JOURNAL OF THE AMERICAN COLLEGE OF CARDIOLOGY VOL.64,NO.25,2014

ª 2014 BY THE AMERICAN COLLEGE OF CARDIOLOGY FOUNDATION ISSN 0735-1097/$36.00

PUBLISHED BY ELSEVIER INC. http://dx.doi.org/10.1016/j.jacc.2014.09.071

Disturbance in Z-Disk Mechanosensitive Induced by a Persistent Mutant Myopalladin Causes Familial Restrictive

Anne-Cecile Huby, PHD,* Uzmee Mendsaikhan, MD,* Ken Takagi, MD,y Ruben Martherus, PHD,*

Janaka Wansapura, PHD,z Nan Gong, MD,* Hanna Osinska, PHD,* Jeanne F. James, MD, PHD,* Kristen Kramer, BS,*

Kazuyoshi Saito, MD,* Jeffrey Robbins, PHD,* Zaza Khuchua, PHD,* Jeffrey A. Towbin, MD,*

Enkhsaikhan Purevjav, MD, PHD*

ABSTRACT

BACKGROUND Familial restrictive cardiomyopathy (FRCM) has a poor prognosis due to diastolic dysfunction and restrictive physiology (RP). Myocardial stiffness, with or without fibrosis, underlie RP, but the mechanism(s) of restrictive remodeling is unclear. Myopalladin (MYPN) is a messenger molecule that links structural and regulatory molecules via translocation from the Z-disk and I-bands to the nucleus in cardiomyocytes. Expression of N-terminal MYPN peptide results in severe disruption of the sarcomere.

OBJECTIVES The aim was to study a nonsense MYPN-Q529X mutation previously identified in the FRCM family in an animal model to explore the molecular and pathogenic mechanisms of FRCM.

METHODS Functional (echocardiography, cardiac magnetic resonance [CMR] imaging, electrocardiography), morpho- histological, , and molecular studies were performed in knock-in heterozygote (MypnWT/Q526X) and homozygote mice harboring the human MYPN-Q529X mutation.

RESULTS Echocardiographic and CMR imaging signs of diastolic dysfunction with preserved systolic function were identified in 12-week-old MypnWT/Q526X mice. Histology revealed interstitial and perivascular fibrosis without overt hypertrophic remodeling. Truncated MypnQ526X was found to translocate to the nucleus. Levels of total and nuclear cardiac ankyrin repeat protein (Carp/Ankrd1) and phosphorylation of mitogen-activated protein kinase/ extracellular signal–regulated kinase 1/2 (Erk1/2), Erk1/2, Smad2, and Akt were reduced. Up-regulation was evident for muscle LIM protein (Mlp), , and heart failure (natriuretic peptide A [Nppa], Nppb, and myosin heavy chain 6) and fibrosis (transforming growth factor beta 1, alpha–smooth muscle , osteopontin, and periostin) markers.

CONCLUSIONS Heterozygote MypnWT/Q526X knock-in mice develop RCM due to persistence of mutant MypnQ526X protein in the nucleus. Down-regulation of Carp and up-regulation of Mlp and desmin appear to augment fibrotic restrictive remodeling, and reduced Erk1/2 levels blunt a hypertrophic response in MypnWT/Q526X hearts. (J Am Coll Cardiol 2014;64:2765–76) © 2014 by the American College of Cardiology Foundation.

From the *Heart Institute, Department of Pediatrics, Cincinnati Children’s Hospital Medical Center, Cincinnati, Ohio; yJikei University, Tokyo, Japan; and the zDepartment of Radiology, Imaging Research Center, Cincinnati Children’s Hospital Medical Center, Cincinnati, Ohio. This work was supported in part by research grants from the American Heart Association, Children’s Cardiomyopathy Foundation (to Drs. Towbin and Purevjav), John Patrick Albright Foundation, and the National Institutes of Health (R01 HL53392 and R01 HL087000; to Dr. Towbin). All authors have reported that they have no relationships relevant to the contents of this paper to disclose. Listen to this manuscript’s audio summary by JACC Editor-in-Chief Dr. Valentin Fuster. You can also listen to this issue’s audio summary by JACC Editor-in-Chief Dr. Valentin Fuster.

Manuscript received June 17, 2014; revised manuscript received August 18, 2014, accepted September 4, 2014. 2766 Huby et al. JACC VOL. 64, NO. 25, 2014 Z-Disk in Restrictive Cardiomyopathy DECEMBER 30, 2014:2765– 76

ABBREVIATIONS estrictive cardiomyopathy (RCM) to determine the pathophysiology and molecular AND ACRONYMS R accounts for approximately 5% of mechanism(s) of FRCM. diagnosed and is CARP/ANKRD1 = cardiac METHODS ankyrin repeat protein characterized by diastolic dysfunction and restrictive physiology (RP), whereas systolic DES = desmin function typically remains normal or near GENERATION OF KNOCK-IN MICE. The study con- ERK = extracellular signal– formed to the protocols approved by the Institutional regulated kinase normal (1).Thevolumeandwallthickness of the ventricles are usually normal or Animal Care and Use Committee at Cincinnati Chil- FRCM = familial restrictive ’ cardiomyopathy small, whereas atrial or biatrial enlargement dren sHospitalMedicalCenter. To generate a murine Mypn-Q526X mutation, we targeted exon 10 in the HF = heart failure occurs due to impaired ventricular filling Mypn gene (Online Figure 1A) using a homologous LV = left ventricular during diastole (2). Particularly in children, recombination method as described previously (10) MLP = muscle LIM protein RCM has the poorest prognosis among all types of heart muscle diseases, with 2- and anddetailedintheOnline Appendix. MYPN = myopalladin 5-year mortality rates of 50% and 70%, RCM = restrictive EVALUATION OF HEART FUNCTION IN MICE. Serial respectively, and the highest rate of sudden cardiomyopathy echocardiography and electrocardiography were per- cardiac death (SCD) (3). Survivors ultimately RP = restrictive physiology formed in mice at 6 and 12 weeks of age (12 animals/ develop heart failure (HF) due to RP, as well TGF = transforming group). Cardiac magnetic resonance imaging was growth factor as pulmonary hypertension; however, the performed in 12-week-old animals when mice showed mechanistic basis of RP with diastolic a markedly increased early (E) and late (A) diastolic fi dysfunction, myocardial brosis, and cardiac stiff- velocities (E/A) ratio, signs of “restrictive filling,” or ness is unclear. diastolic dysfunction by echocardiography. See the

