Research Collection
Doctoral Thesis
Integration of myofibrils in the developing heart and challenges on the intercalated disc stability
Author(s): Hirschy, Alain
Publication Date: 2004
Permanent Link: https://doi.org/10.3929/ethz-a-005001295
Rights / License: In Copyright - Non-Commercial Use Permitted
This page was generated automatically upon download from the ETH Zurich Research Collection. For more information please consult the Terms of use.
ETH Library Dissertation ETH N°15790
Integration of myofibrils in the developing heart and
challenges on the intercalated disc stability
A dissertation submitted to the
SWISS INSTITUTE OF TECHNOLOGY ZURICH
(ETHZ)
For the degree of
Doctor of Natural Sciences
presented by
Alain Hirschy
Biologiste diplômé (Université de Neuchâtel, Switzerland) Born April 15, 1975
Citizen of Neuchâtel
Accepted on the recommendation of
Prof. Dr Jean-Claude Perriard, examiner
Prof. Dr Lukas Sommer, co-examiner
Prof. Dr Thierry Pedrazzini, co-examiner
November 2004 Table of contents
Table of contents I Abbreviations IV Abstract 1 Résumé 3 1 Introduction 5 1.1 The heart 5 1.1.1 Morphological development of the heart 5 1.1.2 Molecular pathways controlling heart development 6 1.1.3 Development of ventricular cardiomyocytes 6 1.2 Myofibrillogenesis and development of cell-cell contacts in the ventricular myocardium 7 1.2.1 The contractile apparatus 7 1.2.2 Assembly of sarcomeric proteins 7 1.2.3 Development of cell-cell contacts 9 1.2.4 Three types of cell-cell contact form the intercalated disc 10 1.2.5 The adherens junction 11 1.2.6 The desmosome 15 17 1.2.7 The gap junction 1.2.8 The costameres: cell to extracellular matrix contacts in contractile cells 19 1.3 Role of ß-catenin in signalling 21 1.3.1 Wnt signalling 21 1.3.2 Other regulators of ß-catenin 24 1.4 The conditional knockout approach 24 1.4.1 Cre/lox technology 24 1.4.2 MLC2v-Cre knock-in mouse 26
1.4.3 ß-catenin floxed mouse 28 1.5 Cardiomyopathies 29 1.5.1 Presentation of cardiomyopathies 29 1.5.2 Mutations in contractile and structural proteins lead to cardiomyopathies 29 1.5.3 MLP KO model 31 1.5.4 DRAL KO model 32 1.6 A im of the study 32 1.6.1 ICD development and integration of myofibrils 32 1.6.2 Reconstruction of ICD in vitro 33 1.6.3 Déstabilisation of the ICD 33
1.6.4 Additional stress on heart cells 34 2 Material and Methods 35 2.1 Cloning methods 35 2.1.1 Plasmid manipulation 35 2.1.2 Transformation of competent cells and bacterial culture 36 2.1.3 Plasmid DNA isolation 37 2.1.4 Sequencing PCR 37 2.2 RNA quantification 38 2.2.1 Total RNA isolation 38 2.2.2 RT-PCR analysis 38 2.3 Immunoblotting 39 2.3.1 SDS-sample preparation 39 2.3.2 Electrophoresis and transfer 39 2.3.3 Antibodies for immunoblot 40 2.3.4 Blotting and immunodetection 41 2.3.5 Densitometrie analysis 41 2.4 Isolation of mouse heart cells and cryosections 41 2.4.1 Isolation of rodent heart 41
I 2.4.2 Dissociation and culture of neonatal rat cardiomyocytes (NRCs) 42 2.4.3 Dissociation and culture of neonatal mouse cardiomyocytes 42 2.4.4 Isolation of adult heart cells 42 2.4.5 Cryosections 43 2.5 Transfection of neonatal rat cardiomyocytes 43 2.6 Fixation, immunofluorescence staining and apoptosis detection 43 2.6.1 Antibodies used in immunofluorescence 43 2.6.2 Immunofluorescence of heart whole mount preparations 45 2.6.3 Immunofluorescence of isolated cells 45 2.6.4 Immunofluorescence of cryosections 45 2.6.5 In Situ apoptosis detection 45 2.7 Microscopy 46 2.7.1 Standard fluorescence microscopy 46 2.7.2 Confocal microscopy 46 2.7.3 Cell measurements and volume reconstruction 46 2.8 Generation of conditional knockout mice and double knockouts by breeding 47 2.8.1 Animal strains, genetic background and maintenance 47 2.8.2 Genotyping 47 2.9 Hypertrophy induction and echocardiography 48 2.9.1 Hypertension-induced hypertrophy (1K1C model) 49 2.9.2 ß-adrenergic stimulation of the heart 49 2.9.3 Echocardiography of the mouse heart 49 2.10 Magnetic resonance imaging (MRI) 49 2.11 Statistical analysis 50 3 Results 51 3.1 Development of the intercalated disc in the heart of mouse embryos 51 3.1.1 Overview of the results 51 3.1.2 Myofibrillar and morphological changes in cardiomyocytes from embryonic stage to adult 51 3.1.3 Appearance of cardiomyocytes during heart development 52 3.1.4 Expression of cell-cell contact and extracellular matrix components in the heart 53 3.1.5 Growth of myofibrils 53 3.1.6 Orientation of myofibrils 54 3.1.7 Change in cell shape during cardiomyocyte development 54 3.1.8 Distribution of the adherens junctions during development 55 3.1.9 Distribution of desmosomes 56 3.1.10 Gap junctions 56 3.1.11 Distribution of the extracellular matrix (ECM) 57 3.2 Labelling of the ICD and myofibrils in vitro 58 3.2.1 Outline of the project 58 3.2.2 Tagging of cDNAs and expression vectors 59 3.2.3 Localisation of transfected catenin proteins 59 3.2.4 Localisation of transfected gap junction proteins 60 3.2.5 Localisation of transfected focal adhesion proteins 60 3.2.6 Expression of red fluorescent constructs 61 3.2.7 Localisation of transfected bicistronic constructs 61 3.3 Analysis of the ß-catenin conditional knockout 62 3.3.1 Outline of the project 62 3.3.2 Generation of a heart spécifie deletion of ß-catenin 62 3.3.3 Specificity and efficiency of the Cre mediated recombination 63 3.3.4 Deletion of ß-catenin through postnatal development 63 3.3.5 Deletion of ß-catenin at neonatal stage inculture 64 3.3.6 Regulation of other intercalated disc proteins in the absence of ß-catenin 65 3.3.7 Possible reasons for the long survival of ß-catenin in ICD 67 3.3.8 Is there a hypertrophic response in conditional ß-catenin KO hearts? 68
II 3.3.9 There is no significant increase of myocyte death in conditional ß-catenin KO hearts 69 3.3.10 Physiological parameters are not significantly altered in basal conditions 69 3.3.11 ß-catenin deletion improves fractional shortening in ß-adrenergic-induced hypertrophy 70 3.4 Importance of ß-catenin in DCM heart 71 3.4.1 ß-catenin knockout in MLP knockout mice 71 3.4.2.Early postnatal lethality associated with double KO mice 71 3.4.3 Hypertrophic response at RNA level 72 3.4.4 ANF and ot/ß-catenin protein expression 73 3.4.5 Sarcomeric organisation 73 3.5 Importance of ß-catenin in DRAL KO heart 74 3.5.1 ß-catenin knockout in DRAL knock out mice 74 3.5.2 Early postnatal lethality associated with double KO mice 74 3.5.3 Intraventricular septum defect in the double KO mice 74 4.1 Development of the intercalated disc: a long process of maturation 76 4.1.1 Critical analysis of the method: 76 4.1.2 How do myofibrils grow ? 76 4.1.3 What is the driving force for the alignment of myofibrils? 77 4.1.4 Embryonic developmental hypertrophy ? 77 4.1.5 Changes in nuclear morphology 78 4.1.6 What does the polarisation of AJ and DJ, compared with GJ mean? 78 4.1.7 The sorting out of cell-cell contacts and ECM contacts in postnatal heart 79 4.2 ICD-myofibril reconstruction in vitro 79 4.2.1 Criticisms of the method 79
4.2.2 a-catenin as a specific marker of cell-cell contact 80 4.2.3 RFP, a promising complement to GFP in dual labelling 80 4.2.4 Labelling of the ICD-myofibril interface in 3D and during myofibrillogenesis 81 4.3 Importance of ß-catenin under physiological and under stress conditions 81 4.3.1 Reasons for the slow disappearance of ß-catenin 81 4.3.2 Adaptation of adherens, desmosomal and gap junction to the ß-catenin deletion 82 4.3.3 No cardiac hypertrophy but cellular hypertrophy. Is this possible? 82 4.3.4 ß-catenin as a regulator of hypertrophy? 83 4.3.5 The sarcomeric structure is preserved 83 4.3.6 Deletion of ß-catenin in hypertrophic conditions: consequences on contractility 84 4.3.7 Early lethality of female cKO, dream or reality? 84 4.3.8 Regulation of ß-catenin in the healthy and hypertrophic heart 85 4.4 Deletion of ß-catenin on the top of LIM protein deletion 85 4.4.1 Critics of the mix in genetic backgound 85 4.4.2 Are there possible explanations for the early lethality of MLP-ß-catenin KOs? 86 4.4.3 Is there a possible explanation for the early lethality of ß-catenin cKO DRAL KO? 86 4.4.4 Ventricular septum defects as a cause of postnatal lethality? 87 4.4.5 Postnatal lethality in relationship with the initiation of developmental hypertrophy? 88 5 References 89 6 Appendix 114 6.1 Database of plasmids expressing Intercalated disc proteins and ES cells selection system 114 Acknowledgements 120 Curriculum Vitae 121
III Abbreviations
ANF: Atrial natriuretic peptide
APC: Adenomatous polyposis coli (tumour supressor protein)
APS: Ammonium persulfate
AR: Androgen receptor
ARC: Adult rat cardiomyocyte
aSK: Alpha skeletal actin
BNP: Brain natriuretic peptide
BSA: Bovine Serum Albumin
ßMHC: Beta-myosin heavy chain
cAMP: Cyclic Adenosine monophosphate
cKO: Conditional knockout
COS: African Green Monkey Kidney (epithelial cells)
Cy2: Cyanine
Cy3: Indocarbocyanine
Cy5: Indodicarbocyanine
DAPI: 4,5 diamino-2- phenylindoldihydrochlorid
DCM: Dilated cardiomyopathy
DNA: Deoxyribonucleic Acid
dNTP's: Deoxynucleosid triphosphate
DRAL: Down regulated in Rhabdomyosarcoma LIM protein
DTT: Dithiothreitol
Ex.x: Embryonic day x.x
E.coli: Escherichia coli
EDTA: Ethylene diamine tetraacetic acid
EGF: Epidermal growth factor
EGTA: Ethylene glycol-bis[ß-aminoethyl ether] N, N, N', N' tetraacetic acid
ES cell: embryonic stem cell
EtBr: Ethidium bromide
FCS: Fetal calf serum
FITC: Fluorescein isothiocyanate
GAPDH: Glyceraldehyde 3-phosphate dehydrogenase
GFP: Green Fluorescent Protein
GSK3ß: Glycogen synthase kinase 3ß
HCM: Hypertrophic cardiomyopathy
HEPES: 2-[4-(2-hydroxyethyl)-l-piperazinyl]-ethansulfone acid
IV HRPO: Horseradish peroxidase
ICD: Intercalated disc
ISO: Isoproterenol
JNK: Janus N-terminal kinase
KO: knockout
LB: Luria-Betani medium
LEF: leukocyte enhancer factor
Luminol: 5-Amino-2,3,-dihydro 1,4 phtalazine-dione, sodium salt
MAPK: Mitogen-activated protein kinase
MDCK: Madin Darby Canine Kidney (cells)
MLC: myosin light chain
MLP: Muscle LIM protein
MRI: magnetic resonance imaging
NGS: Normal Goat Serum
NRC: Neonatal rat cardiomyocyte
P.x.x: postnatal day x.x
PAGE: Polyacrylamid gel electrophoresis
PBS: Phosphate buffered saline
PCR: Polymerase chain reaction
PFA: Paraformaldehyde
PIPES: Piperazine-N,N'-bis[2-ethanesulfonic acid]
PKB: Protein kinase B
PKC: Protein kinase C
PLN: Phospholamban
RNA: Ribonucleic acid
RT: Room temperature
RT-PCR: reverse transcriptase- polymerase chain reaction
SDS: Sodium dodecylsulfate
TCF: T-cell factor
TE buffer: Tris-EDTA buffer
TEMED: N,N,N', N'-tetramethylethylendiamine
UV: Ultra-Violet
Wnt signalling: Wingless-Int signalling (pathway)
WT: Wild-type
X-gal: 5-bromo-4-chloro-3-inolyl-ß-D-galactopyranoside
1K1C: one-kidney one-clip (model)
V Abstract
The intercalated disc (ICD) is a structure composed of three types of junctions: adherens junctions, desmosomes and gap junctions. It is essential for the maintenance of mechanical
and electrical coupling between cardiomyocytes and for the proper integration of myofibrils
into the membrane. So far, little is known about the biogenesis of this organelle and most of
the studies have been made either at postnatal stages (Angst et al., 1997) or with isolated cells
in culture (Kostin et al., 1999).
In the first part of this study, we show for the first time the time-course of intercalated disc
development and pinpoint essential steps in the maturation of this whole structure. In the early
embryonic heart, cardiomyocytes are round-shaped cells with adherens junctions and
desmosomes along their entire plasma membrane. As they elongate, creating two distinct
lateral and distal cell borders, myofibrils get aligned. Adherens junctions and desmosomes
remain at the distal membrane but disappear from the lateral borders and get replaced by cell-
to-matrix contacts. Connexin 43-based gap junctions appear rather late in ventricular
development and are restricted to the intercalated disc only after birth. All these events are
accompanied by an increase of the cell volume, which probably means that the so-called
developmental hypertrophy process, contrary to general beliefs, starts already during the embryonic phase.
Until recently, it was almost impossible to get definitive information about the importance of
individual components of the ICD in mammalian heart because germ line deletions of these
components were often lethal before heart formation had taken place. The Cre/lox technology
enables now to challenge the heart with a tissue-specific deletion to get insight into the
intrinsic resistance of the ICD to these genetic manipulations. After the construction of
plasmids for in vitro labelling of the myofibril-ICD interface, we started with a project to
achieve a heart specific deletion of ß-catenin, a protein which is part of the adherens
junctions. By creating an MLC2v Cre mouse containing the floxed ß-catenin gene, we could
detect the complete removal of this gene from the ICD only two months after birth. The
removal of ß-catenin is well tolerated in adult mice because these animals do not show any
sign of heart dysfunction or cardiac hypertrophy. Nevertheless, we could observe a
compensatory upregulation of other adherens junction proteins in response to ß-catenin
deletion. Quite surprisingly, individual cardiomyocytes, which undergo recombination, have
in average an increased cellular volume. As neither hypertrophic response nor apoptosis could
1 be detected in these animals, we hypothesised that this increase of cellular volume is a compensatory mechanism in response to a decrease of cell proliferation. This hypothesis would then imply that the deletion of ß-catenin is first affecting its signalling and proliferative activity and, as consequence, decreases embryonic cardiomyocyte proliferation.
Challenging the heart further with deletion of either MLP or DRAL (two LIM domain proteins expressed specifically in the heart) in combination with ß-catenin removal has dramatic consequences and about 70% of the double knockout mice die in the first week after birth. Considering that MLP and DRAL knockout mice do not show any lethality at this stage and knowing that ß-catenin is not removed yet from the ICD in the first postnatal week, we conclude that the deletion of the signalling activity of ß-catenin is once again the essential missing factor, which is responsible for this premature heart failure. The question concerning the exact causes of these early heart failure remains to be solved. We were able to demonstrate that MLP-ß-catenin double KO mice have an enlarged heart and upregulate the expression of the hypertrophy-associated marker ANF. In contrary, DRAL-ß-catenin double
KO hearts are apparently normal but detailed examination using magnetic resonance imaging demonstrated the occurrence of a ventricular septum defect. This cardiac malformation results from the incomplete closure of the septum between left and right ventricle during the embryonic heart development and allows the mixing of oxygenated and not oxygenated blood, which is detrimental for heart function. These two phenotypes will be characterised in more detail in future experiments.
2 Résumé
Le disque intercalaire est une structure composée de trois éléments: les jonctions adhérentes, les desmosomes et les jonctions communicantes. Cette structure est essentielle pour la maintenance du couplage mécanique et électrique entre les cardiomyocytes ainsi que pour l'intégration des myofibrilles dans la membrane cellulaire. En l'état actuel, très peu de choses sont connues sur la biogenèse de cet organeile et la plupart des études ont été faites à partir du stade postnatal (Angst et al., 1997) ou avec des cellules isolées mises en culture (Kostin et al.,
1999).
Dans la première partie de cette étude, nous démontrons le développement au cours du temps du disque intercalaire et mettons en évidence les points essentiels qui caractérisent la maturation de l'ensemble de cette structure. Initialement, dans le cœur embryonnaire, les cardiomyocytes sont des cellules approximativement rondes entourées sur l'ensemble de la membrane cytoplasmique par des jonctions adhérentes et par des desmosomes. L'élongation de ces cellules définit une membrane latérale et une membrane distale tandis que les myofibrilles s'alignent à l'intérieur de la cellule. Peu à peu, les jonctions adhérentes et les desmosomes disparaissent de la membrane latérale et sont remplacés par des contacts avec la matrice extracellulaire. Les jonctions communicantes formées à partir de connexin 43 apparaissent relativement tard dans le développement ventriculaire et ne sont restreintes au disque intercalaire qu'après la naissance. Tous ces événements sont accompagnés par une augmentation du volume cellulaire ce qui signifie que le processus appelé hypertrophie développementale commence en réalité déjà dans la phase embryonnaire.
