,000; Millipore), N-cadherin and β-catenin (mouse, 1:1,000; BD Transduction), actin (mouse, 1:3,000; Calbiochem), and Plakophilin-2 (mouse, 1:100; Meridian Life Science).

Genetically engineered mouse models.

Mena^−/− mice were generated as previously described (33). Cardiac-restricted CSQ-overexpressing mice were obtained from Dr. Howard Rockman (Duke University, Durham, NC), with permission from Dr. Larry Jones (Indiana University, Bloomington, IN). Wild-type (WT) C57/Bl6 mice, used for the isoproterenol infusion study, were purchased from Jackson Laboratories. All animal studies were approved by the University of Rochester Medical Center Animal Care and Use Committee.

Human LV tissue procurement.

LV myocardial samples (LV free wall close to the apex) were obtained from the explanted hearts of five cardiac transplant recipients for end-stage HF due to dilated cardiomyopathy. Nonfailing heart tissue was obtained from five donors deemed unsuitable for transplant with no cardiac dysfunction. Samples were immediately frozen in liquid nitrogen and stored at −140°C until use. Patient consent was obtained to use tissue for research in accordance with National Institutes of Health (NIH) guidelines from the HIPAA, under a protocol approved by the Institutional Review Board.

Chronic isoproterenol infusion.

Miniosmotic pumps (model 1007D; Alzet) were implanted in WT mice (6–8 mo old) anesthetized with 2% isoflurane and 40% oxygen and maintained with 0.5% isoflurane and 40% oxygen. Pumps were filled with isoproterenol or vehicle (0.002% ascorbic acid in sterile filtered saline) and delivered at 30 mg·kg⁻¹·day⁻¹ for seven consecutive days (n = 5 animals/group). Mice were subsequently euthanized; excised hearts were rinsed briefly in PBS and cut into two halves, approximately at the suprapapillary level. The bottom part of the heart (ventricles) was snap-frozen in liquid nitrogen for biochemistry analysis, and the top part of the heart (ventricles + atria) was fixed in 10% neutral buffered formalin for histological analysis.

Immunoblotting and immunoprecipitation.

LV tissue was homogenized in RIPA buffer (1% Nonidet P-40, 0.5% sodium deoxycholate, 0.1% SDS, 150 mmol/l NaCl, and 50 mmol/l Tris, pH 8.0) containing protease inhibitor (1836170; Roche) and phosphatase inhibitor (P2850; Sigma) cocktails. Nitrocellulose membranes were incubated in blocking buffer (1% nonfat dry milk or 3% BSA + 0.1% Tween 20 in PBS, pH 7.4) and probed overnight with primary antibodies. Blots were probed with appropriate horseradish peroxidase-linked anti-mouse/rabbit secondary antibody (GE Healthcare) and developed using Amersham ECL chemiluminescence detection reagents. Relative density was quantified in densitometric units using NIH Image J software. Values of each protein were normalized to GAPDH or actin.

For immunoprecipitation, heart lysates (1 mg) from 5-mo-old CSQ mice were precleared and incubated overnight at 4°C with 2 μg mouse anti-Mena antibody or mouse IgG protein. Next, 40 μl washed protein A/G agarose beads (Santa Cruz Biotechnology) were added to the lysate/antibody mixture at 4°C for 3 h. Following washes, beads were resuspended in Laemmli buffer, boiled 5 min, electrophoresed by SDS-PAGE, and probed with mouse anti-phosphorylated Mena and mouse anti-Mena antibodies.

Histology and immunofluorescence.

Tissue was fixed in 10% neutral buffered formalin and embedded in paraffin. Sections were prepared by bisecting the heart across the upper papillary level and cutting 5-μm-thick segments down to the midpapillary level. Sections were stained with hematoxylin and eosin and Masson's trichrome (Leica Autostainer EX) to evaluate changes in overall myocardial structure, fibrosis, and muscle. Histology images were captured using an Olympus BX41 inverted microscope.