SEE PAGE 2777 Online Appendix for experimental details. HISTOPATHOLOGY, IMMUNOHISTOCHEMISTRY, QUANTITATIVE A history of familial RCM (FRCM) is reported in REAL-TIME POLYMERASE CHAIN REACTION, AND ELECTRON approximately 30% of RCM cases, with autosomal MICROSCOPY. Histopathology, including hematoxylin dominant inheritance most commonly noted (4). and eosin, Masson trichrome, immunohistochemical, Several , typically encoding proteins of the transcriptional, and terminal deoxynucleotidyl tran- sarcomere, Z-disk, , or intermediate sferase dUTP nick end-labeling (TUNEL) analysis, filament network, have been associated with auto- was performed to assess structural, fibrotic, hyper- somal dominant FRCM (4). The myopalladin (MYPN) trophic, and/or apoptotic remodeling in the heart. gene, which is located at 10q21.3, en- Transmission electron microscopy was performed codes a 145-kDa protein that participates in linking on glutaraldehyde-perfusedheartsaspreviously regulatory molecules involved in sarcomeric I-band described (11).AfterisolationoftotalRNAfromven- and Z-disk assembly and muscle gene expression tricular tissues, quantitative real-time polymerase (5). The N-terminal MYPN interacts with cardiac chain reaction (PCR) was performed as described in ankyrin repeat protein (CARP/ANKRD1),atran- the Online Appendix.Six12-week-oldanimalsper scriptional coinhibitor of genes involved in the group were used. See experimental details in the development of HF and hypertrophy (6).MYPNhas Online Appendix. dual localization, sarcoplasm and nucleus, similar to PROTEIN EXPRESSION, PULL-DOWN, AND WESTERN that seen with CARP (5). At the Z-disk, MYPN BLOTTING. Human embryonic kidney (HEK293) cells interacts with alpha-actinin and with the SRC were transfected with different chimeras of MYPN– homology 3 domain of nebulette (7). Mutations in green fluorescent protein (GFP) and CARP-V5 com- the MYPN gene cause diverse phenotypes in plementary DNA (cDNA) to evaluate MYPN and CARP humans, including , hyper- interactions using immunoprecipitation and coim- trophic cardiomyopathy, and RCM (8,9).Weprevi- munoprecipitation. Cellular fractionation was per- ously reported a nonsense autosomal dominant formed using the NE-PER kit (Pierce, Rockford, mutation (MYPN-Q529X) that resulted in FRCM in Illinois). Western blotting was used for protein siblings via disturbed myofibrillogenesis and sarco- evaluation, and levels of proteins were quantified meric Z-disk disruption (9). in relative density units using ImageJ software as In this study, knock-in mice carrying a heterozy- described in the Online Appendix. gous and homozygous Mypn-Q526X mutation in exon 10 of the murine Mypn gene, homologous to STATISTICAL ANALYSIS. Statistical analysis, re- the human MYPN-Q529X mutation, were analyzed ported as mean standard error of the mean, was JACC VOL. 64, NO. 25, 2014 Huby et al. 2767 DECEMBER 30, 2014:2765– 76 Z-Disk in Restrictive Cardiomyopathy