Jusqu'à présent, il était pratiquement impossible d'obtenir des informations concrètes sur l'importance de chaque élément du disque intercalaire dans le cœur parce que les deletions de ces composants, introduites dans la lignée germinale, étaient très souvent létales avant la formation du tissu cardiaque. La technologie Cre/Lox nous permet maintenant de confronter le cœur avec des deletions restreintes au tissu cardiaque et d'approfondir ainsi notre connaissance sur la résistance intrinsèque du disque intercalaire face à ces deletions. Après la construction de plasmides pour la visualisation de l'interface myofibrille-disque intercalaire in vitro, nous avons commencé un projet qui consiste à supprimer dans le cœur l'expression de la protéine ß-catenin qui est une composante des jonctions adhérentes. En utilisant une lignée de souris exprimant la Cre recombinase sous le contrôle du promoteur MLC2v et des alleles de ß-catenin floxés, il nous a été possible de détecter une suppression de ß-catenin dans le disque intercalaire seulement à partir du deuxième mois après la naissance. La suppression est
3 bien tolérée chez les souris adultes et ces animaux ne montrent aucun signe de dysfonctionnement cardiaque ou d'hypertrophie. Par contre, nous pouvons observer, par phénomène de compensation, une surexpression d'autres protéines des jonctions adhérentes en l'absence de ß-catenin. De plus, les cardiomyocytes qui ont subi la deletion ont en moyenne un volume cellulaire plus important.
La combinaison de la deletion de ß-catenin avec la suppression soit de MLP, soit de DRAL
(deux protéines exprimées spécifiquement dans le cœur adult et qui contiennent des domaines
LIM) a des conséquences dramatiques car environ 70% des souris double knockout meurent dans la première semaine après la naissance. En considérant que la suppression de MLP ou de
DRAL ne présente pas de létalité à ce stade de développement et en sachant que, au même moment, ß-catenin n'a pas encore disparu du disque intercalaire, nous pouvons conclure que c'est la deletion de l'activité de signal de cette protéine qui est responsable de l'insuffisance cardiaque. Les questions relatives aux causes exactes de l'insuffisance cardiaque restent à déterminer. Nous avons été capables de démontrer que les souris MLP-ß-catenin double knockouts ont un cœur hypertrophié et surexpriment ANF, un marqueur fréquemment associé
à l'hypertrophie. Au contraire, les souris DRAL-ß-catenin doubles knockouts ont un cœur en apparence normal mais un examen plus approfondi, utilisant la résonance magnétique nucléaire démontre la présence d'une ouverture au niveau du septum ventriculaire. Cette malformation cardiaque est causée par la fermeture incomplète du septum entre les deux ventricules pendant le développement du cœur, laissant un passage permettant le mélange du sang oxygéné et non-oxygéné, ce qui est préjudiciable pour la fonction du cœur. Les phénotypes de ces deux doubles knockouts seront analysés avec plus de détail dans le futur.
4 1 Introduction
1.1 The heart
1.1.1 Morphological development of the heart
The heart is the first functional organ in the body (Olson and Srivastava, 1996). In all vertebrates, the heart tube, which consists of an endocardial and a myocardial layer separated by an extracellular matrix known as the cardiac jelly, develops from bilateral populations of precursor cells (Figure 1.1.A) in the anterior lateral plate mesoderm (Fishman and Chicn,
1997). In the mouse, the generation of this tubular heart (Figure l.l.B) starts at embryonic day
8 (E8.0) and the first rhythmic contractions are observed at K8.5. Then the heart tube initiates a looping process (Figurel.l.B) that brings the future atrial region from a posterior to an anterior position. The separation of the common atrioventricular canal into the atria, ventricles and outflow tract is accomplished by the outgrowth of mesenchymal cells, which form structures known as cardiac cushions (Markwald et al., 1996). The expansion of these cushions gives rise to the primordium of septal and valvular structures (Eisenberg and
Markwald, 1995). Together with the ingrowths of the interatrial and interventricular septa the atrio-ventricular cushions delimit the future left and right chambers (Figure 1.1 .D).
Figure 1.1: Schematic diagram of cardiogenesis. Bilaterally symmetrical cardiac progenitor cells (A) are prepattcrncd to form distinct regions of the heart. The precardiac mesodermal cells give rise to a linear tube (B) which forms a rightward loop (C) and begins, to establish the spatial orientation of the four-chambered heart (D)
(Srivastava and Olson, 1997)
Until birth, the maturation of the heart is characterised by the development of trabeculae in the ventricle, the differentiation of a group of ventricular cardiomyocytes into tracts of cells that form the cardiac conduction system and the extensive remodelling of the outflow tract
(with the essential contribution of cells derived from the neural crest as shown by Kirby and
Waldo, 1995) into aorta and pulmonary arteries. Finally, at birth, the closure of the foramen
5 ovale between the left and the right atrium achieves the complete separation between pulmonary and systemic circulation.
1.1.2 Molecular pathways controlling heart development
Four classes of growth factors have been extensively characterised for their inductive role in cardiac lineage differentiation: Wnts (cf. section 1.3.1), transforming growth factors ß
(TGFß) such as activin/Nodal, bone morphogenetic proteins (BMPs) and fibroblast growth factors (FGFs) (Foley and Mercola, 2004). Apart from these inductive signals, several molecule families are also important for the correct formation of the heart and characterise the developmental processes. Among them, the homeobox transcription factor Csx/Nkx2.5
(homologous of Drosophila gene Tinman) is the first known marker of cardiac lineage
(Harvey, 1996). In conjugation with many other factors, e.g the MEF-2 family (myocyte enhancer binding factor-2), the zinc finger protein GATA-4 and the basic helix-loop-helix
(bHLH) transcription factors D-hand and E-hand it determines the commitment of cardiac progenitors into heart cells (for review see Srivastava and Olson, 2000; Fishman and Olson,
1997; Olson and Srivastava, 1996; Mohun and Sparrow, 1997). The different cell populations and chamber formation result from the interaction of these cardiac transcription factors together with the inductive signals mentioned above and many other molecules. Among the best-studied examples are the development of valves, which require the transformation of endocardial cells into mesenchymal cells mediated by Wnt activation (Hurlstone et al., 2003), the differentiation of Purkinje cells from ventricular cardiomyocytes through Endothelin-1
(Hyer et al., 1999), the growth of ventricular trabeculae due to Neuregulin expression from the endothelium (Lee et al., 1995; Gassmann et al., 1995; Hertig et al., 1999) and the septation processes, which are at least partially controlled by the transcription factor Tbx5 (Basson et al., 1997; Li et al., 1997)
1.1.3 Development of ventricular cardiomyocytes
The mechanism determining the chamber identity of cardiac cells is unknown but is likely to involve as well a combination of cardiac and more widely expressed transcription factors.
Although the population of cells giving rise to atrial and ventricle cardiomyocytes are spatially segmented along the anterior-posterior axis of the tubular heart, most of atrial and ventricular specific genes are expressed homogenously early in development and get restricted later. For ventricle specific differentiation, the first marker known is MLC2v. This isoform of the regulatory myosin light chain is under the control of Nkx2.5 (Lyons et al.,
1995; Tanaka et al., 1999) and is expressed from E8.0 (O'Brien et al., 1993; Kubalak et al.,
6 1994). The targeted deletion of MLC2v is lethal at E12.5 (Chen et al., 1998b) indicating that the loss of MLC2v cannot be compensated by the atrial isoform (MLC2a) expressed in both chambers at this stage. It is as well worth noting that the transcription factors E-hand/D-hand, which are predominantly expressed in the ventricles, display a right-left asymmetry.
Consistant with this pattern of expression, the inactivation of D-hand results in the suppression of the right ventricle development (Srivastava et al., 1997) whereas E-hand deletion affects left ventricle development (Firulli et al., 1998; Riley et al., 1998). This suggests that hand proteins may actually provide transcriptional control of genes involved in chamber-specific myogenesis.
1.2 Myofibrillogenesis and development of cell-cell contacts in the ventricular myocardium
1.2.1 The contractile apparatus
Actin is one of the most abundant protein in the organism (1-5% of protein weight in non muscle and more than 10% in muscle cells) and has been highly conserved throughout evolution (Mounier and Sparrow, 1997). In its polymerised state, actin forms the cytoskeletal element known as the microfilament system. In cardiomyocytes, actin filaments are incorporated into a larger structure known as the sarcomere, which is in turn arranged in series to form a functional contractile apparatus called myofibril, delimited and anchored at both ends by a complex of protein known as the adherens junction (see section 1.2.5).
Mechanically, the sarcomere is divided in three filament systems: the actin thin filaments, the myosin thick filaments and the titin elastic filaments. The energy consuming sliding of myosin heads along the thin filament produces a shortening of the whole sarcomeric structure
(Figure 1.2.A) whereas titin filaments are supposed to play the role of a molecular spring
(Granzier and Labeit, 2002). The cell contraction, as observed in isolated cells, is the final result of the coordinated shortening of the sarcomeres along the myofibrils. From the regular striated pattern found in heart and skeletal muscle, as observed in electron microscopy
(Figure 1.2.B), the sarcomere is divided in Z-disc, M-band, A- and I-band where Z stands for zwischen (german for between), M for middle, A for anisotropic, and I for isotropic due to the respective properties of these two bands in polarisation microscopy.
1.2.2 Assembly of sarcomeric proteins
Cardiac myofibrillogenesis is the process of expression and integration of sarcomeric proteins, which results in the formation of regular structures known as myofibrils in the
7 cytoplasm of cardiomyocytes. From different reports, common elements can be highlighted concerning the assembly of the contractile apparatus. First of all, myofibrillogenesis in the heart is an extremely fast process: the delay between protein expression and assembly of sarcomeres does not exceed a couple of hours (Imanaka-Yoshida, 1997; Ehler et al., 1999).
Secondly, the formation of sarcomeres occurs close to the cell membrane (Shiraishi et al.,
1993; Imanaka-Yoshida, 1997; Dabiri et al., 1997). Some degree of divergence remains concerning the exact mechanism of myofibril integration. Published reports using cardiomyocytes in culture showed in the spreading region of the cytoplasm the formation of pre-myofibrils associated with stress fibres-like structures, which then mature into functional myofibrils (LoRusso et al., 1997; Dabiri et al., 1997). In contrast, studies with chicken expiants (Imanaka-Yoshida, 1997) and whole mount preparations using immunohistochemistry and confocal microscopy showed the absence of pre-myofibril
structures, and a sequential assembly of sarcomeric subunits. In this process the giant protein titin acts very likely as a ruler and directs the building of the contractile units (van der Ven et al., 2000). This process is sequential in the sense that Z-disc components (ot-actinin, Z-disc
eiptopes of titin, blue bars in Figure 1.2.A) are localised before components of the M-band
(myomesin, M-protein, M-band epitopes of titin) then followed by thick and thin filament
assembly (green respectively red in Figure 1.2.A, (Ehler et al., 1999; Ehler et al., 2004).
8 mtmmtam MMI^mÊmm^^^^ffffm^ «*MAAAAAAi mMP1 «M«l<
Sarcomere C?
I A-band l-bai
Z-dïsc M-band Z-disc
Figure 1.2: Sarcomeric organisation in schematic drawing (A) and in electron microscopy (B) Note the role of three distinct filament systems for contraction in A): myosin (green), actin (red) and titin (yellow). The regularity of the filament alignment is also useful to distinct substructures in B): Z-disc, A-, I-and M-bands (adapted from
Alberts et al., 1994)
1.2.3 Development of cell-cell contacts
Confocal microscopy reveals the presence of adherens junction proteins as a continuous band around the plasma membrane of embryonic chicken cardiomyocytes (Shiraishi et al., 1993;
Imanaka-Yoshida et al., 1998), which clearly contrasts with the restriction of adherens junctions to intercellular contacts found in cultured embryonic cardiomyocytes in vitro
(Goncharova et al., 1992). From this early development of cardiomyocytes to the neonatal stage, the amount of information is limited. Apart from a study focusing on the relationship
9 between adherens junctions and costameres (Wu et al., 2002) and reports concerning connexin isoform changes (Fromaget et al., 1992; Delorme et al., 1995; Delorme et al, 1997), there is no comparative analysis, which includes both myofibrillar and cellular changes and development of cell-cell contacts in embryonic stages, comparable with what is known for postnatal cardiomyocytes (Angst et al., 1997; Severs, 2000).
1.2 4 Three types of cell-cell contact form the intercalated disc
In adult cardiac muscle, cell-cell contact is maintained by a unique junctional complex termed the intercalated disc (ICD), which consists of three separate junctions; the adherens junction, the desmosome and the gap junction (Figure 1.3). For reason of clarity, the junctions are presented in separated paragraphs below. It is worth to notice that none of the components of the adult intercalated disc is specific to cardiomyocytes and that the three types of junction can be found in many other tissues.
Figure 1 3 The TCD is composed of three types of |unctions visible in electron microscopy Adherens junctions
{Fascia adherens, TA) and desmosomes (D) are easily identified in electon microscopy as electron dense material along the cellular membrane Gap junctions {Nexus, N) bring plasma membranes in close apposition
(Wheateretal,1987)
The specificity of the intercalated disc is due to the spatial arrangement of the complex at the polar ends of the cells and the dual function it fulfils. Adherens junctions and desmosomes maintain the mechanical coupling through the anchorage of the thin and intermediate filaments to the membrane whereas gap junctions insure the electrical and inonic coupling and therefore the propagation of the action potential to the next cell. It is as well important to note that all the germ-line deletions of proteins, which are part of the ICD are embryonically lethal
10 (Table 1.1). When the heart can form, defects are often visible as listed in the second row of
Table 1.1. This fact underlines the importance of each individual component during heart development. It is also worth to note here that the requirement of any particular ICD component during heart development does not necessarly correlates with its importance in the maintenance of the cardiac function when the heart formation is completed.
Table 1.1: Knockout of ICD proteins
KO Description Phenotype of Reference
cardiomyocytes
N-cadherin Lethality at E10, general cell weaker Radiceetal., 1997 aggregation* adhesion defect, the heart fails to
develop normally
aE-catenin Disruption of the Trophoblast n.d Torres etal., 1997
epithelium
ß-catenin Lethality at E7, no mesoderm n.d Haegeletal., 1995
formation
vinculin Lethal at E9.5 with heart and brain normal** Xu etal, 1998
defect
plakoglobin Lethality at El0.5 or El2-16 weaker elastic Bierkamp et al., 1996 properties* heart defect, skin blistering Ruiz et al., 1996
desmoplakin Lethal at E6.5, general disruption of n.d Gallicanoetal.,1998
cell-cell contact
connexin 40 Conduction block n.d Simonetal., 1998
Kirchhoffetal., 1998
connexin 43 Lethal at birth, blockage of the reduction of Reaume etal, 1995 electrical right ventricular outflow tract coupling* connexin 45 defective cardiac vascular n.d Kumai et al, 2000
development Kruger et al, 2000
* Heart expiant from -/- embryos
** ES -/- differentiation
n.d: not determined
1.2.5 The adherens junction
The adherens junction is the anchoring complex of the actin cytoskeleton to the membrane
(Figure 1.4) and, via its transmembrane glycoprotein Cadherins, a mediator of calcium
11 dependent intercellular interactions. Different Cadherins have been characterised and the
restriction to one particular isolorm or group of Cadherins is regulated during development
(Takeichi, 1995). Changes in Cadherin expression and function are correlated with the regulation of different morphologic processes, like establishment of tissue boundaries, metastasis, tissue rearrangement, cell migration and differentiation. In the heart muscle, the main isoform expressed is N-cadherin, (the isoform also expressed in nervous tissue) but other isoforms are present, e.g. T-cadherin (Doyle ct al., 1998; Sacristan et al, 1993) and VE- cadhcrin in the endothelial cells (Telo et al., 1998). As shown in in vitro assays, the linkage of the actin filament to N-cadherin is achieved through a cadherin-associated complex containing a-catenin,. ß-catenin, plakoglobin (previously named y-catenin), vinculin and a- actinin (Figure 1.4). At the cytoplasmic face of the junction, ß-catenin and plakoglobin both contain central repetitive motifs called Armadillo repeats {Armadillo being the Drosophila homologue of ß-catenin), which interact in a mutual exclusive manner with a core region of
30 amino acids within the C-terminus o[ the Cadherin. The link between ß- catenin/plakoglobin and a-catcnin requires N-terminal interactions of both partners (Aberle et al., 1994; Nieset et al., 1997) and finally, a-catenin interacts with the cytoskeleton both directly due to its affinity for actin filaments (Rimm ct al., 1995) and indirectly through actin- binding proteins such as oc-actinin (Nieset et al., 1997) and vinculin (Watabe-Uchida et al.,
1998).
CM CM Plakoglobin
Figure 1.4: Schematic representation of selected components of the adherens junction. The cadherin/catenin complex (cadhenn, a-,ß-catemn, plakoglobin) and actin binding proteins (a-actinin, vinculin) attach the thin filament system to the cellular membrane (CM). Regulation ofthe junctional stability
The regulation of cell-cell adhesion seems to be achieved through phosphorylation of members of the cadherin-associated complex (Figure 1.4). Both ß-catenin and plakoglobin possess a site in the N-terminal domain, which is recognised by the serine/threonine kinase
Glycogen synthase kinase 3ß (GSK-3ß). The phosphorylation by GSK-3ß is the signal which directs ß-catenin and plakoglobin to poly-ubiquitination and subsequent degradation by the proteosomal machinery. As the activity of GSK-3ß is elevated in unstimulated cells, the level of free ß-catenin and plakoglobin is usually quite low (Dihlmann et al., 2003).
ß-catenin and plakoglobin are as well targets of tyrosine phosphorylation (Daniel and
Reynolds, 1997) by several kinases like e.g. Fer tyrosine kinase (Piedra et al., 2003), which causes the dissociation of a-catenin from the complex and thus disrupts adhesion (Tsukatani et al., 1997; Ozawa and Kemler, 1998). The phosphorylation of tyrosine residues by c-Src and the epidermal growth factor receptor (EGFR) causes the disruption of the interaction between
ß-catenin and E-cadherin (Roura et al., 1999) and subsequent loosening of cell-cell contacts.