For immunofluorescence staining, tissue sections were deparaffinized and permeabilized with 0.2% Triton. Antigen retrieval was performed in 10 mmol/l citrate buffer. Nonspecific sites were blocked overnight at 4°C in serum-free DakoCytomation protein block solution (Dako). Primary antibodies, rabbit anti-Mena, rabbit anti-Cx43, and mouse anti-vinculin were prepared in antibody diluent (Dako) and incubated for 1 h at 37°C. After washes, sections were incubated with Alexa fluor secondary antibody (anti-mouse/rabbit 488/568; Molecular Probes) for 1 h at room temperature. Confocal images were acquired using an inverted microscope (Carl Zeiss). As controls for background fluorescence, sections were stained with secondary antibody alone. Images were blinded before analysis, and total immunolabeled Cx43 (pixels/μm²) was measured using an automated computer algorithm (ImagePro version 6.2; Media Cybernetics).

Echocardiography.

Cardiac LV function was assessed in conscious mice using standard transthoracic two-dimensional and M-mode echocardiography with a Visual Sonics Vevo 770 machine equipped with a 30-MHz probe (Visual Sonics, Toronto, Canada). Data were collected from Mena^−/− and WT littermates at 2, 4, and 6–8 mo of age.

Electrocardiograms.

Four-lead surface electrocardiogram (ECG) recordings were obtained from 6- to 8-mo-old conscious mice as described previously (56).

Electron microscopy.

Ventricular tissue sections from WT and Mena^−/− mice were fixed in 2.5% glutaraldehyde, postfixed in 1.0% osmium tetroxide to preserve lipid and lipoprotein structures, processed through a graded series of alcohols, infiltrated in Spurr epoxy resin, embedded, and polymerized at 70°C overnight. Semithin (1.0–2.0 μm) sections placed on grids were stained with uranyl acetate and lead citrate and imaged with a Hitachi 7100 Transmission Electron Microscope with a MegaView III digital camera and “AnalySIS” software.

Optical conduction mapping.

Adult Mena^−/− and WT mice were heparinized (0.5 U/g ip) and then anesthetized with a ketamine (116 mg/kg)/acepromazine (11 mg/kg) mixture injected intraperitoneally. The heart was rapidly excised through thoracotomy and subsequently connected to a Langendorff perfusion system, which utilized warm (36 ± 1°C) bicarbonate-buffered Tyrode's solution (pH 7.4) bubbled with 95% O₂ plus 5% CO₂. The heart was placed in the well of a custom-made plastic chamber maintained at 35–36°C and allowed to equilibrate for 10 min. We recorded changes in cardiac conduction from the anterior epicardium of the ventricles during sinus rhythm at 400–600 frames/s, as described previously (4, 53, 55). The optical mapping system used a high-resolution 64 × 64-pixel charge-coupled device camera recording at a rate of 914 frames/s and a spatial resolution of 109 μm. The camera recorded light at wavelength that was >600 nm. Eighty micrograms of Di-4-ANEPPS (voltage-sensitive dye) dissolved in 1 ml of Tyrode's solution were injected in a 10-ml compliance chamber connected to the perfusion system, and the preparation was illuminated with a 250-W tungsten light source band passed through a 520- to 590-nm filter. No electromechanical uncouplers were used. The dimensions of our field of view were 7 × 7 mm. ECG was recorded simultaneously with optical mapping. Two silver electrodes were positioned in the superfusate, inside the plastic chamber, close to the heart in a position that approximates Lead II. The signal was acquired at 1 kHz, amplified (Axon Instruments Cyber Amp 380) and digitized (Axon Instruments Minidigi 1A), and stored in a computer for later analysis. An average signal was obtained for 7–10 sinus beats. As described previously (41), for each pixel, the activation time was calculated, and 0.5-ms isochronal lines were displayed.

Statistical analysis.

All data are expressed as means ± SE. Statistical significance between groups was determined using unpaired Student's t-test with P≤0.05 considered significant.

RESULTS

Altered mena expression, phosphorylation, and distribution in mouse and human HF.

We previously identified a significant increase in Mena RNA gene expression in failing mouse hearts (8). To confirm concordant changes in Mena protein expression, we performed immunoblots on myocardial tissue from established murine HF models, including chronic stimulation of β-ARs via isoproterenol and cardiac-restricted overexpression of CSQ (11, 32, 37). Chronic infusion of isoproterenol (30 mg·kg⁻¹·day⁻¹, 7 days) significantly increased Mena protein expression threefold compared with age-matched vehicle-treated mice. Likewise, Mena expression increased by 11-fold in CSQ animals compared with WT controls (Fig. 1B). LV tissue from patients with dilated cardiomyopathy also revealed significant upregulation of Mena protein compared with nonfailing samples (Fig. 1C).

Fig. 1.