performed with the Student t test or 1-way analysis the MypnWT/Q526X mice (0.51 0.03) compared with of variance using GraphPad5 software (GraphPad WT animals (0.56 0.03; p ¼ 0.02), consistent with Software, Inc., La Jolla, California). A probability LV remodeling in heterozygotes only. value of p # 0.05 was considered significant. Arrhythmias and conduction abnormalities are reportedly present in 15% to 30% of patients with RESULTS RCM, especially in pediatric patients (13). Electrocar- diography revealed that heart rate, pulse, pulse rate, MANIFESTATION OF DIASTOLIC DYSFUNCTION. Given and QRS durations were similar among mutant and that persistence of the 65-kDa MypnQ526X peptide WT mice, whereas T-wave duration was decreased in vivo may potentially cause a “poison peptide” in MypnWT/Q526X versus WT or homozygotes (Online effect, levels of the Mypn protein were confirmed Table 3). Arrhythmias, including premature atrial by Western blotting; the MypnWT/Q526X line 3 with contractions, premature ventricular contractions, the highest 65-kDa MypnQ526X protein expression and type II second-degree atrioventricular block was selected for further breeding (Online Figure 1B). (Figure 1C), were observed in heterozygotes only. Heterozygous and homozygous recombination of the FIBROSIS IN HETEROZYGOUS MYPNWT/Q526X HEARTS. Mypn-Q526X mutation was confirmed by sequencing Morphological evaluation of revealed of Mypn exon 10 in genomic DNA obtained from no hypertrophy in any animal (Figure 2A). Histologi- mouse tails (OnlineFigure1C). We verified Mypn cally, diffuse interstitial and perivascular fibrosis was mRNA transcription levels in the heart, skeletal found in MypnWT/Q526X ventricular myocardium only. muscle, liver, and kidneys by reverse transcriptase Dystrophin staining revealed intact sarcolemmal PCR. As expected, homozygotes had impaired Mypn integrity and normal cardiomyocyte size among mRNA transcription in heart and the different groups, confirming absence of car- compared with wild-type (WT) and heterozygote diomyocyte hypertrophy (Online Figure 2A). No littermates (Online Figures 1D and 1E). Thus, homo- positive TUNEL staining was seen in any group, zygous MypnQ526X mice were also generated as a excluding the involvement of apoptosis. Ultrastruc- model of Mypn gene ablation. turally, t-tubule enlargement and mild intercalated Both MypnWT/Q526X and MypnQ526X mice were born disk disruption were seen in heterozygous and and grew normally, with no skeletal myopathy homozygous animals (Figure 2B). observed up to 12 weeks of age. At 6 weeks, increased Taken together, these results showed that hetero- E/A ratios, features of RP in humans (12),were zygote MypnWT/Q526X mutants demonstrated fibrotic detected in MypnWT/Q526X mice compared with WT remodeling without an overt hypertrophic response, and homozygotes. The same cohort of mice under- accurately recapitulating the human RCM phenotype. went a serial echocardiography at 12 weeks of age, Homozygous MypnQ526X mice had no or a minimal andtheE/Aratioincreasedsignificantly (Figure 1A). phenotype as result of MYPNQ526X ablation. There- Although systolic function and chamber dimensions fore, further comparative molecular studies were were preserved, MypnWT/Q526X mice showed signs of carried out in MypnWT/Q526X and WT littermates. impaired diastolic filling of the left ventricle (LV), including decreased end-diastolic volume (LVEDV) PUTATIVE Z-DISK STRESS-SENSOR DYSREGULATION IN and internal dimensions compared with WT and MYPNWT/Q526X HEARTS. Immunohistochemical anal- homozygotes (Online Table 1). ysis revealed no differences in Mypn at the Z-disks. Further, cardiac magnetic resonance imaging was Notably, coincident delocalization of Carp and used in the same cohort of 12-week-old mice to pre- desmin (Des) to the periphery was found in cisely evaluate cardiac anatomy and diastolic func- MypnWT/Q526X cardiomyocytes only (Figure 3, arrows). tion. No significant differences between groups were On a protein level, although no MypnWT was seen, the found in circumferential strain, strain rate, torsion, or 65-kDa MypnQ526X was detected in the hearts of het- torsion rate (data not shown). The left atrial cross- erozygous mice as expected (Figure 4A). A significant sectional area of MypnWT/Q526X mice (3.8 0.5 mm2) increase in the expression of the Mypn homolog, pal- waslargercomparedwiththatofWTmice(3.5 ladin (90 kDa), was evident in MypnWT/Q526X hearts and 0.6 mm2)andhomozygotes(3.0 0.3 mm2;p¼ 0.02), confirmed by monoclonal and immunohis- as measured in the 4-chamber view (Figure 1B, Online tochemical staining (Online Figure 2B). Nebulette and Table 2). The LVEDV was significantly lower in the alpha-actinin were mildly up-regulated, whereas Carp MypnWT/Q526X mutants compared with WT mice was considerably down-regulated in MypnWT/Q526X (p ¼ 0.03). The sphericity index, the ratio of the short- (p ¼ 0.03). Levels of calpain 3 were similar in mutants to long-axis dimensions of the LV, was also lower in and WT, excluding the likelihood of Carp degradation 2768 Huby et al. JACC VOL. 64, NO. 25, 2014 Z-Disk in Restrictive Cardiomyopathy DECEMBER 30, 2014:2765– 76

FIGURE 1 Phenotypic Characterization of Mypn Mice

ABWT WT MypnWT/Q526X MypnQ526X

MypnWT/Q526X E A

E A

3 *

2

1 MV E/A (AU) MV E/A

0 WT MypnWT/Q526X MypnQ526X

C 0.5 [sec]

WT 0.5 Volts WT/Q526X Mypn

(A) M-mode images of mitral valve (MV) movement by 2-dimensional echocardiography indicate an increase of early and late diastolic velocities (E/A) ratios in MypnWT/Q526X compared with wild-type (WT) mice or homozygotes, which became significant in 12-week-old heterozygote mutants (blue column) compared with WT and homozygote mice. (B) Cardiac magnetic resonance images in the 4-chamber view demonstrate enlarged left atria in MypnWT/Q526X mice (middle columns) compared with WT (left columns) and MypnQ526X (right columns) mice. (C) Electrocardiography in WT (a) and mutant (b-d) mice. MypnWT/Q526X mice display arrhythmias (arrows), including (b) premature atrial contractions, (c) premature ventricular contractions, and secondary atrioventricular block (d; 1:7 Wenkebach). *p < 0.05. AU ¼ arbitrary units; Mypn ¼ myopalladin. JACC VOL. 64, NO. 25, 2014 Huby et al. 2769 DECEMBER 30, 2014:2765– 76 Z-Disk in Restrictive Cardiomyopathy