In general, activation of tyrosine kinases results in a loss of cadherin-mediated cell-cell adhesion and an increased level of cytoplasmic ß-catenin. Other postulated regulators of the adherens junction are the members of the pi 20 catenin family, e.g. pi 20 catenin (Ireton et al.,
2002; Davis et al., 2003), ARVCF (armadillo repeat gene-deleted in Velo-Cardio-Facial syndrome, Kaufmann et al., 2000) and p0071 (Hatzfeld and Nachtsheim, 1996). Unlike ß- catenin and plakoglobin, pi20 catenin and its relatives do not appear to have a structural role in the junctional complex and their effects on Cadherin adhesive activity are less understood.
Adherensjunctions and signalling
In epithelial cells, it has been shown that the formation of Cadherin mediated cell-cell contacts activates several Rho-GTPase proteins, such as Rho, Rac, Cdc42 (Braga, 1999), which in turn regulate the remodelling of the cortical cytoskeleton leading to the stabilisation of the junction and the binding of actin filaments. Blocking the function of Rho-GTPases by the injection of dominant negative isoforms of Rho or Racl or botulinum toxin C3 removes Cadherins from cell-cell contacts in keratinocytes (Kuroda et al., 1999). The links between the formation of cell-cell contacts and signalling activities are numerous but not always well understood:
Cadherins are participating in many signals involving several tyrosine kinases, e.g. EGF, the vascular endothelial growth factor VEGF (Stefanou et al., 2003) and FGF. Cadherins are also
13 known to negatively regulating elements of the Wnt signalling pathway, which is discussed in more detail in a following chapter.
Deletion ofadherensjunction proteins
The essential functions of N-cadherin become obvious in mice homozygous for the N- cadherin null mutation. Homozygous mutant embryos die by day 10 of gestation (Table 1.1).
Although both neurulation and somitogenesis initiate apparently normally, the resulting embryos suffer from severe malformation. The most dramatic cell adhesion defect is however observed in the primitive heart. The myocardial tissue is initially formed, but myocytes subsequently dissociate and the heart tube fails to develop normally (Radice et al.,
1997).However the differentiation into cardiomyocytes is not impaired (Luo and Radice,
2003) and N-cadherin-/- ES cells develop relatively normal (R. Bugorsky, personal communication). Rescue of the N-cadherin deletion by heart specific expression of E- cadherin leads to survival of animals and to dilated cardiomyopathy characterised by connexin 43 downregulation, cyclin Dl over-expression and concomitantly nuclear replication followed by karyokinesis (Luo et al., 2001; Ferreira-Cornwell et al., 2002) suggesting that the remodelling induced by the exchange of Cadherin isoform is deleterious for cardiomyocytes.
The developmental importance of ß-catenin is also underlined by the study of ß-catenin null mutant mice. Lack of ß-catenin results in embryonic lethality around day 7 of gestation due to a primary defect in the embryonic ectoderm cell layer. It has been suggested that the lack of
ß-catenin affects the adhesive properties of E-cadherin in embryonic ectodermal cells, leading to an improper integration of dividing cells into the ectodermal cell layer, ß-catenin deletion is also characterised by a lack of mesoderm formation (Haegel et al., 1995). In more refined experiments using Cre/lox technology, ß-catenin has been specifically deleted in different tissues (see section 1.4.3 for more details).
Germ-line deletion of plakoglobin is lethal at embryonic day 10.5 to 16 depending on the genetic background of the mice (Ruiz et al., 1996; Bierkamp et al., 1996). The primary defects are observed in the heart and skin where the number of desmosomes are greatly reduced, suggesting that plakoglobin is more important for desmosome assembly than for adherens junction. Studies with cardiomyocytes isolated from plakoglobin -/- embryos have demonstrated that plakoglobin seems to be indeed dispensable for the anchoring of the actin filaments but essential for the mechanical stability of the contractile apparatus by maintaining the passive compliance (Isac et al., 1999).
14 Disrupting the functionality of the epidermal isoform of a-catenin (aE-catenin; the isoform also expressed in heart) by genetic deletion is lethal at the implantation stage due to the rupture of the trophoblast epithelium (Torres et al., 1997) probably caused by a loss of E- cadherin-dependent adhesion, as previously reported (Shimoyama et al., 1992). Vinculin null- mutant embryos died around embryonic day 9.5, with heart structures, which were developmentally abnormal or severely retarded. Interestingly, vinculin null ES cells were able to differentiate in vitro into a variety of cell types, including rhythmically beating cardiomyocytes (Xu et al., 1998). a-actinin deletion has not been reported so far in mice but a point mutation of the a-actinin 2 gene disrupts the normal function of the protein leading to dilated cardiomyopathy (Mohapatra et al., 2003).
1.2.6 The desmosome
Desmosomes are specialised junctions characteristic of tissues, where stable intercellular associations are required. They are especially abundant in heart muscle and skin epidermis.
Like adherens junctions, desmosomes are pulling adjacent cells into close contact, but not by
anchoring the actomyosin filaments at the cell membrane, but the intermediate filaments
(Figure 1.5). Desmosomes are composed of an extracellular core domain (ECD) and two
symmetrical dense cytoplasmic plaques, which are lying in parallel to the plasma membrane
(North et al., 1999). This pair of compact, electron-dense plaques is the most prominent
feature of the desmosome and can be easily seen by electron microscopy (Figure 1.3).
two families of The core of the desmosome is composed of desmocollins and desmogleins,
transmembrane glycoproteins belonging to the Cadherin superfamily of cell-cell adhesion
molecules, also named desmosomal Cadherins (Koch and Franke, 1994). Each desmosomal
Cadherin subclass comprises three different isoforms which are the products of distinct genes,
in the heart among them desmocollin 2 and desmoglein 2 are the major isoforms expressed
(Theis et al., 1993; Schäfer et al., 1994; Angst et al., 1995). The desmosomal Cadherins
contain five extracellular structural domains, which are similar to the ones found in classic
Cadherins. The intracellular domain is however different, containing a binding site exclusively
for plakoglobin (Chitaev et al., 1996), which forms the cytoplasmic dense plaque together
with plakophilins and desmoplakins. As already mentioned, plakoglobin localises not only to
desmosomes, but also to adherens junction, whereas desmoplakins, quantitatively the most
abundant desmosomal proteins, and plakophilins are solely found in desmosomes. The
named desmoplakin gene encodes two similar isoforms generated by alternative splicing,
desmoplakin I and II expressed at various ratio in different tissues with the exception of the
15 heart which lacks desmoplakin II (Angst et al., 1990). Depending on the cell type, the whole desmosomal structure binds to different intermediate filaments. In heart and skeletal muscle, the intermediate filaments are desmin polymers (Capetanaki and Milner, 1998).
CM CM
Figure 1.5: Schematic representation of the desmosomal junction. The desmosomal catenins (desmogleins and desmocollins) and the plakoglobin/dcsmoplakin complex anchor the intermediate filaments to the cellular membrane (CM).
Deletion ofdesmosomal proteins
Of the whole catalogue of desmosomal Cadherins, only desmoglcin 3 has been subjected to genetic manipulation in mice leading to an N-terminal truncation (Allen ct al., 1996) or complete deletion (Pulkkinen et al., 2002). In both cases, the phenotype associated with the lack of desmoglein 3 were skin defects due to the reduction of desmosomal adhesion. As mentioned before in section 1.2.5, plakoglobin disruption is embryonically lethal primarily due to skin and heart defects consistant with the crucial role of the protein in desmosomal function (Ruiz ct al., 1996). Null-mutant embryos for desmoplakin do not survive beyond embryonic day 6.5. Analysis of these embryos showed a critical role for desmoplakin not only in anchoring the intermediate filaments to the desmosomal plaque, but also for desmosomal assembly and/or stabilisation (Gallicano et al, 1998).
Deletion ofdesmin
Deletion of the desmosomal associated intermediate filament desmin leads to cardiac hypertrophy which is later transformed into dilated cardiomyopathy with a thinning of the ventricular walls more pronounced in the right ventricle. Interestingly, the hypertrophic cardiomyopathy can be monitored by the specific increase of the transverse sectional area
16 observed in desmin-/- cardiomyocytes before the transition to dilated cardiomyopathy (Milner etal, 1999).
1.2.7 The gap junction
Gap junctions are regions composed of numerous intercellular channels (Figure 1.6). Each half-channel, also termed connexon, is composed of six identical monomers, the connexins.
Each connexon makes contact, across a narrow extracellular gap, with another connexon in the opposing membrane to form the channel. Connexins contain four transmembrane domains with both amino and carboxy termini lying at the cytoplasmic side of the membrane (Beyer et al., 1987; Zimmer et al., 1987). Examinations of isolated junctions from liver tissue by high- resolution electron microscopy have shown that the connexins do not lie vertically in the membrane, but at a tilted angle, so that only a slight twisting motion is needed to close or open the channel (Unwin and Ennis, 1984).
Connexin
CM CM
Figure 1.6. Gap junctions are formed by the close apposition of two connexons coming from two adjacent cells.
The connexons are hexamers of one particular isoform of connexin crossing the cellular membrane (CM).
More than 20 connexins genes have been identified so far in mammals (Simon and
Goodenough, 1998) and they are now named according to the molecular weight predicted from their cDNA sequence. The same connexin can be expressed in a variety of tissues or organs and more than one type can be detected in the gap junctions of the same cell, with sometimes variations of expression through development. In the rodent heart, six connexins have been detected either at the mRNA or protein level (Delorme et al., 1997). Among them, cx46 is expressed at very low levels (Paul et al, 1991) whereas cx37 and cx50 are restricted to the endothelium and to atrioventricular valves respectively (Reed et al, 1993; Gourdie et al., 1992). The abundance and distribution of the remaining connexins, namely cx40, cx43 and cx45 are regulated during heart development (Delorme et al, 1995; Fromaget et al., 1992;
Alcolea et al, 1999). In the adult heart, cx43 is the major connexin of the ventricular
17 myocardium where it replaces cx40, which is in turn restricted to the conduction system and to the atria. Finally, cx45 is found preferentially in the AV node (van Veen et al, 2001).
Deletion ofgapjunction proteins expressed in the ventricular myocardium
Homozygous null-mutant mice for cx40 have conduction system abnormalities at the level of
His-Purkinje cells, which express preferentially this isoform, resulting in a partial conduction
block. In addition, intraventricular conduction is also slower than normal, possibly due to an
uncoordinated ventricular activation (Simon et al, 1998; Kirchhoffet al, 1998). The germ-
line deletion of cx43 is more severe and leads to an aberrant cardiac development
characterised by the malformation of the conus region overlying the pulmonary outflow tract.
This phenotype could be due to a change in the migratory behaviour of neural crest cells,
which normally express cx43, and participate in the outflow tract formation (Reaume et al,
1995; Lo et al, 1997). In contrast to the germ-line deletion, the conditional inactivation of
cx43 in the heart using either MLC2 or aMHC driven Cre transgenic animals is lethal at one
month respectively two months after birth (Gutstein et al, 2001). Mice suffer from ventricular
conduction slowing and die of arrhythmia but the development of the heart is normal, either
because there is no cardiomyocyte cell-autonomous requirement for cx43 during heart
development or because the inactivation of the cx43 gene by the Cre recombinase takes place
mainly after birth. Finally, two groups have reported the generation of a cx45 knockout. These
mice die in utero, with a dilated heart due to defective vascular development. During the first
24 hours of heart activity, the atria develop severe contraction defects and atrio-ventricular
conduction block has been observed (Kumai et al, 2000; Kruger et al, 2000) and very likely
indicate the requirement of a functional AV node for a normal conduction system.
Regulation ofconnexin 43
Cx43 is the most widely expressed connexin in tissues and cell lines. At the transcriptional
level, cx43 expression in cardiomyocytes is upregulated, among other signals, by cAMP and
the Wnt pathway (van der Heyden et al, 1998; Ai et al, 2000) and downregulated in response
to c-jun N-terminal kinase (INK) activation (Petrich et al, 2002). At post-transcriptional
level, the phosphorylation of cx43, primarily on serine amino acids located in the carboxy-
terminal cytoplasmic end of the protein, appears to influence gap junctional communication in
both a positive and a negative manner. Connexins are modified by numerous kinases, of
which some have been identified: protein kinase C (Doble et al, 2000), p38-MAPK and the
proto-oncogene v-Src (Cottrell et al, 2003). In addition to this control through
18 phosphorylation, cx43 is characterised by a rapid turnover, measured in adult rat heart
(Beardslee et al, 1998), due to efficient proteolytic degradation by both lysosomal and
proteosomal pathways. Finally, many reports demonstrate that close apposition of the
membranes brought by other adhesion molecules is a prerequisite for the formation of gap junctions (Hertig et al, 1996a; Kostin et al, 1999; Wu et al, 2003) and that transfection of
dominant negative N-cadherin with a deletion the of extracellular domains impaired gap junction formation (Hertig et al, 1996b).
Alteration ofconnexin 43 expression in cardiac disease
Altered cell-to-cell electrical coupling has been reported in human ischemic heart disease and
non-ischemic hypertrophic heart disease (Peters et al, 1993). In fact, the perturbation can
occur at two different levels. On one hand, the normal bipolar distribution of cx43-based gap junction is redistributed to the lateral surface of the membrane or internalised in the
cytoplasm. This abnormality was first reported in the myocardial zone bordering infarct scar
tissue in the ventricles of patients with end-stage ischemic heart disease (Smith et al, 1991)
and further characterised in animal models (Matsushita et al, 1999; Daleau et al, 2001). On
the other hand, connexin 43 gene expression is substantially downregulated in the heart of
patients suffering from ischemic heart disease, idiopatic dilated or inflammatory
cardiomyopathies (Dupont et al, 2001; Kostin et al, 2003). Similarly, a reduction of the
number of gap junctions has also been observed in mouse models for dilated cardiomyopathy
(Hall et al, 2000; Ehler et al, 2001).
1.2.8 The costameres: cell to extracellular matrix contacts in contractile cells
Cardiomyocytes and skeletal myotubes adhere to the extracellular matrix (ECM) at vinculin-
rich striated plaques known as costameres (Pardo et al, 1983). In this structure (Figure 1.7),
the extracellular domain of the transmembrane heterodimeric protein integrin binds to the
ECM made predominantly of collagen fibres, laminin and fibronectin secreted by cardiac
fibroblasts (Eghbali et al, 1988; Kim et al, 1999). On the cytoplasmic side, the ß-integrin
subunit binds to an attachment complex containing talin, a-actinin (Otey et al, 1993),
vinculin, paxillin, zyxin and other costameric associated proteins (Hemler, 1998), which in
turn bind to the Z-line of the contractile apparatus (Borg et al, 2000).
Interaction of cells with the ECM is a key element in tissue integrity and organisation. For this
reason, the molecular interaction of integrins with the ECM has been extensively explored,
leading to the discovery of short sequences responsible for the binding of integrins to their
extracellular support. Among them, the RGD motif and the LDV sequence (both found in the
19 fibronectin protein) have been characterised in detail (Mould et al, 1991). As mentioned before, integrins are heterodimeric receptors and a large variation of a and ß subunit combination are found in tissues and define the binding to specific ECM proteins. In myocardium, different a-subunits, are expressed by cardiomyocytes. Among them ai, a3, a5 a6, a-i are the most important (Hornberger ct al, 2000; Maitra et al, 2000) and always found in combination with splice variants of the ßi subunit. All these combinations are subject to developmental regulation. For example, ai and as subunits are expressed only in the embryonic heart whereas ßm is expressed in the postnatal heart and is the predominant ßi splice variant of the adult heart (van der Flier et al, 1997).
Figure 1.7: Schematic representation of selected elements of the costamere. The transmembrane proteins integrins are binding to the extracellular matrix (ECM) and provide stabilisation of the sarcomeric structure via their actin attachment complex (talin, vinculin). Other protein complexes such as dystrophin/ankyrin/spectrin participate also in this cell-to-ECM contacts. The intermediate filament desmin is not represented on this figure.
Deletion of costameric proteins
The germ-line deletion of ßi integrin is lethal soon after implantation (Fässler and Meyer,
1995). However, ßi-null ES cells can differentiate into beating cardiomyocytes via embryoid bodies and adult cardiomyocytes via injection into wild-type blastocysts. The deletion of ßi integrin leads to a delay in the in vitro differentiation and an alteration of the sarcomeric assembly in vivo, particularly the actin filaments, (Fässler et al, 1996). Mice with ßi integrin inactivation restricted to ventricular cardiomyocytes develop intolerance to haemodynamic loading, as consequence of cardiac fibrosis and ultrastructural abnormalities (Shai ct al,
2002). In the same way, expression of dominant negative ßl integrin in the heart resulted in
20 early postnatal lethality, cardiac fibrosis and decreased heart function (Keller et al, 2001). As mentioned before, the knockout of vinculin is lethal at embryonic day 9.5 but ES cells can
still be differentiated into cardiomyocytes in vitro (Xu et al, 1998).
Signalling through integrin
Integrins can transmit signals from the outside of the cell by conformational changes upon binding to ECM proteins. This modification leads to cytoskeletal reorganisation through
GTPase proteins and intracellular signalling through protein kinases such as focal adhesion kinase (Taylor et al, 2000; Torsoni et al, 2003), Akt/PKB, and MAPK pathway (Berken et
al, 2003). In reverse orientation, intracellular signalling can alter the adhesive properties of
integrins and influence the clustering of these receptors to the membrane. In neonatal rat
cardiomyocytes (NRC), forced expression of ßl-integrin induces a hypertrophic response and
expression of cytoplasmic dominant negative domains of the same protein suppresses the
adrenergic-mediated hypertrophic signal (Ross et al, 1998).