Increased cardiac mammalian enabled (Mena) expression in mouse and human heart failure (HF). A: Western blot of ventricular tissue from 6-mo-old wild-type (WT) mice probed with Mena antibody. Isoproterenol treatment (30 mg·kg⁻¹·day⁻¹) for 1 wk significantly increased Mena expression compared with vehicle (0.002% ascorbic acid) in age-matched control animals (n = 3 animals/group; *P < 0.05). B: Western blot of ventricular tissue from 4-mo-old WT and calsequestrin (CSQ)-overexpressing mice probed with Mena antibody shows significant upregulation of Mena expression in the CSQ HF model compared with WT control mice (n = 3 animals/group; **P < 0.001). C: Western blot of ventricular tissue from nonfailing (NF) and idiopathic dilated cardiomyopathy (DCM) human hearts probed with Mena antibody shows that human HF is associated with a significant increase in Mena expression (n = 3 human heart samples/group; *P < 0.05). Densitometric analysis was performed by normalizing Mena expression to simultaneously measured GAPDH. The mean intensity per group is shown in arbitrary units (AU) ± SE. D: Western blot of ventricular tissue from WT and muscle LIM protein knockout (MLP^−/−) mice probed with Mena antibody shows a significant increase in Mena expression in MLP^−/− hearts. Cardiac expression of the carboxy tail of β-adrenergic receptor kinase (βARKct) in MLP^−/− mice reduced Mena expression toward baseline (Rescue). E: left ventricular (LV) tissue from patients with idiopathic dilated cardiomyopathy (IDC) showed enhanced Mena protein expression compared with nonfailing heart samples. Samples loaded in lanes 3–10 (left to right) correspond to paired tissue from four IDC patients before (Pre) and 2 mo following (Post) left ventricular assist device (LVAD) treatment. Similar to genetic rescue with βARKct, LVAD treatment partially normalized Mena levels.

MLP is a critical component in maintaining proper cytostructural organization, and MLP^−/− mice develop cardiomyopathy mirroring the clinical features of human HF (3). As with the other HF models, LV tissue from MLP^−/− mice showed significant upregulation of Mena protein expression compared with WT littermate controls (Fig. 1D). Interestingly, salutary expression of the carboxy tail of β-AR kinase normalized Mena expression in these animals (Fig. 1D) (8). We found similar results following salutary LV assist device support in HF patients (Fig. 1E) (9). Together, these results indicate that myocardial Mena protein expression correlates with LV function, mirroring changes in Mena mRNA. Data from the CSQ model are subsequently presented, although we found comparable results in other HF models.

Mena is functionally regulated through serine phosphorylation by cAMP-dependent protein kinase A (PKA), downstream of β-AR stimulation (35). Alterations in the phosphorylation status of such PKA targets as ryanodine receptors and phospholamban contribute to the pathology of HF (36, 43, 48). Therefore, we hypothesized that changes in the phosphorylation state of Mena may be apparent in HF. To test this, immunoprecipitated Mena from WT and CSQ LV lysates was immunoblotted with a phosphospecific antibody to Mena's conserved PKA phosphorylation site at serine-236 (31, 35). Phosphorylated Mena expression increased nonsignificantly in the LV of CSQ mice compared with WT mice, yet the fraction of phosphorylated Mena relative to total Mena expression was significantly reduced by 40% in CSQ mice compared with WT controls (Fig. 2). These data show that the ratio of phosphorylated Mena to total Mena is significantly reduced in LV tissue of CSQ mice, which may contribute to the cardiac pathology of these animals.

Fig. 2.

Alteration of Mena phosphorylation in HF. A: LV lysates from WT or CSQ-overexpressing mice were immunoprecipitated (IP) with a monoclonal mouse Mena antibody or mouse IgG control. Western blots were then performed with a rabbit phosphorylated Mena (phospho-Mena) antibody (blot on top) or a mouse Mena antibody (blot on bottom). B: blot quantification showing densitometry analysis of phospho-Mena to total Mena expression. Data are expressed as mean arbitrary units ± SE (n = 3 animals/group; *P < 0.05).