FIGURE 2 Morphohistology of Hearts From 12-Week-Old Mice

(A) No significant chamber enlargement or cardiac muscle hypertrophy is seen in all groups (upper panels). Sections stained with hematoxylin and eosin (H&E; 10 magnification) demonstrate significant interstitial and perivascular infiltration in the MypnWT/Q526X myocardium (middle panel), with focal loss of myofibrils (arrow) compared with WT (left panel) and homozygotes (right panel). Masson trichrome staining (20 magnification) reveals interstitial and perivascular fibrosis in MypnWT/Q526X myocardium. Right lower inset panel displays enlarged details. (B) Transmission electron microscopy demonstrates no difference in the sarcomeric and Z-disk organization between groups (left panels, 24,000 magnification). Enlarged t-tubules (arrows) and widened cell-cell junctions with excess convolutions (asterisks) in both mutants are shown. Other abbreviations as in Figure 1. 2770 Huby et al. JACC VOL. 64, NO. 25, 2014 Z-Disk in Restrictive Cardiomyopathy DECEMBER 30, 2014:2765– 76

FIGURE 3 Immunohistochemistry of Mouse Hearts

MYPN (red) was coimmunostained with cardiac ankyrin repeat protein (CARP) (A) and desmin (DES) (B), both shown in green. Right panels represent the merged images with diamidino-2-phenylindole (blue). Although no changes in MYPN localization are noted, the striated costaining pattern in yellow was lost in MypnWT/Q526X hearts, indicating delocalization of CARP and DES from the cytosol to the periphery of cardiomyocytes (arrows). Other abbreviations as in Figure 1. JACC VOL. 64, NO. 25, 2014 Huby et al. 2771 DECEMBER 30, 2014:2765– 76 Z-Disk in Restrictive Cardiomyopathy

FIGURE 4 Protein Analysis of Mouse Hearts

(A and B) Western blotting images of WT (left lanes) and MypnWT/Q526X (right lanes) hearts. Significant changes in protein expression from WT were indicated in fold changes in relative density. Expression of 65-kDa MypnQ526X peptide is indicated by the arrow. Glyceraldehyde 3-phosphate dehydrogenase (GAPDH) was used as a reference protein. (C) Hypothetical pathogenic mechanisms of diastolic dysfunction caused by the heterozygous Mypn-Q526X mutation. The left panel depicts an idealized model of normal diastolic function. MYPN-CARP at the “central I-band” mechano-transmits the titin-PEVK–based myofibrillar stress/strain signals to the nucleus. MYPN connects with NEBL-SH3 and alpha- actinin (ACTN2), further interacting with actin, gamma-filamin, and titin-Z-repeats, which are dependent on ERK2 phosphorylation. In MypnWT/Q526X mice (right panel), levels of mechanosensory molecules are perturbed. CARP and MEK1/2–ERK1/2 levels are reduced; MLP and DES are up-regulated. CARP-dependent genes such as natriuretic peptide A (NPPA), NPPB, and myosin heavy chain (MYH) are up-regulated, whereas ERK1/2 downstream genes SMAD1/2 and AKT are down-regulated. *p < 0.05. **p < 0.01. CALP3 ¼ calpain 3; ERK ¼ extracellular signal–regulated kinase; Mab ¼ monoclonal antibody; MEK ¼ mitogen-activated protein kinase/ERK kinase; MLP ¼ muscle LIM protein; NEBL ¼ nebulette; PALLD ¼ ; SH3 ¼ SRC homology 3; other abbreviations as in Figures 1 and 3. 2772 Huby et al. JACC VOL. 64, NO. 25, 2014 Z-Disk in Restrictive Cardiomyopathy DECEMBER 30, 2014:2765– 76