1.3 Role of ß-catenin in signalling
1.3.1 Wnt signalling
Apart from its structural role in adherens junctions, ß-catenin is also the effector of the Wnt
(wingless/int) signalling pathway. First studies made in Drosophila identified Armadillo, the
Drosophila homologue of ß-catenin as a segment polarity gene controlling anterior-posterior
identity of larval segments during embryogenesis. Soon after, geneticists linked ß-
catenin/Armadillo to the wingless pathway (Peifer et al, 1991; McCrea et al, 1993).
Identified also in mammalian cells, this pathway controls the cytoplasmic concentration of ß-
catenin at the post-transcriptional level (Figure 1.8). As mentioned before, unstimulated cells
maintain low levels of ß-catenin though serine/threonine phosphorylation by a so-called
destruction complex containing among others GSK-3ß, axin (Behrens et al, 1998),
adenomatous polyposis coli protein (APC, Peifer, 1996), and ß-transducin repeat-containing
protein (ß-TrCP, Hart et al, 1999). This phosphorylation is followed by ubiquitination and
subsequent degradation by the proteasome (Aberle et al, 1997; Orford et al, 1997). When
Wnt glycoprotein binds to its transmembrane receptor Frizzled (Fz), GSK-3ß is inactivated by
Dishevelled (Dvl) and ß-catenin is no longer degraded. As a consequence, the cytoplasmic
concentration of ß-catenin increases and ß-catenin can be translocated to the nucleus where it
binds, among others, to a family of TCF/LEF transcription factors and activates a variety of
21 responding genes like CyclinDl, c-Myc, fibronectin and, as specifically demonstrated in cardiac cells, conncxin43 (Ai et al, 2000). ß-catenin can also be transported to the cell membrane and participates in the reinforcement of adherens junctions. Many steps of the Wnt pathways are still poorly understood: what we know about the cascade of events from Fz binding to GSK-3ß inactivation is the involvement of caséine kinase Is (CKIs, Sakanaka et al, 1999) and the subsequent sequestering of axin with phosphorylatcd Dvl from the active destruction complex (Julius et al, 2000). Another poorly understood point in the Wnt signalling pathway is the nuclear translocation capacities of ß-catenin: studies have shown that ß-catenin lacks a defined nuclear localisation sequence and thus must rely on shuttling proteins for transportation in and out of the nucleus (Koike et al, 2004).
* * äC*^*y
Figure 1.8: Simplified view of the Wnt signalling pathway. In response to the binding of Wnt, Frizzled transmembrane receptor activates dishevelled (Dvl, red) which blocks the activity of the serine/threonine kinase glycogen synthase kinase 3 beta (GSK3ß, green). This inactivation prevents the phosphorylation of ß-catenin by its so-called phosphorylation complex (made of GSK3ß axin and APC) and its subsequent destruction by the proteasome. The increase of ß-catenin stability causes both nuclear translocation followed by transactivation of downstream targets and the enhancement of adherens junction formation
Plakoglobin and the Wnt signalling
Plakoglobin and ß-catenin have closely related structures (about 68% of identity in the mouse protein sequence), share a common function in the adherens junction and arc both tightly regulated through serine/threonine phosphorylation. However, in contrast to ß-catenin, plakoglobin interacts also with desmosomal Cadherins and is therefore an important
22 constituent of the desmosome. Another divergence between the two homologous proteins is the transactivation capacity of plakoglobin. Indeed, if the evidence for a TCF/LEF transactivation role of ß-catenin has been clearly demonstrated in many studies and has dramatic consequences in many oncogenic processes (Rubinfeld et al, 1997; Polakis et al,
1999), the participation of plakoglobin in the Wnt signalling as effector is still not clear. As the C-terminal end that contains the transactivation domain is the most divergent sequence between ß-catenin and plakoglobin, it is possible to conceive that the affinity of plakoglobin
for the TCF/LEF DNA binding proteins is low and/or that plakoglobin might activate
different target genes by recruiting different transcription co-factors (Ben-Ze'ev and Geiger,
1998; Zhurinsky et al, 2000). In the same direction, the time shift observed between the
lethality of ß-catenin knockout mice and plakoglobin-null mutation (E7 compared to El0.5-
E16, see Table 1.1) could mirror the unique property of each protein: a signalling molecule
driving mesoderm formation for the former and a desmosomal component for the latter.
Wnt and ß-catenin signalling in the myocardium
Many interesting questions still unanswered concern the participation of Wnt activity in the
development of the heart and/or the involvement of Wnt signalling in normal heart function
and in cardiomyopathies. Data from ES cells, P19 carcinoma cells and chicken embryo
expiants about Wnt activity during heart development are conflicting (Marvin et al, 2001;
Nakamura et al, 2003; Schmidt et al, 2001). Perhaps one way to reconcile these results with
each other is to postulate an early requirement of Wnt activity for the induction of mesoderm
(Liu et al, 1999; Morkel et al, 2003) followed by a rapid downregulation allowing the
differentiation of progenitor cells into cardiomyocytes. Consistent with this downregulation, a
transgenic mouse with a TCF ß-galactosidase reporter element failed to show any Wnt
activity in the heart region with the exception of the septum transversum (persistant up to
E10.5) and the outflow tract (from E10.5 to E12.5) likely corresponding to neural crest
derived cells (Maretto et al, 2003).
Recent evidence suggests that ß-catenin is stabilised in human cardiac hypertrophy (Rezvani
and Liew, 2000), in neonatal rat cardiomyocytes under ß-adrenergic stimulation and in adult
rats with trans-aortic constriction (Haq et al, 2003). In these cases however, GSK-3ß
inactivation and ß-catenin stabilisation is not depending on Wnt activity but occurs via serine-
9 phosphorylation of GSK-3ß predominantly through the PI3K/Akt pathway. Moreover,
cardiac hypertrophy can be suppressed both in vitro and in vivo by the expression of a
23 constitutive active form of GSK-3ß (Haq et al., 2000; Antos et al., 2002) leading to phosphorylation and subsequent degradation of ß-catenin. All together, these results place
GSK-3ß in a central position for the regulation of hypertrophic events in the heart biology field.
1.3.2 Other regulators of ß-catenin
The ubiquitously expressed prolyl isomerase Pinl binds to a phosphorylated serine-proline motif next to the APC binding site in ß-catenin and inhibits the interaction between both proteins leading to elevated levels of free ß-catenin. Reciprocally, ß-catenin is considerably reduced in several organs of Pinl -/- mice including the heart. Even if these mice display several abnormalities including decreased body weight and testicular atrophy, it is interesting to note that no report has shown so far any cardiac dysfunction (Ryo et al, 2001; Liou et al,
2002). Recently, a new striated muscle specific ubiquitin ligase, Ozz-E3, has been discovered.
This protein interacts with ß-catenin and promotes its degradation by the proteosomal machinery. As expected, the targeted deletion of this gene leads to an increased stability of ß- catenin in heart and skeletal muscle but surprisingly, this stability is restricted in myoblasts to the membrane-bound fraction of ß-catenin. More surprising in this Ozz-E3 knockout is the sarcomeric alteration found in skeletal tissue associated with a relocation of the nucleus towards the centre of the myotube (Nastasi et al, 2004), which is a defined pathological trait known as centronuclear myopathy (Riggs et al, 2003).
1.4 The conditional knockout approach
1.4.1 Cre/lox technology
The conditional knockout approach, which is a refinement in the field of transgenic manipulation, relies on known methods for DNA transfer combined with the new powerful properties of recombinases and their DNA recognition sequences. This technique allows the specific inactivation of a target gene in a chosen tissue. This restriction is useful to distinguish the effect of the deletion in a given organ from the pleiotropic effect of the germ-line deletion.
From far the most popular system used is the combination of the Cre recombinase (Figure
1.9) discovered in bacteriophage PI, together with sequences flanked by specific 34bp recognition sequences called loxP sites (Hoess et al, 1984; Rajewsky et al, 1996).
24 —
- loxP Transgenic loxP Target loxP Gene A Target v^eno
Figure 1.9: Schematic representation of the specificity of the Cre/lox technology adapted for the conditional inactivation of genes in the laboratory mouse (Sauer, 1998). The system consists on one side of a target gene partially or totally surrounded by two LoxP sites and on the other side of the Cre recombinase expressed under the control of a tissue-specific promoter. Upon Cre activation, the sequence of the targeted gene "floxed" by
LoxP sites is removed leading to a loss of function.
When expressed, Cre recombinase catalyses the excision of the sequence between the loxP sites. Therefore the deletion can be controlled in terms of tissue specificity by the choice of the promoter driving the Cre expression, whereas the location of the loxP sites defines the exact extent of the deletion. The conditional knockout technique has been very successfully applied to mouse transgenesis (for review see Lewandoski, 2001) and many mouse strains have been generated for the specific expression of Cre in almost every tissue. In the cardiac field, a restricted number of transgenic Cre-mice exist and each of them has its own specificity in term of activity and tissue restricted expression (Table 1, Ruiz-Lozano and
Chien, 2003).
25 Table 1.2: Conditional mutagenesis of diverse alleles in distinct cardiovascular cell lineages
(adapted from Ruiz-Lozano and Chien, 2003)
Cre mouse Targeted Tissue Reference
MLC2v (Knock-in) Ventricular myocytes (Chen etal, 1998b)
MLC2a (Knock-in) Atrial and ventricular myocytes (Huang et al, 2003) aMHC Atrial and ventricular myocytes (Agah etal, 1997)
MCK Skeletal and cardiac muscle (Bruning etal, 1998)
Pax3 Neural crest (Li et al, 2000)
Wnt-1 Neural crest (Danielian etal, 1998)
Tie-2 Endothelium (Kisanuki etal, 2001)
SM-22 Smooth muscle (Kuhbandner et al, 2000)
MLC: myosin light chain, MHC: myosin heavy chain, MCK, muscle (isoform of) creatine kinase,
SM: smooth muscle
Further improvements in the Cre/lox technology are now available to provide an inducible
Cre expression, either by the use of doxycycline activated promoters (Utomo et al, 1999) or by fusion of the Cre recombinase with mutated versions of the estrogen-receptor ligand- binding domains, which can be induced by tamoxifen injection (Sohal et al, 2001).
1.4.2 MLC2v-Cre knock-in mouse
of Prof. K. Chien in The MLC2v-Cre mouse was generated by Dr Ju Chen in the laboratory
which San Diego. This transgenic mouse was created with a knock-in strategy (Figure 1.10),
of the MLC2v means that the Cre coding sequence is placed under the control endogenous
instead of the normal promoter, creating a transgenic allele (further refered as MLC2vCre) coding sequence (Chen et al, 1998b).
26 1kb ATG Wt allele X BEH B X B X X màmkhmàAi—[ J_L E1 E2
MLC2v5'
Cre allele IRES
BEH E EXB H X X L J L Cre GFf] §^GK-Neoflx/
-loxP-
Cre5'Cre3' pA
MLC2v5'
Figure 1.10: 5' Structure of the mouse myosin light chain 2v (MLC2v) gene (adapted from Chen et al, 1998). In the Cre allele, the two first cxons (El, V2 in WT allele) are replaced by the sequence of the Cre recombinase.
The PGK-Neo cassette shown between the two LoxP sites (red triangles) is removed upon Cre expression
rc/t K As MLC2v is essential for heart function, MLC2vc is lethal and thus Cre-expressing animals must always have one copy of the wild type allele. In this heterozygous situation, the amount of MLC2v mRNA is reduced to 50% of the wild-type situation while MLC2v protein accumulates at levels comparable to wild-type heart. Moreover, MLC2vCre/wt and wild type animals have similar cardiac functions and exhibit no difference in response to induced hypertrophy (Minamisawa ct al, 1999a). From this report, we can conclude that both Cre expression and MLC2v hemizygosity are not deleterious to the heart.
MLC2v-Cre in practice
So far, MLC2v was used for the heart specific deletion of many floxed genes: the retinoic acid receptor-a RARa (Chen et al, 1998a), the interleukin receptor gpl30 (Hirota ct al, 1999), connexin 43 (Gutstein et al, 2001), the epidermal growth factor receptor ErbB2 (Ozcelik ct al, 2002), ßi-integrin (Shai et al, 2002), Notch 1 (A. Croquclois, personal communication) and Nkx2.5 (Pashmlbroush et al, 2004). It is worth to note that MLC2v-Cre mediated deletion can be traced in slow skeletal muscle, which is in accordance with the MLC2v expression pattern. Remarkably, all these conditionally inactivated genes did not show any embryonic lethality and the postnatal phenotypes were found in most cases in mice older than one month of age.
27 1.4.3 ß-catenin floxed mouse
As ß-catenin germ-line deletion displays embryonic lethality before most cell types emerge, little can be learned about the specific requirement of different organs or tissues for this protein. The Cre/lox approach was therefore chosen to control the tissue-specific deletion of the gene and thus circumvent the lethality problem. A transgenic mouse harbouring loxP sites in front of exon 2 and behind exon 7 was created in the laboratory of Prof R. Kemler (Figure
1.11). Upon Cre expression, the ATG start codon and more than half of the coding sequence is removed, leading to the complete inactivation of the gene (Brault et al, 2001).
Wt allele ww4a^, a re;
^J2M3T— 4 sUsL, [7] JÖi-
El 5 nU LJ I—ILJ LJe n LJ nU
floxed allele a re; 4**-BM4^ Er-*
floxdel allele
! g —^^^^—
1kb
Figure l.l l: 5' structure of the murine ß-catenin locus (adapted from Brault et al, 2001). The translation start
(ATG) is located on the second exon. In the floxed allele, genomic sequence between exons 2 and 6 (blue boxes) is surrounded by LoxP sites (red triangles). Upon Cre activity, the N-terminal coding sequence between the
LoxP sites is removed leading to a non-functional gene (black exons in the floxdel allele).
As previously mentioned, this transgenic mouse has been successfully used for different conditional knockouts, using the Cre recombinase under the control of the Wntl promoter to direct the recombination predominantly in neural tissue, the Kl9 promoter for an early deletion in primitive endoderm and the Tie-2 promoter for driving the recombination in endothelial cells. Wntl-mediated deletion leads to dramatical brain malformation (Brault et al, 2001). When looking at this conditional knockout in more detail, one can see that some neural crest-derived structures like melanocytes and the sensory neurons of the dorsal root ganglia are missing (Hari et al, 2002). As, the differentiation of neural crest cells into sensory neurons is indeed directed by the interaction of Wnt signalling/ß-catenin (Lee et al, 2004), this demonstrates the role of this signalling pathway in fate specification during neural development. The removal of ß-catenin in the embryonic endodermal cells causes the trans- differentiation of endoderm to precardiac mesoderm and results in the formation of multiple hearts in the endodermal layer (Lickert et al, 2002), which is a demonstration of the requirement of ß-catenin for the correct etablishment of the three germ layers. Last but not
28 least, ß-catenin deletion in endothelial cells resulting in vascular defects and remodelling of the adherens junction and displays also an embryonic lethality (Cattelino et al, 2003).
1.5 Cardiomyopathies
1.5.1 Presentation of cardiomyopathies
Cardiomyopathies are by definition diseases of the heart muscle. They can be subdivided into ischemic and non-ischemic cardiomyopathies. Ischemic cardiomyopathies result from coronary artery disease, such as heart attacks whereas the aetiology of non-ischemic cardiomyopathies is much less clear and several types have been defined based on pathological observations. In humans, the two main common types are dilated cardiomyopathy (DCM) and hypertrophic cardiomyopathy (HCM). From a clinical point of view, DCM is by far the most common non-ischemic cardiomyopathy with a prevalence almost twice as high as HCM (Codd et al, 1989). The disease is characterised by the dilation of one or both ventricular chambers together with a thinning of the chamber walls. This dilation reduces the pumping ability of the heart (systolic function) and therefore a reduced ejection fraction is a hallmark of the disease, which strongly reduces exercise capacities of the patients and leads to heart failure. Epidemiologic data show that DCM occurs most often in middle-aged people and more often in the male population. By contrast, HCM is a relatively rare disease characterised by an abnormal thickening of the heart walls, predominantly in the left ventricular wall and the septum. The thickening reduces the size of the pumping chamber and prevents the heart from properly relaxing between beats and therefore properly filling with blood. The disease affects men and women of all ages and autosomal dominant inheritance is frequently observed, termed familiar hypertrophic cardiomyopathy.
1.5.2 Mutations in contractile and structural proteins lead to cardiomyopathies
In addition to signalling pathways triggered by extracellular stimuli or problems in the calcium handling during the cycles of contraction, a common cause of both DCM and HCM is mutations affecting the contractile apparatus and structural cytoskeletal proteins. Initial studies of human familiar HCM and DCM linked mutations of contractile proteins, which impair force generation to HCM whereas mutations of structural components, which affect force transmission were linked to DCM (Chien, 1999; Seidman and Seidman, 2001). This general hypothesis of defined mutations leading to predictable cardiac disease has been challenged by the study of different transgenic mice, where mutations in contractile and cytoskeletal proteins have been created. Although the hypothesis is partially true, some results
29 are surprising (Table 1.3, see also Chu et al, 2002) for a complete list of references). In particular, a point mutation introduced into the ß-myosin heavy chain gene leads to HCM, which then progresses to DCM (Freeman et al, 2001). All together, these data imply a certain level of molecular interconnection between the two historically separated cardiomyopathies, with the idea of DCM being the end stage of certain cases of HCM (Figure 1.12 and Chien,
1999).
Table 1.3: Non-exhaustive list of genetically altered mouse models of HCM and DCM,
(adapted from Chu et al, 2002).
Contractile proteins
Gene Genotype Phenotype
a-MHC a-MHC (-/-) Death at El 1-12 d of gross heart defects
a-MHC (+/-) Cardiac hypertrophy, depressed function, fibrosis, and sarcomeric
abnormalities
Arg403Gln (-/-) Death at 1 week of age
Arg403Gln (+/-) Cardiac hypertrophy, disarray, and fibrosis increased with age
ß-MHC R403Q Cardiac hypertrophy that progressed to dilated cardiomyopathy overexpression
function and a- Aspl75Asn Mild cardiac hypertrophy with impaired myocyte
Tropomyosin disorganisation.