Cardiac Mena is localized to ICD structures (19), which connect single myocytes to mechanical and electrochemical syncytium. To assess the spatial distribution of Mena in murine ventricles, immunostaining was performed in WT and CSQ mouse heart sections (Fig. 3, A and B, respectively). Consistent with prior reports in WT mice, Mena was predominantly expressed at the ICD, along the short edge of neighboring myocytes (Fig. 3, A and B). In contrast to WT mice, Mena staining in CSQ mice appeared brighter in ICD regions with more diffuse distribution toward the interior of cardiomyocytes, corroborating the increased Mena expression evident in Western blots (Fig. 1). Furthermore, CSQ mice displayed myocyte disarray characteristic of failing myocardium, in sharp contrast to a fairly regular arrangement of muscle cells in WT LV tissue. Our well-validated Mena antibody (33) allowed us to confirm the absence of Mena expression in Mena^−/− mice (Fig. 3C).

Fig. 3.

Mena upregulation occurs in the intercalated discs (ICD). Laser scanning confocal images of LV sections from WT (A), CSQ (B), or Mena knockout (Mena^−/−, C) mouse hearts incubated with polyclonal rabbit Mena antibody followed by Texas Red-conjugated goat anti-rabbit IgG. At the top right corner of each image, tissue autofluorescence (green) shows longitudinal orientation of the cardiomyocytes. Samples were visualized by confocal microscopy using identical light exposure and image capture settings. Immunostainings show that Mena is localized at the ICD in WT heart (A), whereas Mena expression is increased both at the ICD and throughout the cardiomyocytes of CSQ failing heart (B). Absence of Mena expression in Mena^−/− confirms the specificity of the antibody (C). Arrows show Mena expression predominantly at the ICD, along the short edge of neighboring myocytes. Asterisks show where Mena staining in CSQ mice appeared brighter in ICD regions with more diffuse distribution toward the interior of cardiomyocytes.

Mena^−/− mice have reduced basal cardiac function.

Mena^−/− mice are viable and fertile, although delays in the formation and maturation of brain nerve fibers have been observed (33, 39). We evaluated the cardiac phenotype of Mena^−/− mice to investigate the role of Mena in myocardial function.

Cardiac function and dimensions were assessed in conscious WT and Mena^−/− mice by echocardiography at 2, 4, and 6–8 mo of age [Fig. 4 and Supplemental Tables 1, 2, and 3 (Supplemental data for this article may be found on the American Journal of Physiology: Heart and Circulatory Physiology website.)]. LV systolic function, as determined by fractional shortening (FS, %), was ∼15% lower in 2-mo-old Mena^−/− mice compared with age-matched WT littermates. FS further deteriorated to ∼83% of WT animals by 6–8 mo, despite no differences in heart rate, end-diastolic dimensions, or wall thickness between groups. Additionally, mean velocity of circumferential fiber shortening rates were significantly slower and LV end-systolic dimensions were significantly larger in Mena^−/− mice (Supplemental Tables 1, 2, and 3). These results show that Mena^−/− mice exhibit impaired cardiac performance, with the greatest dysfunction apparent in the oldest animals.

Fig. 4.

Reduced cardiac function in Mena^−/− mice compared with control WT mice. A: heart function was assessed in WT and Mena^−/− mice via conscious M-mode echocardiography at indicated ages. Mena^−/− mice (filled bars) showed significant reduction in fractional shortening at all time points compared with age-matched, littermate WT animals (open bars); n = 6–7 animals for each group. *P ≤ 0.05 compared with WT mice. B: representative echocardiographic images.

Mena deletion results in cardiac hypertrophy with morphological and ultrastructural abnormalities.

As assessed by the heart weight to body weight ratio, Mena^−/− mice have significantly larger hearts than WT mice, indicative of cardiac hypertrophy (Fig. 5A). Importantly, the average body weights of the two groups were quite similar (23.21 ± 7.54 g for WT and 22.11 ± 4.75 g for Mena^−/− mice). Histological comparison of WT and Mena^−/− LV sections revealed clear morphological distinctions, highlighted by conspicuous fibrotic areas (Fig. 5Cb), poorly aligned myocytes (Fig. 5Cd), and the presence of perivascular cellular infiltrates (Fig. 5Cf) in Mena^−/− ventricular tissue (Fig. 5C).

Fig. 5.