due to increased protease activation. Des and muscle was detected in the nuclear fraction of heterozygote LIM protein (Mlp/cysteine-rich protein 3 [Csrp3]), the hearts, but very little was detected in homozygote putative Z-disk stress-sensors, were significantly hearts, suggesting a potential cause of the RCM up-regulated in heterozygotes versus WT (p < 0.05). phenotype. A concordant decrease in total Carp levels Mild down-regulation of alpha-tubulin, caveolin 3, was detected in MypnWT/Q526X hearts. and vinculin occurred in MypnWT/Q526X compared with For further clarification of MYPN-CARP in- WT, suggesting early recruitment of t-tubules and teractions, different combinations of human MYPNWT adherens junctions, respectively (Online Figure 2B). and MYPNQ529X cDNAs fused with GFP or mixtures of Intercalated disk proteins such as desmoplakin 2, both MYPNWT and MYPNQ529X (MYPNWT/Q529X)were connexin 43, or N-cadherin were not affected. Levels cotransfectedwithahumanCARP-V5constructin of cardiac troponin I (cTnI) and phospho-cTnI were HEK293 cells and pulled down. As shown in Figure 5B, normal, consistent with those of intact sarcomeres. the GFP antibody pulled down the 65-kDa MYPNQ529X as did the 145-kDa MYPNWT in all transfected ANALYSIS OF DOWNSTREAM SIGNALING PATHWAYS. cells, except controls and CARP-transfected ones. In Because cardiac stretch is regulated mainly at the contrast, levels of V5-pulled 40-kDa CARP were Z-disks by extracellular signal–regulated kinase 2 reduced in MYPNQ529X cellscomparedwiththosein (ERK2)–dependent phosphorylation of titin at the MYPNWT. Further, the 40-kDa CARP-V5 was unde- Zis1/Z-repeats (14), we sought to determine whether tectable in MYPNWT/Q529X mixedlysates,presumably the ERK cascade is altered. Reduced levels of due to greater reduction of CARP expression phosphorylated–mitogen-activated protein kinase compared with mutant MYPNQ529X cells. These (MAPK)/ERK1/2 (pMek1/2) and pErk1/2 were seen in effects support the in vivo data of reduced murine MypnWT/Q526X heartscomparedwiththoseofWT Carp levels in heterozygous hearts and the idea (Figure 4B), whereas other potential MAPK-associated that nuclear 65-kDa MypnQ526X may suppress Carp targets, including p53, c-Jun N-terminal kinase, p38, expression. focal adhesion kinase, nuclear factor kappa–light Subsequently, levels of Carp, Mlp,andDes mRNA chain enhancer of activated B cells, signal transducer were evaluated in the myocardial tissues of hetero- and activator of transcription 3, ERBB4, and trans- zygote and homozygote mice to further elucidate forming growth factor (TGF)-b1, were not affected whether differential phenotypes were associated with (Online Figure 2C). However, there was less phos- nuclear MypnQ526X-mediated alteration in transcrip- phorylation of Smad2 and Akt in MypnWT/Q526X ani- tional activities of mechanosensitive molecules. As mals.Further,CARP-dependent genes, including shown in Figure 5C, divergent changes in CARP mRNA natriuretic peptide A (Nppa), Nppb,andmyosinheavy levels were detected in MypnWT/Q529X and MypnQ529X chain 6 mRNA (Online Table 4), were more highly hearts compared with WT controls. Carp transcription expressed in MypnWT/Q526X hearts, supporting the was reduced in MypnWT/Q529X,whereasMypnQ529X idea of diminished inhibitory effects of CARP due to mice presented Carp up-regulation, and this variation reduced levels of CARP protein (Figure 4C). Signifi- between mutants was statistically significant. In cant increases in the expression of fibrotic (a-smooth contrast, no significant difference in transcription of muscle actin, Tgfb1, periostin, and osteopontin), Mlp and Des was revealed in any groups. inflammatory (vascular cell adhesion protein, inter- cellular adhesion molecule, and interleukin-1b), and DISCUSSION antiapoptotic (B-cell lymphoma 2) genes were iden- tified in MypnWT/Q526X hearts. The characteristic features of autosomal dominant PERSISTENCE OF NUCLEAR 65-KDA MYPN PEPTIDE FRCM result in a high-risk disorder with poor clinical ASSOCIATED WITH RCM PHENOTYPE. Mouse hearts outcomes, including HF, arrhythmias, and SCD (2). from all groups (WT and heterozygote and homozy- These features include diastolic dysfunction and gote mutants) were fractionated and levels of total, restrictive ventricular filling, with limited ability to nuclear, and cytoplasmic Mypn, palladin, and Carp augment cardiac stroke volume due to increased proteins were quantitated to determine whether the myocardial stiffness and interstitial fibrosis; how- location and expression levels of the proteins were ever, a mechanistic etiology remains unclear (4). associated with the RCM phenotype in vivo. Compa- MYPN is a sarcomeric protein localized at the Z-disk, rable expression of nuclear palladin was documented I-band, and nucleus. Expression of the N-terminal in all groups (Figure 5A), excluding the likelihood of region of MYPN in cardiomyocytes resulted in severe pathological effects of increased palladin levels in the disruption of the sarcomere (5), whereas N-terminal nucleus. Interestingly, the 65-kDa MypnQ526X peptide mutation MYPN-Y20C reduced CARP-MYPN binding JACC VOL. 64, NO. 25, 2014 Huby et al. 2773 DECEMBER 30, 2014:2765– 76 Z-Disk in Restrictive Cardiomyopathy