Tropomodulin Overexpression Dilated cardiomyopathy in juvenile mice with severe myofibrillar disorganisation.
Cytoskeletal proteins
Gene mutation Phenotype
MLP MLP (-/-) Dilated cardiomyopathy with hypertrophy, disrupted myocyte
cytoarchitecture, and heart failure after birth.
AËP ALP (-/-) Right ventricle dilated cardiomyopathy
Desmin Desmin (-/-) Dilated cardiomyopathy, concentric cardiomyocyte hypertrophy.
Dominant negative Desmin-related cardiomyopathy with abnormal intra-sarcoplasmic
aggregates.
For References and further explanation of abbreviations, see (Chu et al, 2002).
30 Sarcomeric j> ^j^Ä^V Cytoskeletal
Genetic modifiers ? *-
Hypertrophic Dilated cardiomyopathy cardiomyopathy (HCM) (DCM)
Figure 1.12: Defects in sarcomeric or cytoskeletal components lead either to hypertrophic cardiomyopathy
(HCM) or dilated cardiomyopathy (DCM). For reason not clearly understood, some initially HCM pathologies like ßMHC R403Q (see Table 1.3) can be modified with time to DCM (Chien, 1999)
1.5.3 MLP KO model
The striated muscle-specific LIM-only protein MLP was first described as a positive regulator of myogenic differentiation (Arber et al, 1994). In skeletal muscle, MLP expression correlates with the onset of cell differentiation and accumulates during myotube formation but is then down regulated in adult tissues and can be re-expressed upon muscle denervation. In contrast, heart expresses high levels of MLP both during development and in the adult stage.
Like other LIM proteins, MLP has both a nuclear and a cytoplasmic localisation. It has been demonstrated that the nuclear fraction associates with MyoD and enhances its activity, which explains the regulatory involvement of MLP in differentiation (Kong et al, 1997). In the cytoplasm, MLP localises with the cortical actin-based cytoskeleton and the sarcomeric Z- disc structure via interactions with ßi spectrin (Flick and Konieczny, 2000), a-actinin and zyxin (Louis et al, 1997). New investigations suggest a role of stress-sensor for this Z-disc complex, which is affected in a subset of human DCM (Knoll et al, 2002; Knoll et al, 2003).
Indeed, the targeted deletion of MLP in adult mice shows a profound alteration of the cardiac structure and the functional characteristics of DCM (Arber et al, 1997). It is worth to note that the first generations of MLP knockout mice displayed a phenotype visible between postnatal day 5 to 12, characterised by ANF upregulation and lethal increase in heart size but this early lethality disappeared after further crossing, suggesting that the genetic background
31 could modulate this acute hypertrophic response. Moreover, the MLP adult phenotype can be rescued by the ablation of phosholamban, the inhibitor of the sarcoplasmic reticulum calcium
(SERCA) pump. This suggests that MLP-/- cardiomyocytes have a defective calcium handling system (Minamisawa et al, 1999b). Interesting for the context of our research, MLP knockout mice display a general defect in the structure of the intercalated disc combined with the upregulation of adherens junction proteins and downregulation of gap junctions (Ehler et al, 2001).
1.5.4 DRAL KO model
DRAL (down regulated in rhabdomyosarcoma LIM protein) also called FHL2 (four and a half
LIM domain protein 2) is expressed specifically in the developing heart. During late stages of development and in the adult, DRAL expression is mainly restricted to the ventricular myocardium. As MLP, DRAL seems to have a dual function: although it interacts with the cytoplasmic domain of several integrin chains (Wixler et al, 2000), the giant sarcomeric protein titin and metabolic enzymes (Scholl et al, 2000; Lange et al, 2002), DRAL is also a transcriptional co-activator of the androgen receptor (AR) and of ß-catenin (Muller et al,
2000; Martin et al, 2002). A specific trimeric interaction exists between DRAL, AR and ß- catenin (Yang et al, 2002; Mulholland et al, 2002), which seems to be specific because the other members of the steroid hormone receptor family and the FHL1/3 proteins do not participate in this interaction. The net transactivation result of the binding is less clear and could be either positive or negative for ß-catenin and AR responding genes, depending on the tissue context (Pawlowski et al, 2002; Wei et al, 2003). Mice with ablation of DRAL are much less affected than MLP-/- animals and cardiac parameters, even after aortic banding, are comparable to wild-type mice treated the same way (Chu et al, 2000). The only phenotype observed is an exaggerated hypertrophic response after adrenergic stimulation with isoproterenol (Kong et al, 2001) therefore DRAL knockout cannot be considered as a model of cardiomyopathy sensu stricto.
1.6 Aim of the study
1.6.1 ICD development and integration of myofibrils
The central starting point of the project is the description of ICD development in relation with myofibril integration and changes in cellular polarisation. So far, the ICD organisation has been investigated in the postnatal and adult myocardium (Angst et al, 1997), in cultivated cells and in some rare cases in the embryonic stages (Tokuyasu and Maher, 1987; Shiraishi et
32 al, 1993; Wu et al, 2002) but a comparative description of all the components of the ICD in relationship with myofibrils development is still missing. Early embryonic cardiomyocytes are round-shaped cells surrounded by adherens junctions with myofibrils running along the cell membrane in criss-cross conformation while in postnatal and adult stages, the cells are rod-shaped with a perfect parallel alignment of sarcomeric subunits and a clear separation between lateral costameric contacts to the ECM and restricted cell-cell contacts at the bipolar
ends. What remains to be discovered is the transition between the embryonic and the adult
organisation and the mechanisms which trigger the maturation of cardiomyocytes.
Immunohisochemistry and confocal microscopy of whole mount heart preparations has been
selected as the methods of choice in order to stay as close as possible to the in vivo situation
and avoid possible artefacts created by isolation procedures or cell culture. Combination of
specific antibodies will allow the visualisation of each type of junction, as well as
extracellular matrix components and sarcomeric proteins. Comparison of confocal
micrographs at different time point of the embryonic development together with
morphometric measurements should then reveal critical steps in the maturation of
cardiomyocytes.
1.6.2 Reconstruction of ICD in vitro
This part of the project is dedicated to the expression of fluorescent fusion protein in
the cardiomyocytes. Pioneer work in our laboratory (Komiyama et al, 1996) has shown
possibility to transfect neonatal cardiomyocytes with expression plasmids and the correct
localisation of transfected GFP fusion proteins in the sarcomeric structure (Auerbach et al,
1997). The aim here is to reconstruct in vitro the interface between cell-cell contact and
sarcomeric elements. For this task, we need to evaluate several candidate genes, which would
with sarcomeric localise at cell-cell contacts upon transfection and express them together
fusion proteins. If this approach works, it should enable us to get more insight into the
dynamics of ICD assembly.
1.6.3 Déstabilisation of the ICD
In the third part of the thesis, we want to challenge the ICD stability by removing one of its
important components. As shown in Table 1.1, the germ-line deletion of ICD components is
often lethal and to circumvent this problem we have to use a conditional knockout approach
(Figure 1.8), which restricts the choice of candidate to available floxed genes. For this reason
of the cadherin/catenin ß-catenin was chosen although this protein has a dual function as part
complex and as the effector of the Wnt signalling pathway, which can of course seriously
33 complicate the analysis of a possible phenotype. In the first approach, we assume that ß- catenin does not play a major role in transducing Wnt signals in the heart ventricles. This assumption is based on the fact that TCF-ßgal transgenic mice, which express the reporter gene ß-galactosidase upon ß-catenin nuclear translocation and TCF binding, failed to detect any ß-catenin activity in the embryonic and postnatal myocardium (Maretto et al, 2003).
Moreover, the absence of nuclear localisation of ß-catenin in cardiac tissue or isolated neonatal cells, under physiological conditions, corroborates this hypothesis.
In this part of the work, the first preoccupation is to show the viability (or lethality) of mice bearing the ß-catenin deletion recombination. A secondary element is to demonstrate the specificity and efficiency of the deletion both at the genomic and at the protein level. After this, we have to analyse the possible effects of the ß-catenin removal on other ICD proteins and on the heart functionality in general.
1.6.4 Additional stress on heart cells
In the case that ß-catenin deletion does not display any cardiac phenotype per se in normal conditions, it is worth thinking that a phenotype could still be induced when an additional stress is applied to the heart (Kong et al, 2001). For this, several stress-models are established
of based on further genetic modifications, which means breeding with knockout models cardiomyopathies (Chu et al, 2002) or surgical interventions like hypertension-induced hypertrophy (Wiesel et al, 1997) or hypertrophy induced either by transaortic ligation
(Tarnavski et al, 2004) or ß-adrenergic stimulation (Tsukazaki et al, 1995). As we are dealing with a protein with dual function, it would be interesting to induce on one hand a stress on the contractile apparatus (leading to HCM) or on the force transducing apparatus
(leading to DCM) to test the cytoskeletal properties of ß-catenin. On the other hand, deleting
ß-catenin together with a specific co-activator would bring more insights into its signalling activity.
34 2 Material and Methods
2.1 Cloning methods
2.1.1 Plasmid manipulation
Plasmid digestion
For subsequent cloning, 2 pg of plasmid DNA were digested with 10 U of restriction
enzyme(s) (New England Biolabs, Bioconcept, Allschwil, Switzerland) in the appropriate buffer for 1-2 hours. When needed, digestion mixes were supplemented with lx bovine serum
albumin (BSA, supplied by New England Biolabs). For analytical purposes, the digestion was
essentially scaled down, using 200 ng-1 u.g of plasmid DNA with 5-10 U of enzyme for 1 hour.
Cloning PCR
PCR reactions were carried out in 50 u.1 volumes with the following components: 10 ng template DNA, lx Pfu polymerase buffer (Promega, Catalys AG, Wallisellen, Switzerland), 1
uM of each primer, 0.2 mM dNTPs, 2 U Pfu polymerase (Promega). Reactions were run on
an Eppendorf Mastercycler Gradient (Vaudaux-Eppendorf AG, Schönenbuch, Switzerland)
using the following cycling conditions: 95°C for 2 minutes, 35 cycles of 95°C for 30 seconds,
annealing at 55-65°C for 30 seconds (temperature depending on primers), elongation at 68°C
for variable time (depending on the length of the amplicon, usually 2 minutes per one kb) and
a termination step at 68°C for 10 minutes.
Standard agarose gel (0.5-2%)
Agarose (Eurogentec SA, Seraing, Belgium) was melted in 0.5x Tris-borate-EDTA buffer
(TBE buffer), cooled to 50°C and mixed with 0.7 ug/ml EtBr before pouring in
electrophoresis chambers (OWL, Axon Lab AG, Baden-Dättwil, Switzerland). DNA probes
were mixed with DNA loading buffer (6x TBE buffer, 0.25% bromophenol blue, 0.25%
xylene cyanol FF, 15% Ficoll Type 400 (Amersham Biosciences, Otelfingen, Switzerland) in
water) 4:1 and electrophoresis was run for 30 minutes to 2 hours at 100-150 V.
0.5x TBE buffer (for llitre): 5.4 g Tris-Base, 2.75 g Boric acid, 0.46 g Na2EDTA
35 Isolation ofDNA fragments from agarose gels and DNA clean up
Digested DNA was separated on standard agarose gels according to size and DNA bands were visualised with a long-range UV lamp to avoid DNA damages. Scalpels were used to excise the bands corresponding to the expected fragments and the DNA was purified using
NucleoSpin Extract kit (Macherey-Nagel AG, Oensingen, Switzerland) according to the manufacturer's instructions. Yield of the isolated fragments was judged by running aliquots of them on standard agarose gels and comparing the band intensity with the standard lkb DNA ladder (Eurogentec SA).
Ligation ofDNA fragments
DNA fragments were quantified on agarose gels as mentioned above. Typically, 50 ng of linearised vector was used for ligation. Three different ligations were set up: a ratio of vector to insert of 1:1 and 1:3 and a control without insert. The ligation was made in a final volume of 10 pi containing lx of T4 DNA ligase buffer and 200 U of T4 DNA ligase (New England
Biolabs) for 2 hours at 22°C or overnight at 4°C.
Oligo linker ligation
When small modifications of plasmids were required, for example for the modification of multicloning sites, the oligo linker method was performed as follows: the vector was linearised with 2 restriction enzymes leaving non-cohesive ends. 2 oligonucleotides (HPLC purified) of about 60 bases, which are complementary to produce a core of double-strand
DNA and 2 overhangs necessary for the insertion are mixed in a total volume of 20 ul at a concentration of 50 uM, warmed for 5 minutes at 65°C and cooled down to RT. From this mix lui was used for ligation with the linearised vector.
2.1.2 Transformation of competent cells and bacterial culture
Competent cells and transformation
Calcium competent cells were made according to the method of Inoue (Inoue et al, 1990).
For transformation, competent cells were thawed on ice and 5 ul of the ligation mix was added to 100 ul of bacterial solution. The tubes were incubated for 30 minutes on ice and then the cells were heat-shocked at 42°C for 90 seconds, then returned on ice for 5 minutes. 400 pi of SOC medium was added to the tubes and incubated for 1 hour at 37°C under constant agitation. 50-200 pi of the final bacterial solution (depending on the quality of the competent cells and the size of the final construct) were plated on LB-Agar plates containing the
36 appropriate antibiotic for selection (100 pg/ml for ampicillin, 50 pg/ml for kanamycin) and left overnight at 37°C. The next day, single colonies were grown in liquid culture of LB containing the appropriate antibiotic (see above) overnight at 37°C with agitation.
White-blue screening
When inserts were ligated in pBluescript or pGEX-T vectors containing the a complementing
fragment of the LacZ gene, a supplemental selection was made by using XL 1-blue
competent cells (Stratagene, Amsterdam, The Netherlands) and by covering the LB-Agar plates with 20 ul of X-Gal solution (50 mg/ml in dimethylformamide) and 100 pi of 100 mM
IPTG 30 minutes before plating the bacteria.
The colonies containing inserts appeared white because the insertion destroyed the
complementation of the a fragment to form functional ß-galactosidase, which hydrolyses X-
gal into a blue precipitate.
SOC medium (for 1 litre): 20 g bacto-peptone, 5 g bacto-yeast extract, 0.5 g NaCl; autoclave;
add KCl to 2.5 mM, MgCl2 and MgS04 tolO mM, glucose to 20 mM.
LB medium (for llitre): 10 g bacto-peptone, 5 g bacto-yeast extract, 10 g NaCl, adjust pH
to7.0 with NaOH; autoclave.
15 of litre of LB agar plates: prepare LB medium as described above and add g agar per
solution; autoclave; let cool to 50°C, then add the appropriate antibiotic and pour on plates;
store poured plates at 4°C.
2.1.3 Plasmid DNA isolation
For the isolation of plasmid DNA at small scale, miniprep kits (Macherey-Nagel) were used
with 5 ml of bacterial culture according to the manufacturer instruction. For larger amounts of
DNA suitable for transfection, midiprep kits or maxiprep kits (Macherey-Nagel) were used
with 25, respectively 100 ml of bacterial culture. At the end of the purification procedure,
plasmid DNA was resuspended in 10 mM Tris-HCl [pH8.5]. The integrity of plasmid was
Plasmid judged on an agarose gel and quantification was made by UV spectrophotometry.
DNA was considered to be clean if the ratio A26o: A2go was above 1.8. The amount of DNA
at -20°C until was calcultated using 1 OD26o=50 ug/ml DNA. Plasmid DNA was stored
further use.
2.1.4 Sequencing PCR
200 PCR reactions were carried out in 20 pi volumes with the following components: ng
template DNA, 3ul 5x sequencing dilution buffer (400 mM Tris-HCl [pH9.0], 10 mM MgCl2)
37 0.15 uM primer, 2 pi termination ready reaction mix (PE Applied Biosystems). Reactions were overlaid with mineral oil and run on an Eppendorf Mastercycler Gradient (Vaudaux-
Eppendorf) using the following conditions: denaturation at 96°C for 2 minutes, followed by
25 cycles of 96°C for 10 seconds, annealing at 50°C for 5 seconds, extensions at 60°C for 4 minutes.
2.2 RNA quantification
2.2.1 Total RNA isolation
Tissue samples were homogenised in Trizol Reagent (Gibco BRL, Life Technologies AG,
Basel, Switzerland) using a Polytron mixer (Kinematica AG, Luzern, Switzerland). The next steps of the isolation were carried out according to the manufacturer's instructions. With this method applied for heart tissue, a yield of 1 pg of total RNA/mg of tissue can be expected. The RNA pellet was resuspended in DEPC-treated water and quantified by UV spectrophotometry at 260 and 280 nm. RNA was considered to be clean if the ratio A26o/A28o was above 1.7. The amount of RNA was calculated using 1 OD26o = 40 pg/ml RNA. For convenience, RNA was diluted to 1 ug/ul and stored in small aliquots at -80°C.
2.2.2 RT-PCR analysis
Reverse transcriptase reaction cDNA synthesis was made with the Thermoscript RT-PCR system kit (Invitrogen, Basel,
Switzerland). Briefly, denaturation step was carried out in PCR tubes (Axon Lab AG, Baden-
Dättwil, Switzerland) in a volume of 12 pi containing 1 pg of total RNA, 50 ng of random hexamers and 1.66 mM dNTPs. This mix was heated for 5 minutes at 65°C then placed on ice. 8 pi of cDNA synthesis mix (containing 2.5x cDNA buffer, 15 U of Thermoscript RT,
12.5 mM DTT and 40 U of RNAse OUT) was added and the annealing step was initiated at
25°C for 10 minutes, followed by cDNA synthesis at 50°C for 40 minutes. cDNA was stored at -20°C until used for PCR amplification.