Hypertrophy and myocyte disarray in Mena^−/− mouse hearts. A: heart weight-to-body weight ratios of WT and Mena^−/− mice at 8 mo of age. Data are expressed as means ± SE; n = 6 animals for each group, *P ≤ 0.05. B: hematoxylin and eosin (H & E) staining of representative WT and Mena^−/− mouse hearts at ×1.25 magnification. C: Masson's Trichrome staining (a-d) and H & E staining (e and f, magnifications of the boxed areas indicated in B) of 6-mo-old WT (a, c, and e) and Mena^−/− (b, d, and f) heart sections. Note the enhanced expression of connective tissue (blue staining), indicative of fibrosis (b vs. a), myocyte disarray (d vs. c), and increase in nuclei (blue staining), showing cellular accumulation around vessels (f vs. e), in Mena^−/− mice. Magnifications: a, b, e, and f = ×10; c and d = ×20.

Mena's localization in the ICD suggests that it may be important for linkage to the cytoskeleton and potential stabilization of cell-to-cell interactions. Ultrastructural analysis was performed in Mena^−/− LV tissue using electron microscopy with representative photomicrographs displayed in Fig. 6. The arrows indicate the interface of adjacent myocytes, which appears as an undulating line with ICD structures evident by the presence of dark electron-dense regions (desmosomes). Figure 6A shows that WT tissue is characterized by fairly regular arrangement of rectilinear Z-lines, parallel to the ICD, and oblong mitochondria aligned with the myofilaments. This pattern is altered in Mena^−/− tissue, since the ICD and Z-lines are crooked, poorly defined, and arranged at odd angles to each other and the mitochondria. Additionally, mitochondria appear shrunken and more irregularly shaped (Fig. 6B). At higher magnification, glycophospholipid accumulation [which has been associated with myocardial hypertrophy and conduction dysfunction (18)] near ICDs is apparent in Mena^−/− tissue but not in WT tissue (Fig. 6, C and D). These results may suggest that loss of Mena leads to cardiac hypertrophy and fibrosis along with disorganized myofibrillar and ICD structure.

Fig. 6.

ICD disorganization in Mena^−/− mouse hearts. Representative electron micrographs from the LV of WT (A and B) and Mena^−/− (C and D) mice. Bar scales are shown on the bottom right corner of each image. White arrows (WT) or black arrows (Mena^−/−) are placed on either side of the myocyte-myocyte junction. Selected mitochondria, Z-lines, and glycophospholipids are indicated by M, Z, and G, respectively.

Cardiac conduction abnormalities in Mena^−/− mice.

ICD structures facilitate the propagation of electrical signals throughout the myocardium. Altered ICD in Mena^−/− mice could predispose these animals to abnormal cardiac conduction and arrhythmias. To test this, ECG recordings were obtained from conscious adult Mena^−/− and WT mice. As shown in Fig. 7, A and B, Mena^−/− mice displayed prolonged PR and QRS intervals compared with controls, providing evidence for abnormal conduction. No changes in heart rate or RR intervals were observed between groups.

Fig. 7.

Cardiac conduction abnormality in Mena^−/− mice. A: representative electrocardiogram recordings obtained from conscious 6-mo-old WT and Mena^−/− mice. B: graphs showing PR and QRS interval durations (ms). Data are expressed as means ± SE (n = 5 mice/group; *P < 0.01 and **P < 0.001). C: representative optical cardiac conduction activation maps. On left is a cartoon of the heart illustrating its position in relation to the charge-coupled device camera. RV, right ventricle. Middle and right show activation maps of epicardial wave propagation during sinus rhythm in WT and Mena^−/− mice, respectively. The stars indicate the sites of earliest breakthrough on the LV and RV.

We specifically evaluated excitation wave propagation during intrinsic rhythms using optical mapping of Langendorff perfused, isolated Mena^−/−, and WT hearts (n = 3/group). Representative epicardial activation maps are shown in Fig. 7C. Mena^−/− mice displayed abnormal ventricular depolarization highlighted by slowing of wavefront propagation from the apex across the heart. Note that a significant portion of the Mena^−/− heart is colored blue or purple, corresponding to a longer conduction interval. Indeed, right bundle branch block was observed in two Mena^−/− mice, suggesting rhythm disturbance and increased arrhythmia propensity in these animals. Furthermore, the electrical wavefront often penetrated the epicardium at only one site in Mena^−/− hearts, highlighting the suboptimal propagation and the potential for uncoordinated contraction across the ventricles. Collectively, these data suggest that loss of Mena results in slowed electrical conduction and may predispose Mena^−/− mice to cardiac arrhythmias and contractile dysfunction.

Loss of Mena affects expression and localization of ICD proteins.