and caused hypertrophic cardiomyopathy in vivo FIGURE 5 Expression and Interaction of MYPN and CARP (Central Illustration)viaperturbedMYPNnuclear shuttling and abnormal assembly of terminal Z-disks within intercalated disks (9). Mutations in MYPN, A Cyto Nucleus Heart C Carp/Gapdh therefore, cause diverse cardiomyopathic phenotypes 2.0 p=0.05* p=0.25 within a critical “final common pathway” (9,15). 1.5 Q526X Q526X WT/Q526X WT/Q526X WT/Q526X Q526X The objectives of this study were to uncover the 1.0 Mypn Mypn WT Mypn WT Mypn WT Mypn molecular pathogenetic mechanism(s) and discover Mypn 0.5 n=4 n=4 n=4 the associated key signaling molecules in MYPN- PALLD Changes Fold → 0.0 Q529X mutation–induced restrictive remodeling 65kDa WT MypnWT/Q526X MypnQ526X using a knock-in mouse model. We demonstrated that → Q526X Csrp3/Gapdh homozygous Mypn mice displayed no phenotype CARP 1.5 due to ablation of the mutant protein. Heterozygous TBP WT/Q526X Mypn mutants, corresponding to a human GAPDH 1.0 heterozygous MYPN-Q529X mutation, recapitulated 0.5 the clinical and pathological features of human B n=4 n=4 n=4 Fold Changes Fold FRCM, including diastolic dysfunction with pre- WT Q529X WT/Q529X 0.0 served systolic function by 12 weeks of age. Intersti- WTMypnWT/Q526X MypnQ526X WT Q529X WT/Q529X tial and perivascular fibrosis was seen before overt MYPN MYPN CARP CARP+MYPN control MYPN CARP+MYPN cardiac contractile dysfunction, whereas myocyte CARP+MYPN Des/Gapdh 1.5 hypertrophy, apoptosis, and necrosis were absent. 145kDa Pull-GFP From these data, we hypothesize that the RCM WB-GFP 1.0 phenotype results from persistence of dysfunctional ←65kDa Q526X 0.5 truncated Mypn protein and consequent multi- Pull-V5 * 40kDa Changes Fold n=4 n=4 n=4 Q526X WB-V5 ple pathological “hits.” First, Mypn translocates 0.0 to the nucleus, probably perturbing levels of Carp WT MypnWT/Q526X MypnQ526X expression (Central Illustration). Second, MypnQ526X causes alterations in Mlp/Csrp3 and Des and blunts (A) Levels of MYPN and CARP in cytoplasmic and nuclear fractions and whole lysates of the phosphorylation of Mek1/2–Erk1/2, likely via mouse hearts. GAPDH was used as a cytosolic reference and TATA-box binding protein Q526X negative nuclear feedback. Further, activation of (TBP) as a nuclear reference protein. The 65-kDa Mypn peptide is detected (arrow) in the nucleus, and total CARP levels decrease in MypnWT/Q526X hearts. The 65-kDa Smad2 and Akt are reduced, likely to compensate for MypnQ526X peptide is degraded, whereas levels of nuclear CARP are increased in MypnQ526X “fi ” the mutation-induced nal common pathway of samples. (B) Immunoprecipitation studies in human embryonic kidney 293 cells. The RCM (15). expressed human proteins were pulled down with anti–green fluorescent protein (GFP) or Balanced levels of CARP are reported to be essential for anti-V5 and detected by Western blotting with anti-GFP (upper panel) or Q529X proper adaptive responses to different stresses in striated anti-V5 (lower panels). Reduced levels of pulled-down 65-kDa MYPN compared with 145-kDa MYPNWT protein in MYPNQ529X cells and MYPNWT/Q529X mixtures are indicated by muscle (Central Illustration), including exercise (16),stretch arrows. Double arrows indicate reduced levels of 40-kDa CARP-V5 in MYPNQ529X cells (17), cytokines (18), hypoxia (19), adrenergic stress (20), and compared with MYPNWT cells, suggesting suppressive effects of mutant MYPNQ529X on anthracycline (doxorubicin) toxicity (21). Multiple mecha- CARP levels. Asterisks indicate a failure of anti-V5 in pulling down 40-kDa CARP-V5 in the nisms, including 26S proteasome degradation, TGFb- MYPNWT/Q529X mixture. (C) mRNA levels of Carp, Mlp/Csrp3, and Des in mouse hearts Gapdh SMADs, and caspase 3 and calpain 3 proteases, are reported expressed in fold changes relative to levels of in WT mice. Results were obtained in triplicate from 4 mice per group. Csrp ¼ cysteine and glycine-rich protein 3; other to be CARP regulators (22). Ablation of all forms of muscle abbreviations as in Figures 1, 3, and 4. ankyrin repeat proteins (MARPs), including CARP, en- hances myogenic differentiation protein and MLP expression in skeletal muscle (23). Although ablation of CARP or MARPs has been shown to not affect cardiac translocated to the nucleus, reducing the expression of function similarly to our MypnQ526X homozygotes (24), Carp. Transcription of a key fibrotic molecule, Tgfb1,was cardiac-restricted CARP overexpression results in an inhi- up-regulated, but Tgfb1 protein levels were not changed, bition of hypertrophic and fibrotic remodeling via reduced and Smad2 was down-regulated, suggesting an alternative levels of TGFb and ERK1/2 (18). Our previous MypnY20C TGFb-SMAD–independent fibrotic pathway. Increased transgenic mouse model (Central Illustration) suggested a expression of 90-kDa palladin was shown to induce crucial role of nuclear WT Mypn by demonstrating the a-smooth muscle actin through Tgfb1 and Erk1/2 activation development of hypertrophic cardiomyopathy in vivo as a (25), whereas decreased Ankrd2 and Mlp/Csrp3 levels were result of perturbed nuclear shuttling of Mypn (9).Inthe reported in forced skeletal muscle inactivity in mice (26). present MypnWT/Q526X model, the mutant MypnQ526X Therefore, decreased Carp and elevated Mlp/Csrp3 and 2774 Huby et al. JACC VOL. 64, NO. 25, 2014 Z-Disk in Restrictive Cardiomyopathy DECEMBER 30, 2014:2765– 76

CENTRAL ILLUSTRATION Model of MYPN Mutation–Induced Abnormal Response to Physiological and Pathological Stimuli

In this model, the middle panel represents normal, balanced MYPN interactions that regulate adequate responses to various stresses. In MYPN-Q529X mutation–induced RCM, the MYPNQ529X peptide (right panel) translocates to the nucleus, decreases CARP expression, and increases MLP and DES expression on the protein level, leading to fibrosis via up-regulated POSTN and OPN. The hypertrophic response is abolished via reduced expression of ERK1/2, SMADs, and AKT. Limited or loss of adaptive responses to various stimuli results in the development of RCM. Conventional cardiovascular drugs, including angiotensin-converting enzyme inhibitors, angiotensin II receptor blockers, and beta-blockers thus have limited effect in patients with RCM. The left panel represents the mechanism of MYPN-Y20C mutation–induced HCM as a result of reduced binding of CARP to the MYPNY20C peptide and impaired nuclear translocation of MYPN. Subsequent alterations in DES, ACTN2, VINC, DSP, and Cx43 were evident. ACTN2 ¼ alpha-actinin; AngII ¼ angiotensin II; a,b-AR ¼ alpha- and beta-adrenergic receptors; CARP ¼ cardiac ankyrin repeat protein; Cx43 ¼ connexin 43; DES ¼ desmin; Dox ¼ doxorubicin; DSP ¼ desmoplakin; ERK1/2 ¼ extracellular signal–regulated kinase 1/2; HCM ¼ hypertrophic cardiomyopathy; MLP ¼ muscle LIM protein; MYPN ¼ myopalladin; NEBL ¼ nebulette; OPN ¼ osteopontin; POSTN ¼ periostin; RCM ¼ restrictive cardiomyopathy; TGF ¼ transforming growth factor; VINC ¼ vinculin.