Semi-quantitative PCR reactions
PCR reactions were carried out in 50 pi volumes with the following components: 2 pi of template cDNA, lx Taq polymerase buffer (Gencraft GmbH, Lüdinhausen, Germany), 0.2 pM of each primer, 0.2 mM dNTPs, 2.5 U Taq polymerase (Gencraft). Reactions were run on an Eppendorf Mastercycler Gradient (Vaudaux-Eppendorf) using the following cycling conditions: 95°C for 2minutes, cycles of 95°C for 30 seconds, annealing at 60°C for 30
38 seconds, elongation at 72°C for 45 seconds. The number of cycles to fit into the exponential
amplification was determined for each set of primers in a trial experiment. The amplification
products were run on standard agarose gels, isolated and sequenced to ascertain accurate gene
amplification. The list of amplified genes, amplification size, cycle numbers and primers pairs
are shown in Table 2.1
Table 2.1: Primers and settings for the amplification of hypertrophic markers and control
genes ,.r,Jg§p|pi 1111 üäI ||lfel»illlfj||ilgl llislfe ANF (atrial natriuretic factor) 23 ANF5' CTCTGAGAGACGGCAGTGCT 402bp
ANF3' TATGCAGAGTGGGAGAGGCA
BNP (brain natriuretic peptide) 23 BNP5' GAGGCGAGACAAGGGAGAACA 375bp
BNP3' CGATCCGGTCTATCTTGTGCC
GAPDH (glyccraldehyde 3- 19 GAPDH5' CTTGAAGGGTGGAGCCAAACG 532bp
phosphate dehydrogenase) GAPDH3' GCTGTTGAAGTCGCAGGAGACAA
PLN (phospholamban) 19 PLN 5' CACTGTGACGATCACCGAAGC 303bp
PLN3' GGCGGCAGCTCTTCACAGA
ctSK («-skeletal actin) 25 aSK5' CCCTGGACTTCGAGAATGAGATG 399bp
aSK3' GGAAGGTGGACAGCGAGGC
ßMHC (ß-myosin heavy chain) 25 ßMHC5' AAGAACCTACTGCGGCTCCA 310bp
ßMHC3' TCCACCTAAAGGGCTGTTGC
N-RAP (ncbulin-rclatcd 23 N-RAP5' GCGTCCCGAGGAGTTTCATA 596bp
anchoring protein) N-RAP3' CTGCTGCTGTTCACGGTCr
Cre recombinase Cre5' GTTCGCAAGAACCTGATGGACA 340bp
Cre3' CTAGAGCCTGTTTTGCACGTTC
2.3 Immunoblotting
2.3.1 SDS-sample preparation
35 mg of frozen tissue was homogenised by freeze-slamming and solubilised in 500 pi of
SDS-sample buffer (3.7 M Urea, 134.6 mM Tris-HCl [pH6.8], 5.4% SDS, 2.3% NP-40,
4.45% ß-mercapto-ethanol, 4% glycerol, 60 mg/1 bromophenol blue) boiled for 5 minutes,
sonicated in a water bath and spun down briefly to avoid pipetting large cellular debris.
Samples were stored at -20°C until loading on acrylamide gels.
2.3.2 Electrophoresis and transfer
The SDS-PAGE electrophoresis was performed with minigel apparatus (Bio-Rad Laboratories
AG, Life Science Research, Reinach, Switzerland) according to the method of Laemmli
39 (Laemmli, 1970). For equal loading of protein, a test gel was run and stained with Coomassie blue solution (0.1% (w/w) Coomassie Brilliant-blue R-250, 50% Methanol, 10% Acetic acid) followed by destaining with a 10% Acetic acid solution and intensity of protein bands was compared to equalise loadings in following experiments. 10 or 15% Acrylamide gels were used depending on the size of the proteins to be separated. Kaleidoscope Prestained Standard
(7-200 kD) was used as marker (Bio-Rad). The gels were run at 120 V for the stacking gel and 200 V for the separating gel until the bromophenol blue line reached the bottom of the casting chamber.
2.3.3 Antibodies for immunoblot
Table 2.2: Primary antibodies for immunoblots HHHNHHNMR iiPSfPlllillj Wq^Mamm'S^iMMSIXîSM&M^M^MiM^m^m^^Sâ
mM a plakoglobin 1 :2000 Transduction Laboratories, Lexington, UK
mM a vinculin, clone hVin-1 1:1000 Sigma, Buchs, Switzerland
mM a ß-catenin 1:2000 Transduction Laboratories
mM a desmoglein, clone 62 1:1000 Transduction Laboratories
pR a a-actinin 1:1000 provided by Dr M. Gautel, King's College, London, UK
pR a aE-catenin 1 :1000 Sigma
pR a aT-catenin, serum n°952 1:500 Obtained from Dr F. Van Roy, University of Gent, Belgium
pR a ANF 1 :500 ANAWA, Wangen bei Dubendorf, Switzerland
pR a connexin43 1:250 Zymed, P.H Stehlin & Co. AG, Basel, Switzerland
pR a laminin 1:1000 Sigma
pR a pan-cadherin 1:1000 Sigma
pR a all actin 1:1000 Sigma
pR a desmoplakin 1:500 Serotec, Oxford, UK
Table 2.3: Secondary antibodies for immunoblots
ifillilÄ
HRPO G a M IgG 1:3000 DAKO Diagnostics AG, Zug, Switzerland
HRPOGa R IgG 1:2000 Calbiochem, Juro supply, Luzern, Switzerland
40 2.3.4 Blotting and immunodetection
Blotting was performed overnight at 60 mA in a wet chamber blotting apparatus (Mini- electrophoretic Blotting System; CBS Scientific Company Inc., Axon Lab AG) onto a nitrocellulose membrane (Hybond-C-extra, Amersham Life Science). Blotting buffer consisted of (for 1 litre): 192 mM glycine, 25 mM Tris-HCl [pH8.3], 0.01% SDS, 20%
Methanol. Protein transfer was checked by staining the blots with Ponceau red (Ponceau S solution; Serva, Heidelberg, Germany). Blots were incubated in blocking solution* for at least
1 hour, followed by incubation with the primary antibody diluted in First antibody solution** for 1 hour at RT in rolling 50 ml Falcon tubes. After washing the blots 3x 5 minutes with blocking solution, the incubation with the secondary antibody was performed for 1 hour in low salt buffer (9 mM Tris-HCl [pH7.4], 154 mM NaCl, 0.1% Tween-20). After three more washes with low salt buffer, chemiluminescence detection was performed by adding to the membrane 8 ml of 125 mM Tris-HCl [pH7.5] 0.05% hydrogen peroxide, 1 ml of 25 mM
Luminol and 1 ml of 5 mM Iodophenol. X-Ray films were exposed in the dark on the membrane for different period of time (1 second to 15 minutes, depending on the signal intensity) and developed in a X-Ray film developer apparatus.
*Blocking solution consists of either 5% non fat dry milk powder or 3% BSA in low salt buffer
**First antibody solution consists of either 2.5% non fat dry milk powder or 1.5% BSA in low salt buffer
2.3.5 Densitometrie analysis
X-Ray films were scanned in grey scale at 300 dpi with a HPScanjet Ilex hardware (Hewlett- packard Company, Palo Alto, USA) and the digital image was submitted to densitometric analysis using Image J software (National Institutes of Health, Washington, USA). The densitometric value represents the mean of the grey intensity of all the pixels present in a selected area. For normalisation, a densitometric analysis is performed on a selected reference, e.g. actin content for standardisation.
2.4 Isolation of mouse heart cells and cryosections
2.4.1 Isolation of rodent heart
Mice or rats were killed by cervical dislocation and the hearts were immediately excised and placed on ice in PBS to wash the blood from the ventricular and atrial chambers. For subsequent culture, atria were discarded with scalpels.
41 2.4.2 Dissociation and culture of neonatal rat cardiomyocytes (NRCs)
Newborn hearts were digested with collagenase (Worthington Biochemical Corp., Freehold,
USA) and pancreatin (Gibco BRL) and cultured as described elsewhere (Rothen-Rutishauser et al., 1998; Sen et al., 1988; Komiyama et al., 1996) modified for rat cells according to
of (Auerbach et al., 1999). Cells were plated onto fibronectin-coated dishes with a density
the medium was either 0.4xl06 cells per 35 mm culture dish in plating medium. The next day, changed to a culture medium or a transfection procedure was carried out.
Plating medium: 68% Dulbecco's MEM (Amimed AG, Basel, Switzerland), 17% Medium
Ml99 (Amimed AG), 10% horse serum (Gibco BRL), 5% fetal calf serum (Gibco BRL), 4 mM Glutamine (Amimed AG) and 1% penicillin-streptomycin (Amimed AG).
Maintenance medium: 75% DBSS-K buffer (116 mM NaCl, 1 mM NaH2P045 0.8 mM
MgS04, 5.5 mM Glucose, 32.1 mMNaHC03, 1.8 mM CaCl2), 20% medium M199,1% horse
serum, 1% penicillin-streptomycin, 4 mM Glutamine and 0.1 mM phenylephrine (Sigma).
2.4.3 Dissociation and culture of neonatal mouse cardiomyocytes
The culture of neonatal mouse cardiomyocytes was made essentially as for neonatal rat
cardiomyocytes except that Verapamil (Sigma) was added during the isolation procedure and
the Ficoll gradient step was obmitted, which increased the percentage of non-cardiomyocytes
cells in the culture. For this reason, 10 pM Ara-C (ICN Biochemicals, Cleveland, USA) was
added to the maintenance medium to prevent the proliferation of contaminating fibroblasts.
2.4.4 Isolation of adult heart cells
After cannulation of the aorta, the heart was mounted on a Langendorf setup with a flow rate
20.5 mM 1.8 mM of 2 ml per minute. Tyrode's solution, (137 mM NaCl, 5.4 mM KCl, MgCl,
CaCb 11.8 mM Na-HEPES, 10 mM glucose; pH 7.4) was used for washing out the blood.
1.2 After this, the heart was perfused with calcium-free Tyrode (130 mM NaCl 5.4 mM KCl,
mM KH2P04 1.2 mM MgS04 6 mM Na-HEPES; pH 7.2) for about 6 minutes. The next step
dissolved in the was the perfusion with 0.1 mg/ml of Blendzyme 3 (Roche Diagnostics AG)
calcium-free Tyrode, and this solution was usually recirculated for 12 minutes. During this
time the tissue became soft and friable. On the last step, the enzyme was washed out with
calcium-free Tyrode supplemented with 0.18 mM CaC^ for about 15 minutes. The ventricles
cell was were removed, and gently minced and agitated in a beaker until a cloudy suspension
obtained. Cells were allowed to settle and were washed twice with the same solution. After
the isolation, 100 pi of cell suspension were immobilised on gelatinised microscope slides by
short centrifugation on a Cytopsin apparatus (Shandon Southern Inc, Pittsburgh, USA).
42 2.4.5 Cryosections
Hearts rinsed in ice cooled PBS were frozen in Isopentan cooled in liquid nitrogen, 'fhe heart was mounted in Tissue tek OCT medium (Piano W.Plannet AG, Wetzlar, Germany) and cryosection (20 pm) were cut in an HM560 MV cryostat (Microm, Walldorf, Germany) at -
20°C, then retrieved on gelatinised microscope slides for immunostaining.
2.5 Transfection of neonatal rat cardiomyocytes
Prior to transfection, cells were grown overnight in plating medium in incubators with 10%
CO2. Two hours before transfection the cells were changed to transfection medium
(73%DBSS-K buffer 20% M199, 4% horse serum, 4 mM glutamine, 1% penicillin- streptomycin). Transfections were carried out as following: 1 pg of plasmid DNA was diluted inlOOpl of OptiMEM (Gibco BRL) and added to 100 ul of OptiMEM containing 4 pi of
Escort III reagent (Sigma). After 12 minutes of incubation at RT, 800 pi of transfection medium was added to the mix, vortexed and given to the cells, previously washed twice with
Optimem. The transfection was carried out for 4 hours in 5% CO2 then the medium was replaced by maintenance medium and cells were cultured for 2 days before fixation.
2.6 Fixation, immunofluorescence staining and apoptosis detection
2.6.1 Antibodies used in immunofluorescence
Table 2.4: Primary antibodies for immunofluorescence
mM a sarc. a-actinin, clone EA53 1:500 Sigma
mM a IgM titin clone 9D10 1:2 Developmental Studies Hybridoma Bank, Iowa City, USA mM a myomesin, clone B4 1:50 Obtained from E.Perriard, ETH, Zürich, Switzerland pR a EH-myomesin 1:1000 Obtained from I. Agarkova, ETH, Zürich, Switzerland mM a ot-cardiac actin, clone ad-20.4.2 1:20 Progen, Heidelberg, Germany
pR a MyBP-C 1 :100 Obtained from Dr M. Gautel, King's College, London, UK
mM a desmin, clone D33 1 :100 DAKO Diagnostics AG
pR a pan-cadherin 1:100 Sigma
pR a ß-catenin 1:200 Sigma
43 mM a ß-catenin 1:100 Transduction Laboratories pR a aE-catenin 1 :250 Sigma
mM a vinculin, clone hVin-1 1:20 Sigma
mM a plakoglobin, clone 15 1:100 Transduction Laboratories
pR a desmoplakin 1:200 Serotec
pR a connexin 43 1:100 Zymed
pR a laminin 1:100 Sigma
Table 2.5: Secondary antibodies for immunofluorescence
G a M FITC 1:50 ICN Cappel, Eschwege, Germany
G a M IgM (|i-chain specific) FITC 1:100 Sigma
G a R Cy2 1:500 Jackson laboratory, Bar Harbor, Maine, USA
G a M IgG (y-chain specific) Cy3 1:200 Jackson laboratory
G a M Cy3 1:1000 Jackson laboratory
G a R Cy3 1:200 Jackson laboratory
G a M Cy5 1:100 Jackson laboratory
G a R Cy5 1:100 Jackson laboratory
Table 2.6: Non-immune dyes
^^MjEß^^&Ä^il Sil Jillll:
Alexa488-phalloidirt 1:50 Molecular probes, Eugene, OR, USA
Alexa633-phalloidin 1:50 Molecular probes
DAP I 1:100 Molecular probes
Picogreen 1:250 Molecular probes
44 2.6.2 Immunofluorescence of heart whole mount preparations
Heart whole-mount preparations were fixed with 4% PFA (in PBS) for 1 hour, then rinsed 3x
5 minutes with PBT (PBS with 0.002% Triton X-100). After 30 minutes of permeabilisation with 0.1% Triton X-100 (in PBS) and 2 washes with PBT, cardiac jelly was digested with 1 mg/ml of hyaluronidase (in PBS). This digestion was followed with 3 washes with PBT and
30 minutes of blocking with 5% normal goat serum in gold buffer (10 mM Tris-HCl [pH7.2],
155 mM NaCl, 2 mM EGTA, 2 mM MgCl2, 1% BSA). Both primary and secondary antibodies were diluted in gold buffer. Incubation with the primary antibody solution was performed overnight at 4°C, followed by 5x 20 minutes washes with PBS. As before, the
overnight incubation with the secondary antibody was followed by 5x washes with PBS.
Finally, the cells were mounted in Lisbeth's medium (0.1M Tris-HCl [pH9.5]: glycerol (3:7)
including 50 mg/ml of N-propyl-gallate) to prevent the fading of the samples.
2.6.3 Immunofluorescence of isolated cells
Cells were fixed 10 minutes with 4% PFA (in PBS), rinsed once with PBS than blocked 5
minutes with 0.1 M glycine (in PBS). In order to permeabilise the membrane, cells were
incubated 10 minutes with 0.2% Triton X-100 (in PBS) and washed 3 times. Primary and
secondary antibodies were diluted in gold buffer. The cells were incubated in the primary
antibody for 1 hour at RT, and then washed 3x followed by the incubation with the secondary
solution for 1 hour at RT. After three subsequent washes, the cells were mounted in Lisbeth
medium.
2.6.4 Immunofluorescence of cryosections
The procedure for the staining of cryosections is the same as for isolated cells except that the
sections were blocked with 5% normal goat serum (in gold buffer) before the incubation with
antibodies and that the incubations with primary and secondary antibodies were made
overnight at 4°C.
2.6.5 In Situ apoptosis detection
Detection of apoptosis was made with the in situ cell death detection AP kit (Roche
Diagnostics AG, Rotkreuz, Switzerland) on tissue cryosections according to the
manufacturer's instructions. Both positive (DNAse I treatment) and negative (no terminal
deoxynucleotidyl transferase) controls were performed together with the analysed samples.
45 2.7 Microscopy
2.7.1 Standard fluorescence microscopy
Conventional fluorescence microscopy was done on a Zeiss AxioLab microscope (Carl Zeiss
AG, Feldbach, Switzerland) equipped with an AxioPlan 100x/1.25 oil objective. Pictures were taken with an HRm Axiocam CCD camera and recorded using Axiocam software. Final processing of the images was done with Photoshop software (Adobe systems, San Jose,
USA).
2.7.2 Confocal microscopy
Analysis of the stained cells or cryosections was carried out using a confocal microscope. The
imaging system consisted of a Leica inverted microscope DM IRB/E (Leica Microsystems
AG, Glattbrugg, Switzerland) a Leica confocal scanner TCS NT/SP1 and a PC workstation.