Because Mena^−/− mice displayed conduction abnormalities and alterations in ICD structure, we wondered whether changes in expression or localization of adherens junction, cell-matrix, or gap junction proteins contributed to these characteristics. Accordingly, several major ICD proteins were examined via immunoblot (Fig. 8) and confocal analyses (Fig. 9). LV lysates revealed a significant decrease in vinculin expression (Fig. 8), a known binding partner of Mena, in parallel with loss of Mena expression. On the other hand, the gap junction protein Cx43 was significantly upregulated in Mena^−/− compared with WT mice (Fig. 8). Expression of adherens junction proteins β-catenin, N-cadherin, and zyxin, a well-defined Mena-interacting partner (20), as well as desmosome protein plakophilin-2 (47) did not differ between groups (Fig. 8, Supplemental Fig. 1, and data not shown). Furthermore, we did not detect compensatory changes in expression or phosphorylation of the Mena family member VASP in Mena^−/− myocardium (Fig. 8).

Fig. 8.

ICD protein expression in myocardium of WT and Mena^−/− mice. A: representative Western blots performed on LV lysates from WT and Mena^−/− mice. Vertical line between connexin 43 (Cx43) bands indicates removal of an empty lane from the image. VASP, vasodilator-stimulated phosphoprotein; p-VASP, phosphorylated VASP; TCx43, total Cx43. B: densitometric analysis was performed by normalizing expression of each protein of interest to simultaneously measured GAPDH. Data are expressed as mean arbitrary units ± SE; n = 4–5 mice/group; *P < 0.05 and **P < 0.01.

Fig. 9.

Cx43 lateralization in Mena^−/− mice. A: laser scanning confocal microscopy images of LV sections from WT and Mena^−/− mice. Cx43 is pseudocolored in yellow, and the nucleus is stained blue (TOPRO). Green autofluorescence shows longitudinal orientation of the cardiomyocytes. Cx43 is predominantly expressed on the ends of myocytes (asterisks) in WT tissue (left). In contrast, Mena^−/− hearts displayed enhanced Cx43 expression along the lateral sides of myocytes (arrows). B: analysis of Cx43 expression was performed on a composite Z-stack image from 10–11 images taken at 0.5-μm intervals. Measurements were obtained from three separate images from different regions of the LV for each heart (5 WT and 5 Mena^−/−) by an investigator blinded to the phenotype. Cx43 expression was quantified in pixel/μm² using ImagePro 6.2 software. Lateralized Cx43 expression (Cx43 staining parallel to the orientation of the cells) was normalized to the total Cx43 expression in the image field. Mena^−/− hearts displayed a trend toward increased Cx43 lateralization compared with WT hearts (P = 0.13).

To evaluate the spatial distribution of Cx43 and vinculin, we performed confocal microscopy on WT and Mena^−/− myocardial sections (Fig. 9 and Supplemental Fig. 2). In WT hearts, Cx43 was typically present in ICD regions along the short sides of apposing cardiac myocytes (Fig. 9A). In Mena^−/− hearts, however, we found a heterogeneous Cx43 distribution and lateralization of Cx43 along the side-to-side junction of myocytes (Fig. 9A) that has been associated with cardiac conduction abnormalities and HF (2, 22, 27). To further evaluate Cx43 lateralization on Mena deletion, we used a pixel-counting algorithm to calculate lateralized Cx43/total Cx43 in LV cardiac sections. Figure 9B shows that Cx43 lateralized more frequently in Mena^−/− hearts compared with WT, yet the difference was not significant when normalized to total Cx43 expression, perhaps because of the heterogeneity and concomitant increase in total Cx43 (Fig. 8). Vinculin localized to ICDs and costameres of cardiac muscle in WT hearts; a general widening and disorganization of ICDs were apparent in Mena^−/− mice (Supplemental Fig. 2). Overall, these results suggest that Mena deletion results in remodeling of vinculin and Cx43, which may, in part, explain the cardiac conduction and dysfunction phenotypes.

DISCUSSION

There has been a paucity of knowledge regarding the role of Mena in the heart. Here, we demonstrate an important role for Mena in regulating cardiac function, conduction, and myocardial ICD structure, with the following observations: 1) cardiac Mena protein expression was upregulated in human HF and in two mouse models of HF; 2) decreased proportion of phosphorylated Mena in failing myocardium; 3) decreased cardiac function in Mena^−/− mice, coupled with hypertrophy; and 4) cardiac conduction slowing in Mena^−/− mice, correlated with altered ICD structure, cardiac fibrosis, and remodeling of ICD proteins.