Des expression likely led to fibrosis via periostin and hypertrophic response in heterozygotes (Figure 5C). osteopontin in our model (17). Moreover, up-regulation of Because Carp has been shown to regulate the Erk2- intercellular adhesion protein, vascular cell adhesion pro- dependent titin-PEVK spring force at the I-band, we tein, and interleukin-1b in mutants may contribute to assumethatreductioninMek1/2–Erk1/2 phosphory- perivascular fibrosis (27). lation might have a negative feedback origin through We hypothesized that the net result of decreased the nuclear MypnQ526X-induced perturbation of Erk1/2, Smads, and Akt signaling may blunt the mechanosensitive proteins, Carp, Mlp, and Des (28). JACC VOL. 64, NO. 25, 2014 Huby et al. 2775 DECEMBER 30, 2014:2765– 76 Z-Disk in Restrictive Cardiomyopathy

STUDY LIMITATIONS. An assessment of diastolic for providing antibodies against myopalladin and dysfunction ideally includes invasive measurements. nebulette. This study adopted an integrated approach using echocardiography and CMR imaging in the same REPRINT REQUESTS AND CORRESPONDENCE: Dr. cohort of mice and provides compatible data compa- Enkhsaikhan Purevjav, Cincinnati Children’sHospital rabletoinvasivemeasurements. Medical Center, MLC7020, 3333 Burnet Avenue, Cincinnati, Ohio 45229. E-mail: enkhsaikhan.purevjav@ CONCLUSIONS cchmc.org.

We demonstrated that the 65-kDa MypnQ526X peptide PERSPECTIVES caused diastolic dysfunction without contractile impairment via net changes in the mechanosensory proteins. To date, no pharmacological therapies have COMPETENCY IN MEDICAL KNOWLEDGE: Restrictive been shown to clearly improve outcomes in patients myocardial remodeling is responsible for the diastolic ventricular withRCM,withhearttransplantbeingthedefinitive dysfunction that characterizes FRCM, which carries a poor treatment option for many, especially in childhood prognosis, particularly in affected children. No pharmacological (29).Ifourmurinefindings mirror the pathological therapies have been found that clearly improve clinical outcomes hallmarks of RCM in humans, it could explain why the in patients with this disorder, leaving few alternatives to cardiac TGFb1 suppressor angiotensin-converting enzyme transplant. inhibitors, beta-blockers, and angiotensin receptor TRANSLATIONAL OUTLOOK: blockers are ineffective in the treatment of patients This study discovered early with RCM (3). Further studies on time-dependent molecular markers of restrictive remodeling in a murine model of expression changes in CARP, MLP, DES, and ERK1/2 FRCM that caused diastolic dysfunction without contractile in patients with RCM may provide useful information impairment. This points the way toward future studies in which for discovering diagnostic and therapeutic targets. the expression of genes regulating mechanosensory proteins could be modified as potential therapeutic targets. ACKNOWLEDGMENT The authors thank Dr. Siegfried Labeit (Klinikum Mannheim, Mannheim, Germany)