The images were recorded using a Leica PL APO 63x/1.4 oil objective from the bottom to the
top with a Z-axis step size of 0.2-0.6 pm. Image processing was done on a Fujitsu Siemens
computer (Fujitsu Siemens computer Bv., Regensdorf, Switzerland) using Imaris (Bitplane
AG, Zürich, Switzerland) a 3D multi-channel image processing software specialised for
confocal images (Messerli et al., 1993)
2.7.3 Cell measurements and volume reconstruction
Imaris 4.0 (Bitplane AG) was used to visualise single section out of the confocal data set and
to measure the cell length and width. In whole mount preparations and tissue sections, the
labelling of ICD components and/or laminin staining identified the cell length and width. For
the calculation of cell cross sectional area in embryonic and neonatal cardiomyocytes, length
and width measured in the middle of the confocal stack were considered as the long and short
= axis of an ellipsoidal object. Therefore, the cell cross sectional area (CSA) 7iLW/4. Because
not all the confocal data used here had a nuclear staining, nuclear CSA was calculated from
the apparent hole in the cytoplasm which is devoid of sarcomeres. A pilot experiment in order
to validate this method was successfully made by calculating the nuclear CSA as mentioned
here and compare it with DAPI labelling (data not shown).
3D Volume rendering and volume calculation was made using the Surpass programme
(Bitplane AG). The application of this software to the data set allows the quantification of the
volume of particles above a specific threshold. Surpass uses the marching cube algorithm that
creates triangle models of constant density surfaces. Applied to isolated cardiomyocytes with
F-actin staining, this programme precisely delineates the volume of the cell.
46 2.8 Generation of conditional knockout mice and double knockouts by breeding
2.8.1 Animal strains, genetic background and maintenance
Cre/+ flox/flox MLC2v (Chen et al., 1998b), ß-catenin (Brault et al., 2001) MLP knockout (Arber et al., 1997) and DRAL knockout mice (Kong et al., 2001) were maintained at the animal facility of the Institute of Cell Biology at the Swiss Federal Institute of Technology (Zurich,
Cre/+ Switzerland). The genetic background of the mice is C57B1/6 for MLC2v and ß-catenin flox/flox whereas MLp ko and DRAL KO mice have a mixed B6;129 genetic background.
Maintenance and procedures followed were in accordance with the Swiss Federal Veterinary
Office (BVET) guidelines.
2.8.2 Genotyping
Mouse identification and DNA extraction
cut for Mice at weaning age were ear-clipped for identification and a piece of the tail was
DNA extraction. Briefly, 0.5 ml of lysis buffer (0.1M Tris-HCl [pH 8.5], 5 mM EDTA, 0.2%
SDS, 0.2M NaCl, 0.2 mg/ml Proteinase K (Sigma)) were added to the tail biopsy and
digestion took place for 2 hours at 55°C under constant agitation. After vortexing, extracts
was collected were centrifuged to separate hairs and bones from digested tissues. Supernatant
down and DNA was precipitated by addition of one volume of Isopropanol and spinned by
in 0.5 ml of centrifugation. After a wash with 70% Ethanol, DNA pellets were resuspended
TE buffer (10 mM Tris-HCl [pH7.5], 10 mM EDTA [pH8.0] and heated at 80°C until
complete dissolution.
were killed cervical The same procedure was used for neonatal mice except that pups by
dislocation and the centrifugation to remove hairs and bones was not necessary.
Genomic PCR
PCR reactions were carried out in 25 pi volumes with the following components: 1 pi of
genomic DNA extract, lx Taq polymerase buffer (Gencraft), 0.25 pM of each primer, 0.2
mM dNTPs, 1.25 U Taq polymerase (Gencraft). Reaction mixes were overlayed with mineral
oil (Sigma) and PCR tubes were placed on an Eppendorf Mastercycler Gradient (Vaudaux- Eppendorf) using the conditions described in Table 2.7
47 Table 2.7: Settings for genomic PCRs and interpretations of PCR amplification
Mi » «in wwJMm MLC2v- lx 2min 95°C Cre5' GTTCGCAAGAACCTGATGGACA 340bp for
Cre MLC2v: Cre/+
40x lmin lmin Cre3' CTAGAGCCTGTTTTGCACGTTC 95°C, 60°C, no amplification
2min 72°C in WT SRY lx 2min 95°C SRY5' GTTCAGCCCTACAGCCACA'l 197bp for male
mice
SRY3' CAGCTGCTTGCTGATCTCTG 40x lmin95°C, lmin60°C, no amplification
in female ß-catenin lx 2min 95°C RM41 AAGGTAGAGTGATGAAAGTTGTT 500bp Tor floxdel 324bp
40x lmin95°C, lmin60°C, RM42 CACCATGTCCTCTGTCTATTC for flox 221bp
2min 72°C for WT RM43 TACACTATTGAATCACAGGGACTT
MLPWT lx 2min 95°C GACCCAGGGCTGTTTGC l.SkbpforWT MLPS no amplification
40x 30scc 95°C, 45sec ACAATATTGACCTGTCCCC in MLP KO
63.5°C, 2min 72°C MLPdel
MLP KO lx 2min 95°C GACCCAGGGCTGTTTGC 1.5 kbp for MLP
MLPs KO,
30sec 45sec GTTCAATGGCCGATCCC 40x 95°C, no amplification MLPneo 63.5°C, 2min 72°C inWT
DRAL lx 2min 95°C TGACTGAACGCrrrGACTGC 400 bp for
DRAL1647 DRAL KO
40x 30sec 95°C, 45sec ATGGGTGTTCCACACTCCTC 200 bp for WT DRAL1648 58.1°C, 45sec 72°C CCCATTACGGTCAATCCGCCG
GCCTCCAGTACAGCGCGGCTG
LacZ as
ex 43 lx 2min 95 °C cx43ct5' CCTGCTGAGAACCTACAT C 366 bp, used
as internai 40x lmin 95°C\ lmin60°C, cx433' GAGCAGCCATTGAAGTAAGC control 2min 72°C
2.9 Hypertrophy induction and echocardiography
The hypetrophy inductions and echocardiography reported in this section 2.9 were perfomed
in the laboratory of Professor Dr T. Pedrazzini in the Hypertension division of the University
of Lausanne, Switzerland.
48 2.9.1 Hypertension-induced hypertrophy (1K1C model)
One kidney-one clip surgery (Wiesel et al., 1997) was performed on anaesthetised (1-2%
Halotan in oxygen) male mice that were 6 months of age. Briefly, a 0.11-mm clip was inserted on the left renal artery to chronically reduce the perfusion pressure, and a right nephrectomy was performed at the same time. Echocardiographic measurements were recorded before and 4 weeks after surgery. After this, the animals were sacrificed and the heart weight-to-body weight ratio was measured (milligrams/gram).
2.9.2 ß-adrenergic stimulation of the heart
Miniosmotic pumps (Alzet model 2002) were filled with Isoproterenol, and were set to deliver this drug at 30 mg kg"1 day"1 for 14 days. Pumps were then implanted in 9-10 weeks
old male and female mice under Isofluran anaesthesia. At the end of the stimulation phase, the heart contractility of animals was measured by echocardiography under anaesthesia. After this, the animals were sacrificed and the heart weight-to-body weight ratio was measured
(milligrams/gram).
2.9.3 Echocardiography of the mouse heart
Left ventricular dimensions were assessed in anesthetised mice (90 mg/kg ketamine, 5 mg/kg
xylazine, intraperitoneal) by echocardiography using an ATL HDI 5000 ultrasound machine
(Philips Medical Systems, Bothell, Washington, USA) equipped with a 12-MHz phase array
linear transducer (LI2-5). M-mode images were used for measurement of left ventricular
anterior wall (LVAW) thickness, left ventricular (LV) dimension, intraventricular septum
(IVS) thickness at end diastole (D) and end systole (S).
2.10 Magnetic resonance imaging (MRI)
The MRI study was performed on El5.5 embryos, which were exsanguinated in ice-cold PBS
and genotyped as mentioned before. Embryos were then fixed for 2 weeks in 4% PFA (in
PBS), placed in individual Eppendorf tubes containing PBS and shipped with Fedex express
mail to the laboratory of Professor Dr Bhattacharya (Department of Cardiovascular Medicine,
Oxford, UK) where the analysis was performed. Embryos were embedded in 1% agarose
(Seakem) containing 2 mM gadolinium-diethylenetriamine pentaacetic anhydride (Gd-DTPA)
in nuclear magnetic resonance tubes and the MRI was recorded as stated by (Schneider et al.,
2003).
49 2.11 Statistical analysis
All Statistical analyses and tests were carried out using Excel software (Microsoft, Redmond,
USA). Data are given as mean +/- standard deviation. Bars in graphs represent standard errors and differences analysed with a two-tailed T test with a P value below 0.05 were considered significant.
50 3 Results
3.1 Development of the intercalated disc in the heart of mouse embryos
3.1.1 Overview of the results
In the first section of the results, the emergence of the first ICD like structures will be analysed by immunocytochemistry and confocal microscopy from the first stage where heart differentiation becomes obvious (E8.5) to the situation of the heart of newborn and postnatal mice. The integration of myofibrils is essential for the mechanical coupling and heart function and thus special attention will be given to myofibril anchoring at the developing cardiomyocyte contacts, which later become the ICD. As development proceeds, the myofibril orientation changes and an alignment emerges parallel to the long axis of the cardiomyocytes. The alignment of myofibrils is a critical feature of striated muscle and in the heart must be achieved in single cells and coordinated with the neighbouring cells. It is very likely that this progression is also mediated at the ICD but presently this process of integration of the contractile organelles is not clear.
3.1.2 Myofibrillar and morphological changes in cardiomyocytes from embryonic stage to adult
The development of the heart of vertebrates is characterised by a dramatic change in cellular architecture. This fact is illustrated by the series of confocal micrographs shown in Figure 3.1 where embryonic, neonatal and adult preparations are immunostained with antibodies against the cell-cell contact proteins Cadherins, a myofibrillar marker (titin or myomesin). The fluorescent dye DAPI was also added to visualise nuclei. We can observe that the embryonic heart cells are polygonal in shape surrounded by adherens junctions (Figure 3.1.A) while the nascent myofibrils are not oriented (Figure 3.1.B). At neonatal stages, as illustrated in Figure
3.1.E, the adherens junction is no longer surrounding the cells but gets restricted to the longitudinal cell border whereas the alignment of the myofibrils is increasing (Figure 3.1.F).
The adult cardiomyocytes are extremely polarised cells, with brighter ICD at the longitudinal cell border (Figure 3.1.1) and well aligned myofibrillar structures (Figure 3.1.J). When looking at the DAPI staining (Figure 3.1.C,G,K) the density of nuclei is strikingly decreasing from embryonic to adult heart, as a consequence of the developmental hypertrophy.
In the Figure 3.2, embryonic, neonatal and adult heart preparations are immunostained with antibodies against the adherens junction protein ß-catenin and the ECM component laminin to
51 Figure 3 1. Myofibrillar and morphological changes are affecting cardiomyocytes throughout embryonic and postnatal development. E14.5 whole mount (A,B,C,D), P2 (E,F,G,H) and 6 months old (I,J,K,L) cryosections are immunostained against the adherens junction protein Cadherin (A,EJ green channel in overlays) and against a sarcomeric marker to show the myofibrillar structures (titin 9D10 in B, myomesin in F and J, red channel in overlays). The nuclei are visualised by DAPI staining (C,G,K, blue channel in overlays). Note the changes in adherens junction distribution during development (A with E and I), the progressive alignment of the myofibrillar structures (B with F and J) and the decrease in nucleus density due to the increase of the cellular volume (C with G and K). Bar, 15 urn. Figure 3.2. Changes between cell-cell contacts and cell-to-extracellular matrix contacts during cardiac development. El 0.5 whole mount (A,B,C), P2 (D,E,F) and 6 months old (G,II,I) heart sections are immunostained against the adherens junction protein ß-catenin (A,D,G green channel in overlays) and against laminin to show the extracellular matrix distribution (B,h ,H, red channel in overlays). As seen in figure 3.1, the restriction of the surrounding adherens junctions towards bipolarity is obvious (compare A with C). During this time, the deposition of extracellular matrix allows the formation of cell- to-matrix lateral contacts as visualised at birth (IÎ) and in adult cells (H). Bar, 15um. compare cell-cell and cell-to-ECM contacts. As stated before, adherens junctions are
surrounding all the embryonic cardiomyocytes (Figure 3.2.A). In contrast, laminin deposition
is not present around all cells (Figure 3.2.B) a situation commented later in section 3.1.11.
After birth, the elongation of the cells creates a junctional and a lateral side where cell-cell
contacts respectively cell-to-ECM contacts are going to be restricted (Figure 3.2.D,E). In
adult cells, the ECM deposition has excluded the cell-cell contacts components from the
lateral side of the membrane. As a consequence cells are now bipolar, with bright ICDs at the junctional sides of the membrane and lateral membranes embedded in the ECM (Figure
3.2.G,H).
3.1.3 Appearance of cardiomyocytes during heart development
The first contractions of the mouse heart can be observed between embryonic day 8.5 to 9.
This fact indicates that cardiomyocytes, which are by definition the beating cells of the heart,
have already assembled enough sarcomeres to form functional myofibrils. These myofibrils
represent the contractile machinery of the cells and thus the hallmark of cardiomyocyte
differentiation, which can therefore be used to distinguish cardiomyocytes from other cells in
tissue preparations (see also Figure 3.1.B,F,J). We know from previous work made on
chicken whole mount preparations that embryonic cardiomyocytes are polygonal cells
surrounded by adherens junctions (Shiraishi et al., 1993 see also Figure 3.1.A), which rapidly
develop their growing myofibrils along the cell border in a random orientation (Ehler et al.,
1999 see as well Figure 3.1.D) and then elongate to form in neonatal stage spindle-shaped
cells (Leu et al., 2001) with adherens and desmosomal junctions concentrated at both ends in
a bipolar fashion (see Figure 3.1.B-C), followed by the restriction of gap junction with a delay
(Angst et al., 1997) and the definitive acquisition of the rod-shaped structure by adult
cardiomyocytes. What is missing in these reports is how these changes affect the organisation
of myofibrils and how is this then correlated with the changes in cell polarity. In other words,
how a polygonal cell with growing non-oriented myofibrils knows where to restrict cell-cell
contacts made of adherens, desmosomal and gap junctions and where to establish the lateral
borders. Another unsolved question is how the growth of myofibrils affects embryonic
cardiomyocytes in terms of morphologic parameters, i.e. the changes in cell length and width
and the increasing importance of the cytoplasmic fraction with growing number of
sarcomeres. Previous work in our laboratory showed that cardiomyocytes are undergoing a
so-called developmental hypertrophy beginning at P4 (Leu et al., 2001) but little is known
about earlier time points and this study is also supposed to fill this gap.
52 3.1.4 Expression of cell-cell contact and extracellular matrix components in the heart
cell- The purpose of this experiment is to verify the expression of the different proteins of the cell contacts used later in immunofluorescence staining. A compilation of immunoblots
shown in Figure 3.3 demonstrates for each junction the expression of one specific representative: desmoplakin for the desmosome (first row), laminin for the ECM (second row), Cadherins for the adherens junction (third row) and connexin 43 for the gap junction
(fifth row), a-actinin, which is both a sarcomeric and an adherens junction protein is also
shown here (fourth row). Although submitted to increased accumulation during heart
development, the expression of actin is used here as a loading control (sixth row). All the
markers of intercalated disc and extracellular matrix components presented in Figure 3.3 are
expressed in the heart from in the earliest stage of investigation with the notable exception of
connexin 43 (fifth row, column 1-2). As mentioned above in the introduction and in a separate
paragraph below, this isoform is expressed only later in development (El 6.5) and other
connexins are expressed at this time in the heart ventricle. A second remark concerns the
protein concentration of the different markers. Laminin concentration is rapidly increasing
after birth indicating the deposition of the ECM ensheating the cells along the lateral
membrane not involved in ICD formation (Figure 3.2.E,H). Likewise the connexin 43
continues its accumulation and replaces the embryonic isoforms of connexin 40 in the ICD of
adult cardiomyocytes. Finally, a-actinin concentration increases indicating the formation of
new sarcomeres, which is discussed below.
3.1.5 Growth of myofibrils
As stated before, cardiomyocytes are by definition the contractile cells of the heart. This
contractile capacity is due to the presence of assembled myofibrils, which can be recognised
by their characteristic striated pattern in microscopic analyses. As we are interested to know
how fast these myofibrils are growing during embryonic development, we isolated hearts
from mouse embryos staged according to Theiler's criteria (Theiler, 1972). For young stages,
whole mount analyses were performed, while for stages older than El 8.5 cryosections were
carried out because of the limiting penetration of the antibody in large hearts. The samples
were subsequently stained for triple immunofluorescence with antibodies and analysis was
performed by confocal microscopy. The myofibrils were revealed at their major transverse
structures in the sarcomeres using a combination of EH-myomesin for the embryonic M-band,
sarcomeric a-actinin for the Z-disc and a titin epitope near the Z-disc (titin 9D10). In addition
adherens to one of these myofibril markers, an antibody that recognises the transmembrane
53 Figure 3.3. Immunoblotting of intercalated disc and extracellular matrix proteins. SDS extracts of hearts isolated at the indicated time during development were immunoblotted with specific antibodies. The figure shows that at the earliest stage investigated (E9.5) adherens junctions, desmosomes and extracellular matrix are already detectable. In comparison, connexin43 can only be detected at El 6.5. Note the multiple bands corresponding to the different phosphorylation states of connexin (PO^unphosphorylated, PI= 1 phosphorlylation, P2 - 2 phosphorlyation sites). The membrane was blotted with an antibody which recognizes all actin isoforms as loading control. junction proteins Cadherins and decorates the outline of the myofibrils, was also used in the
preparation. When looking at the confocal micrographs (see Figure 3.5.D,F,H), the changes in
myofibril length during embryonic development are obvious. For statistics, the length of
myofibrils was measured for at least 20 cells using five or more different stacks of confocal
data and results are shown in Table 3.1. From 12.19 pm on average at E8.5 where they run
along the cell membrane, the myofibrils are regularly growing to reach 27.44 pm at El6.5,
38.18 pm at birth and finally reach 42.70 pm at postnatal day 4 (P4) where they fill most of
the cytoplasm. This result means that the so-called developmental hypertrophy observed in
neonatal heart (Leu et al., 2001) starts very likely already in the embryonic phase and
contributes to the increase of heart size together with the still ongoing mitotic division (Ahuja
et al., 2004). It is as well important to realise that new sarcomeric subunits must be added
inside the growing myofibrils (about one sarcomere per 2 pm) to permit the expansion of the
whole structure.