Altered Mena expression and phosphorylation is associated with HF.

Our prediction that Mena protein levels would increase correspondingly with its mRNA expression pattern was confirmed by Western blot and immunofluorescence in human and murine HF (Figs. 1 and 2). Consistent with our prior microarray data, as well as with protein data presented herein, increased Mena mRNA expression has been identified by microarray in other HF models, including HF induced by cardiac-restricted P38 and JNK activation (40), as well as the G_qα-overexpressing mouse cardiomyopathy model (38). Recently, increased Mena protein expression was also identified in WT mice after ascending aorta constriction (46), suggesting a role for Mena in the hypertrophied heart. Because Mena expression correlated directly with the HF phenotype, we were somewhat surprised to discover that Mena knockout resulted in cardiac structural abnormalities, hypertrophy, and dysfunction (Figs. 4–7). Furthermore, three of four Mena^−/− mice (n = 4) subjected to ischemia-reperfusion injury succumbed to fatal perisurgical cardiac arrhythmias, whereas all WT mice survived (n = 3, data not shown). To further probe Mena's role in the development of HF while avoiding surgically induced mortality, Mena^−/− were bred with CSQ/Mena^+/− mice. Despite 87 offspring, we were unable to generate any viable CSQ/Mena^−/− offspring. Because the Mena locus (chromosome 1) lies outside the susceptibility loci (chromosomes 2 and 3) for CSQ mice (52), these findings suggest increased susceptibility to cardiac dysfunction on Mena deletion.

Numerous cytoarchitectural proteins, which perform the critical function of transmitting mechanical and electrical stimuli between cells, are known to be upregulated in failing myocardium (21). The fact that alterations of the cytoskeleton are commonly associated with HF (23) indicates that this may be a compensatory mechanism in response to the loss of contractile machinery and related proteins (e.g., titin and α-actinin). Likewise, we hypothesize that increased expression of Mena, a cytoskeleton-associated protein, is an initially adaptive response, since Mena levels tend to normalize with targeted HF treatments (Fig. 1).

Phosphorylation of Mena is critical for its functional regulation of the cytoskeleton and cell motility (31, 35). Our data indicate that, while Mena expression is enhanced in HF, the proportion of phosphorylated Mena is substantially reduced (Fig. 3). Excessive activation of the adrenergic response is a hallmark in HF (11); pathological desensitization and dysregulation of adrenergic signaling during HF may be associated with decreased PKA activity (20), resulting in reduced Mena phosphorylation. Intriguingly, VASP phosphorylation (7, 46) was identified as an important modulator of cardiac function and cytoskeletal remodeling. Future studies will assess whether Ena/VASP proteins, like several other cytoarchitectural proteins, may interact with various A-kinase anchoring proteins, including Ezrin (14) and Synemin (44). These interactions may aid in the regulation of Mena phosphorylation and therefore cardiac function. Further illumination of the role of Mena, its phosphorylation state, and protein-protein interactions in these processes is worthy of future investigation.

Mena^−/− mice develop cardiac dysfunction and hypertrophy.

We demonstrate that Mena deletion causes cardiac hypertrophy, dysfunction, and conduction slowing (Figs. 4–7), suggesting Mena may be required for normal cardiovascular function. Interestingly, VASP null mice have no such baseline cardiac phenotype. Furthermore, they have no apparent alterations in cardiac morphology, nor do they respond different from WT animals to cardiac pressure overload induced by aortic constriction (46). However, combined Mena and VASP knockouts die during embryonic development (5). On the other hand, mice develop dilated cardiomyopathy upon cardiac-specific overexpression of the Ena/VASP homology domain, EVH1, to displace VASP and Mena from cardiac ICDs (15). Although this domain would nonspecifically target a host of EVH1 domain-binding proteins [e.g., fat/cadherin, zyxin, vinculin (31)], the observed effects may in part result from Mena mislocalization. Regardless, all data indicate that Mena may be critical for cardiac cytoskeletal stability.

Altered ICD proteins and slowed conduction in Mena^−/− mice.