REFERENCES

1. Maron BJ, Towbin JA, Thiene G, et al. Con- 7. Ma K, Wang K. Interaction of SH3 pediatric patients with restrictive cardiomyopathy. temporary definitions and classification of the domain with titin PEVK and myopalladin: impli- Circ Heart Fail 2012;5:267–73. cardiomyopathies: an American Heart Association cations for the signaling and assembly role of titin 14. Gautel M, Goulding D, Bullard B, Weber K, Scientific Statement from the Council on Clinical and nebulin. FEBS Lett 2002;532:273–8. Furst DO. The central Z-disk region of titin is Cardiology, Heart Failure and Transplantation 8. Duboscq-Bidot L, Xu P, Charron P, et al. assembled from a novel repeat in variable copy Committee; Quality of Care and Outcomes Re- Mutations in the Z-band protein myopalladin gene numbers. J Cell Sci 1996;109:2747–54. search and Functional Genomics and Translational and idiopathic dilated cardiomyopathy. Cardiovasc Biology Interdisciplinary Working Groups; and 15. Bowles NE, Bowles KR, Towbin JA. The “final Res 2008;77:118–25. Council on Epidemiology and Prevention. Circula- common pathway” hypothesis and inherited tion 2006;113:1807–16. 9. Purevjav E, Arimura T, Augustin S, et al. cardiovascular disease. The role of cytoskeletal Molecular basis for clinical heterogeneity in proteins in dilated cardiomyopathy. Herz 2000; 2. Maskatia SA, Decker JA, Spinner JA, et al. inherited cardiomyopathies due to myopalladin 25:168–75. Restrictive physiology is associated with poor mutations. Hum Mol Genet 2012;21:2039–53. outcomes in children with hypertrophic cardio- 16. Chen YW, Nader GA, Baar KR, Fedele MJ, myopathy. Pediatr Cardiol 2012;33:141–9. 10. Liu P, Jenkins NA, Copeland NG. A Hoffman EP, Esser KA. Response of rat muscle to fi fi 3. Webber SA, Lipshultz SE, Sleeper LA, et al. highly ef cient recombineering-based method acute resistance exercise de ned by transcrip- fi Outcomes of restrictive cardiomyopathy in child- for generating conditional knockout mutations. tional and translational pro ling. J Physiol 2002; – – hood and the influence of phenotype: a report Genome Res 2003;13:476 84. 545:27 41. from the Pediatric Cardiomyopathy Registry. Cir- 11. Maiellaro-Rafferty K, Wansapura JP, 17. Mohamed JS, Boriek AM. Loss of desmin culation 2012;126:1237–44. Mendsaikhan U, et al. Altered regional cardiac triggers mechanosensitivity and up-regulation 4. Sen-Chowdhry S, Syrris P, McKenna WJ. wall mechanics are associated with differential of Ankrd1 expression through Akt-NF-kappaB Genetics of restrictive cardiomyopathy. Heart Fail cardiomyocyte calcium handling due to nebu- signaling pathway in smooth muscle cells. FASEB – Clin 2010;6:179–86. lette mutations in preclinical inherited dilated J 2012;26:757 65. cardiomyopathy. J Mol Cell Cardiol 2013;60C: 5. Bang ML, Mudry RE, McElhinny AS, et al. 18. Song Y, Xu J, Li Y, et al. Cardiac ankyrin repeat 151–60. Myopalladin, a novel 145-kilodalton sarcomeric protein attenuates cardiac hypertrophy by inhibi- protein with multiple roles in Z-disc and I-band 12. Jean-Charles PY, Li YJ, Nan CL, Huang XP. tion of ERK1/2 and TGF-beta signaling pathways. protein assemblies. J Cell Biol 2001;153:413–27. Insights into restrictive cardiomyopathy from PLoS One 2012;7:e50436. clinical and animal studies. J Geriatr Cardiol 2011; 6. Mikhailov AT, Torrado M. The enigmatic role of 19. Lee MJ, Kwak YK, You KR, Lee BH, Kim DG. 8:168–83. the ankyrin repeat domain 1 gene in heart devel- Involvement of GADD153 and cardiac ankyrin opment and disease. Int J Dev Biol 2008;52: 13. Walsh MA, Grenier MA, Jefferies JL, Towbin JA, repeat protein in cardiac ischemia-reperfusion 811–21. Lorts A, Czosek RJ. Conduction abnormalities in injury. Mol Med 2009;41:243–52. 2776 Huby et al. JACC VOL. 64, NO. 25, 2014 Z-Disk in Restrictive Cardiomyopathy DECEMBER 30, 2014:2765– 76

20. Zolk O, Marx M, Jackel E, El-Armouche A, muscle. Am J Physiol Cell Physiol 2007;293: molecules in hearts with congestive and ischemic Eschenhagen T. Beta-adrenergic stimulation in- C218–27. cardiomyopathy: immunohistochemical evaluation duces cardiac ankyrin repeat protein expression: of inflammatory endothelial cell activation. Eur J 24. Bang ML, Gu Y, Dalton ND, Peterson KL, involvement of protein kinase A and calmodulin- Cardiothorac Surg 2005;27:122–7. Chien KR, Chen J. The muscle ankyrin repeat dependent kinase. Cardiovasc Res 2003;59: proteins CARP, Ankrd2, and DARP are not essential 28. Witt SH, Labeit D, Granzier H, Labeit S, 563–72. for normal cardiac development and function at Witt CC. Dimerization of the cardiac ankyrin pro- 21. Chen B, Zhong L, Roush SF, et al. Disruption of basal conditions and in response to pressure tein CARP: implications for MARP titin-based a GATA4/Ankrd1 signaling axis in cardiomyocytes overload. PLoS One 2014;9:e93638. signaling. J Muscle Res Cell Motil 2005;26:401–8. leads to sarcomere disarray: implications for 25. Ronty MJ, Leivonen SK, Hinz B, et al. Isoform- 29. Denfield SW, Webber SA. Restrictive cardio- anthracycline cardiomyopathy. PLoS One 2012;7: specific regulation of the actin-organizing protein myopathy in childhood. Heart Fail Clin 2010;6: e35743. palladin during TGF-beta1-induced myofibroblast 445–52. 22. Kanai H, Tanaka T, Aihara Y, et al. Trans- differentiation. J Invest Dermatol 2006;126: – forming growth factor-beta/Smads signaling 2387 96. induces transcription of the cell type-restricted 26. Roberts MD, Childs TE, Brown JD, Davis JW, KEY WORDS CARP/ANKRD1, ERK1/2, ankyrin repeat protein CARP gene through CAGA Booth FW. Early depression of Ankrd2 and Csrp3 fibrosis, remodeling motif in vascular smooth muscle cells. Circ Res mRNAs in the polyribosomal and whole tissue 2001;88:30–6. fractions in skeletal muscle with decreased voluntary running. J Appl Physiol 2012;112:1291–9. 23. Barash IA, Bang ML, Mathew L, Greaser ML, APPENDIX For a supplemental Methods Chen J, Lieber RL. Structural and regulatory roles 27. Wilhelmi MH, Leyh RG, Wilhelmi M, section as well as figures and tables, please of muscle ankyrin repeat protein family in skeletal Haverich A. Upregulation of endothelial adhesion see the online version of this article.