Alteration of ICD proteins, which affect both the structural integrity of the ICD and/or transmission of intercellular signals, generally promote dilated cardiomyopathy in mice (16, 58). Electron microscopy in Mena^−/− mice demonstrated widened and disorganized ICDs with reduced desmosomes and glycophospholipid accumulation near the ICD structure (Fig. 6). These alterations appear to occur postnatally, since the embryonic lethal triple-null Mena/VASP/Evl embryos have apparently normal cardiac and ICD structure (17). Importantly, Mena^−/− mice displayed slowed electrical conduction at intrinsic rhythm in vivo (Fig. 7). Correlated with these findings was significantly altered expression of two ICD-associated proteins, vinculin and Cx43, in Mena^−/− mice (Fig. 8). Moreover, Cx43 (but not vinculin) expression appeared to be increased at the ICD and redistributed to the lateral membranes of cardiac myocytes (Fig. 9). No change in expression was noted for any of the other ICD proteins assayed, including VASP (Fig. 8). Cardiac dysfunction in Mena^−/− mice, despite normal VASP expression in the heart, suggests there may be Mena-specific cardiac functions. Indeed, cardiac dysfunction appears to result when cardiac Ena/VASP (particularly cardiac Mena) expression is reduced below a certain threshold, although this level is not reached in either Mena^+/− (where there was no morphometric or functional divergence from WT) or VASP^−/− animals (46).

Ventricular remodeling due to cardiac overload is characterized by changes in expression and redistribution of Cx43 (25), in part contributing to conduction slowing and arrhythmia propensity in HF (28, 30). Furthermore, chronic adrenergic stimulation (resulting in hypertrophic response) can increase myocardial Cx43 expression and gap junction conductance in vivo and in vitro (45). Changes in gap junction proteins are also associated with changes in cell-cell and cell-extracellular matrix interactions, as well as alterations in cell motility involving remodeling of the cytoskeleton (13). Disruption of adherens junctions directly contributed to rearrangement of Cx43 distribution, resulting in Cx43 lateralization (51). Given Mena's localization in the ICD and the effects on Cx43 remodeling upon genetic Mena deletion, it could be speculated that Mena modulates a signal(s) from cardiac sarcolemma to the actin cytoskeleton at ICD that modulates the expression and localization of Cx43.

Additional factors that may undermine cell-to-cell communication are tissue architecture [cell shape and size (50), fibrosis, dilatation, fat infiltration, and hypertrophy (54)], myocyte-fibroblast coupling (29), and sodium channel density (47), although we did not observe changes in sodium channel expression (Supplemental Fig. 3). Phosphorylation of Cx43 is important for gap junction regulation; dephosphorylation results in altered cell coupling and cardiac conduction abnormality in rabbit and human HF (1). Using an antibody that specifically identifies the nonphosphorylated isoform of Cx43, we were unable to detect any significant changes in Cx43 phosphorylation state (data not shown). Future studies will entail a systematic analysis of the His-Purkinje system of Mena^−/− mice, as well as intercellular coupling, and should provide further valuable insight into the mechanism of our conduction observations.

Decreased vinculin expression, which is associated with increased susceptibility to HF (58, 59), likely results from structural alteration of the ICD in Mena^−/− mice (Figs. 8 and 9). Interestingly, the Mena^−/− mice, vinculin^+/− mice (vinculin^−/− are embryonic lethal), and cardiac-restricted vinculin-null mice share highly similar phenotypes, including prolonged QRS complexes, conduction delays, widened ICDs, and lateralized Cx43, suggesting vinculin may play a key interactive role with these proteins in the control of normal cardiac conduction and maintenance of cardiomyocyte junctions (58, 59).

In summary, the current study identifies Mena protein expression as a novel indicator of murine and human HF. We suggest that increased Mena expression may initially be a cardioprotective adaptive mechanism, since deletion of Mena resulted in cardiac dysfunction, possibly mirroring the decreased proportion of phosphorylated Mena observed in HF. Future work, including study of cardiac-restricted Mena deletion and overexpression, may help further our understanding of how Mena, its phosphorylation, and potential interacting partners modulate cardiac conduction and function.

GRANTS

This work was supported by an American Heart Association (AHA) Postdoctoral Fellowship (09POST2190063, S. L. Belmonte); National Heart, Lung, and Blood Institute (NHLBI) Grants R01-HL-89885, 3R01-HL-089885-02S1, and R01-HL-091475; an AHA Scientist Development Grant (B. C. Blaxall); and NHLBI Grants PO1-HL-039707, PO1-HL-070074, and RO1-HL-080159 (J. Jalife).

DISCLOSURES

None.