3.1.6 Orientation of myofibrils
A second question regards the organisation of the myofibrillar structures during embryonic
heart maturation. Using the same set of data as for section 3.1.5, we observed a pronounced
change in the orientation of myofibrils throughout development, from a criss-cross pattern to
a more parallel one. In order to generate a numerical interpretation of these changes of
organisation, we measured the angle between individual myofibrils and the assumed long axis
of the cell. This value should be indicative of the degree of myofibrillar orientation. Ideally
the value would be close to 45° for round-shaped cells with a criss-cross pattern of myofibrils
and approaches 0° when the myofibrils align more in the final parallel orientation. The results
are shown in the second row of Table 3.1. At E8.5 myofibrils are mostly associated with the
cell membrane and cannot be easily identified (Figure 3.5.B, arrowheads) therefore the
calculation was made with only 8 cells. Nevertheless, the trend towards a better alignment of
myofibrils is later remarkable: the average angle deviation to the longitudinal axis is reduced
from 31.15° at E9.5, to 9.55° at E16.5 and finally 5.00° at P4, which is already close to the
perfect alignment observed in adult cells.
3.1.7 Change in cell shape during cardiomyocyte development
Here we want to correlate the increasing alignment of the myofibrils with changes in cell
shape. Therefore we decided to determine the relative elongation of the cells by measuring the
ratio between length and width and estimated the increase of the cellular dimensions by
calculating the ratio of cell cross-section area versus nucleus cross-section area. The data used
54 Table 3.1. Measurement of cardiomyocytesduringembryonic and earlypostnataldevelopment
E8.5 E9.5 E12.5 E16.5 E18.5 PO P4 Myofibril length(urn) 12.19+/-3.41 12.20+/-3.31 18.53+/-2.70 27.44+/-5.46 38.18+/-5.34 38.18+/-5.76 42.70+/-5.66 Myofibril Orientation (°) 35.33+/-15.82* 31.15+/-8.50 15.33+/-5.79 9.5S+/-5.76 6.14+/-4.78 6.07+/-4.23 5.00+/-3.04 Cell length (um) 15.60+/-3.05 18.25+/-3.28 25.59+/-4.09 27.86+/-2.80 42.88+/-6.48 42.39+/-4.93 48.18+/-8.69 Cell width (um) 9.37+/-1.87 10.66+/-1.79 9.69+/-1.62 9.42+/-1.54 8.90+/-1.70 9.50+/-1.12 9.13+/-0.83 Cell CSA (urn2) 114+/-30 153+/-40 194+/-43 208+/-49 302+/-84 315+/-53 343+/-S8 Cell LAV 1.73+/-0.48 1.75+/-0.42 2.71+/-0.61 3.00+/-0.36 4.94+/-1.03 4.56+/-1.33 5.35+/-1.20 Nucleus length (urn) 12.27+/-1.95 12.39+/-1.58 13.73+/-2.02 12.53+/-1.52 12.89+/-1.83 11.24+/-1.46 11.34+/-1.31 Nucleus width
(urn) 9.05+/-1.66 8.61+/-1.28 7.73+/-1.30 7.49+/-0.99 6.30+/-0.77 6.02+/-0.93 5.88+/-0.91 Nucleus CSA (urn2) 88+/-25 84+/-21 83+/-16 74+/-16 63+/-9 53+/-10 52+/-10 Nucleus LAV 1.38+/-0.27 1.46+/-0.22 1.83+/-0.44 1.69+/-0.21 2.09+/-0.45 1.92+/-0.46 1.98+/-0.38 Ratio Cell/Nucleus 1.35+/-0.27 1.84+/-0.22 2.38+/-0.55 2.81+/-0.32 4.75+/-1.10 6.21+/-1.75 6.81+/-1.95
Values are mean +/- SD. ttLW »rea*l • nrientatinn CSA (cross section = Mvofih ril = and If»difference made h v the mvnfihril cnrm îareH tn thf".IrmirihiH inal axis nf the cell n > 20 with the exceptionof *)where n=8
CSA, LAV ratios and Ration Cell/Nucleus are means of measurements on individual cells.For this reason, for example (CellLAV)mean* LmeanAVn 60 -i T60
;f so »0 S.
V s « 4ft o o >. 30 30 ig w c Ol .£ ft 9 20 20 3
• 10 - 10
E8.S E9.5 E12.S E16.5E18.5PO P4
E8.5 E9.5 E12.5 E16.5E18.5PO
10 -i t 10
^J 6 E8.S E9.S E12.5 E16.6E18.5PO P4 Figure 3.4. The myofibril alignment, the elongation of the cells and the restriction of cell-cell contacts to the intercalated disc are related during development. From series of immunostainings with sarcomeric and cell-cell contact markers, growth parameters of cardiomyocytes can be calculated (see Table 1.) Myofibril length and alignment (A), cell length and width (B), ratio between length and width or between cell area and nucleus area (C). here are the same as discussed in sections 3.1.5 and 3.1.6. Looking at Table 3.1 (row 3-6), we is From 15.60 at can see that during embryonic development the cell growth longitudinal. pm E8.5, the length increases about three times to reach 42.39 pm at birth. The growth curve of cell length (Figure 3.4.B, triangles) resembles that of myofibrillar length (Figure 3.4.A, was observed triangles). However, during the whole process no change in cell width (3.4.B, which is also squares). This means that the cells undergo a tremendous unidirectional growth, remarkable when looking at relative cellular dimensions: the length to width ratio triples from 1.73 to 5.35 (Figure 3.4.C, triangles), whereas the ratio of cell cross-sectional area to nuclear to 75% of the cross sectional area changes from 1.35 to 6.21 (Figure 3.4.C, squares). Thus, up taken the nucleus area of the cardiomyocytes at the beginning of heart development is up by cell membrane et and it is no longer surprising that the myofibrils appear close to the (Ehler al., 1999), simply because the area close to the border is the only available cytoplasmic space This steric constraint for the of were they can assemble (see Figure 3.1.D). growth myofibrils is rapidly removed with the elongation of the cell and at birth the nucleus represents a less important restrictive element for the assembly of new sarcomeres (Figure 3.1.H). 3.1.8 Distribution of the adherens junctions during development In view of the changes occurring in myofibrillar orientation within the cell, our interest is the ICD. Adherens focused on the developmental changes of cell-cell contacts to yield bipolar of transmembrane junctions are of major interest because these complexes composed proteins related (the tissue specific Cadherins), a set of catenins (i.e. a- ß- or y-catenin), and catenin proteins (vinculin, a-actinin, pl20cas, ARVCF, p0071) link the actin cytoskeleton to the membrane. In the context of cardiomyocytes, this means that adherens junctions are actually a- and the integration sites of myofibrils. In our studies, antibodies against Cadherins, ß- similar results catenin were used in whole mount specimen and tissue cryosections and gave in immunofluorescence confocal microscopy (data not shown), y-catenin (also called as well in desmosomal structures. In plakoglobin) was not used because this protein is found order to unambiguously identify cardiomyocytes and adherens junction in tissue preparations, was used in with a an antibody specific against sarcomeric a-actinin always conjugation pan- almost cadherin specific antibody. As we can see in Figure 3.5, adherens junctions are completely surrounding the round-shaped cardiomyocytes at early stages (Figure 3.5.A-F). As mentioned above, the myofibrils are found in close contact with the cell membrane (Figure 3.5.B, arrowheads). As the cells elongate, myofibrils get aligned in a parallel orientation (Figure 3.5.E-H). This elongation creates a lattice devoid of Cadherins at the lateral side of the 55 Figure 3.5. Adherens junctions are restricted to the intercalated disc only after birth. Single confocal sections of E8.5 (A,B)* E9.5 (C,D) EI2.5 (E,F) EI6.5 (G,H) E18.5 (1,J) whole mount and PO (K,L), P4 (M,N), P10 (0,P) cryosections. The preparations were double stained with pan-cadherin (A,C,E,G,1,K,M,0) and a-actinin (B,D,F,H,J,L,N,P) antibodies. We can observe the transition from round-shaped cells surrounded with adherens junctions and few myofibrils running along the plasma membrane at E8.5 (arrowheads in A,B), then the appearance of numerous myofibrils at E9.5 (C). Further in development, cells elongate but are still surrounded with adherens junction (E12.5-16.5 (E,G)). The lateral degradation of cadherin-based contacts at El 8.5 (arrows in I) leads to the restriction of cell-cell contact at polar ends after birth (K,M) and finally condensed structures running perpendicularly to myofibrils reprensenting mature intercalated disc (arrows in O). Bar, lOum. cells (Figure 3.5.1-J), which is then followed by the restriction of the junction towards a bipolar organisation at birth and postnatal stage P4 (Figure 3.5.K-N). At the oldest stage recorded here (PIO, Figure 3.5.0-P) the first compact adherens junction localisations, corresponding to more-advanced intercalated discs are visible. The orientation of these adherens junctions is now perpendicular to the long axis of the cell (Figure 3.5.0, arrows), which is also indicative of the progression of cardiomyocytes towards a rod-shaped and bipolar structure. 3.1.9 Distribution of desmosomes Desmosomes are similar structures as adherens junctions but have different function. The transmembrane components of the desmosomes (desmogleins and desmocollins) are linking the intermediate filament system to the membrane via the binding of several intermediate proteins, e.g. desmoplakins, plakoglobin, plakophilin. In striated muscles, the intermediate filaments consist of polymers of desmin, which surround the Z-discs, interlink them together and integrate the contractile apparatus with the plasma membrane. Thus, desmosomes are believed to be responsible for the mechanical stabilisation of the contractile apparatus and it is therefore interesting to compare both cellular contacts during development. In our study, we used an antibody specific for desmoplakin together with a titin antibody (anti 9D10 titin), which recognises an epitope of the protein in the PEVK region near the Z-disc. The results in Figure 3.6 show the dotted pattern characteristic of desmosomes during cardiomyocyte development. At E9.5-E12.5 (Figure 3.6.A-D) the dotted localisation is surrounding the cells in a distribution reminiscent of epithelial cells. This distribution is not much affected by the elongation of the cells that takes place at E16.5-E18.5 (Figure 3.6.E-H). Restriction in a bipolar fashion at the nascent ICD becomes apparent at birth (Figure 3.6.1-J) and large condensed desmosomal structures become visible, very similar as the pattern of adherens junctions at PIO (Figure 3.6.K-L). Although the embryonic distribution of desmosomes is different from adherens junctions (punctuation versus complete surrounding), the restriction towards bipolarity occurs in the same way. This means that the changes affecting adherens junctions and desmosomes during development are co-ordinated probably because both junctions provide together the mechanical coupling which is essential for the work of the cell. 3.1.10 Gap junctions After adherens junction and desmosomes, gap junctions represent the third type of cell-cell contact in the ICD. As gap junctions form ion channels, they maintain the electrical coupling of individual cardiomyocytes, which then work as a functional syncytium. So far more than 56 Figure 3.6. Desmosomes are punctuate structures restricted to the intercalated disc after birth. Single confocal sections of E9.5 (A,B), EI2.5 (C,D) E16.5 (E,F) E18.5 (G,H) whole mount and PO (I,J), PIO (K,L) cryosections. The preparations were double stained with desmoplakin (A,C,E,G,I,K) and titin9D10 (B,D,F,H,J,L) antibodies. Desmoplakin shows a punctuate pattern typical for desmosomes at the plasma membrane (A,C) which get restricted to the polar junctions (E,G,I) and to the future intercalated disc after birth (arrows in K). Bar, 10pm. Figure 3.7. The development of cx43~based gap junctions is delayed to later stages of embryonic development. Single confocal sections of E9.5 (A,B), E12.5 (C,D) E16.5 (E,F) E18.5 (G,H) whole mount and PO (1,J), PIO (K,L) cryosections. The preparations were double stained with connexin43 (A,C,E,G,I,K) and a-actinin (B,D,F,H,J,L) antibodies. Cx43 forms distinct punctuations visible by immunofluorescence at El8.5 (G). Even after birth, cx43 does not localize at the intercalated disc and shows preferentially lateral staining (I,K). Bar, lOum. 20 connexin genes have been identified in mammals and all these proteins are encoded by separate genes and are named according to their respective molecular weight. The main connexins expressed in the heart are connexin 37, 40, 43, and 45. In ventricular cardiomyocytes, the embryonic isoforms consist of connexin 40 and to a lower extend connexin 45 but it is known that both isoforms are replaced during development by connexin 43 in the adult ventricular myocardium (Delorme et al., 1997; Fromaget et al., 1992; Alcolea etal., 1999). Due to the lack of a reliable antibody against connexin 40 at the time we started this study, we restricted the investigation of gap junction proteins to the connexin 43 isoform. As before, the immunofluorescence study of the connexin 43 antibody was made in conjugation with an antibody specific for sarcomeric a-actinin. As shown in Figure 3.7, we detect the presence of connexin 43 only at El8.5 (Figure 3.7.A-G) whereas the same protein can be detected already at E16.5 using immunoblotting technique, (see Figure 3.3, fifth row). At birth the punctuations observed are larger and brighter (Figure 3.7.1-J), but another clear discrepancy with regard to adherens junctions and desmosomes comes from postnatal stage (Figure 3.7.K- L) where connexin 43 is still found at the lateral membrane and does not form dense structures as observed with Cadherins or desmoplakin. These findings are in accordance with published data, which showed that the restriction of adherens junctions and desmosomes to the intercalated disc are preceding the restriction of gap junction (Angst et al., 1997). The gap junction assembly appears to be independent from the alignment of the myofibrils and thus also from adherens junctions, an idea already suggested by others (Gutstein et al., 2003). 3.1.11 Distribution of the extracellular matrix (ECM) In the myocardium, the extracellular matrix is composed mainly of collagen (Borg and Caulfield, 1981), laminin and fibronectin fibres (Kim et al., 1999), which serve as a support for the anchoring of integrin-based costameres. It is important to realise that this lateral cell- to-matrix contact does not exist at the beginning of cardiomyocyte maturation simply because the cells are forming cell-cell contacts with their surrounding neighbors and this lateral connection must be created during the elongation of the cells as illustrated by Figure 3.2. For the study of extracellular matrix deposition, a laminin antibody was used in conjugation with a sarcomeric a-actinin antibody. As shown in Figure 3.8, the laminin staining is first restricted to regions of the heart without myofibrils, possibly non-cardiomyocytes or more likely not yet differentiated myocytes (E9.5-E12.5, Figure 3.8.A-D). After further development, the signal concentrates at the lateral border, i.e. costameric regions of the 57 confocal sections of Figure 3 H. The development of extracellular matrix in the embryonic heart . Single E9.5 (A,B), E.12.5 (C,D) E16.5 (E,F) E18.5 (G,H) whole mount and PO (1,J), PIO (K,L) cryosections. The preparations were double stained with laminin (A,C,E,G,I,K) and a-actinin (B,D,F,H,J,L) antibodies. At early stages (E9.5, A), laminin seems to delineate the border of the cardiomyocyte area. At E12.5, laminin is occasionaly expressed around cardiomyocytes. From E16.5 (E), laminin is expressed at the lateral sides, corresponding to the future costameres. Bar, 10p.m. cardiomyocytes (Figure 3.8.E-J). At birth and even more pronounced at PIO, most of laminin concentrates at the lateral cell membrane (Figure 3.8.K-L) and the nascent ICDs are devoid of this protein. Taken together with the distribution of adherens junctions (Figure 3.5), these results document a transition in the lateral membrane between the still existing cadherin-based attachment of two neighbouring cells in embryonic stages (Figure 3.5.E,G,I) and the exclusive cell-to-matrix attachments as seen in the postnatal heart (Figure 3.8.I,K) and even more clearly in adult cells (Figure 3.1.1). 3.2 Labelling of the ICD and myofibrils in vitro 3.2.1 Outline of the project The idea in this project is to reconstruct partially the ICD-myofibril interface in vitro using transfection of tagged protein in neonatal cardiomyocytes, a method already used in our laboratory to follow the localisation of sarcomeric proteins (Komiyama et al., 1996). What we want to learn from this experiment is the dynamics of the ICD-myofibril assembly and the resemblances (and differences) in the organisation of myofibril attachment in vivo and in cultured cells. For this, catenin constructs (Figure 3.10) are transfected in cardiomyocytes and the localisation of each fluorescent construct is always compared with the immunostaining of Cadherins (Figure 3.10, third column), which is used as a bona fide marker of adherens junctions. The expression of gap junction proteins (Figure 3.11) and focal contact constructs (Figure 3.12) are also presented here for comparison. The final aim of this project is the expression of one particular construct which labels the adherens junction together with a myofibrillar marker in an in vitro system, which gives a three dimensional view of the myofibrillar organisation. An example of such in vitro model is shown in Figure 3.9, with the expression of a-actinin in a chicken heart expiant. In this test experiment, the heart expiant was electroporated with an a-actinin-GFP construct (Figure 3.9, green channel) then fixed and immunostained with a myomesin antibody which decorates the M-band (Figure 3.9, red channel). The localisation of a-actinin at the Z-disc in transfected cells can be appreciated by fhe regular alternation of the green (a-actinin-GFP, Z-disc) and red striation (myomesin, M-band). The picture shown here is a shadowed visualisation of confocal sections, which gives a better rendering of the object in 3D. This heart expiant model is also helpful to demonstrate the diversity in myofibrils organisation found in the cardiac tissue at embryonic stages (see arrows and arrowheads in Figure 3.9). 58 Figure 3.9. Transfection of sarcomeric a-actinin cDNA in chicken heart expiant. 3- D reconstruction and shadowed visualisation of confocal images. A three days-old heart was explanted and electroporated with a construct expressing a-actinin-GFP and cultivated for three more days on a laminin coated dish. After fixation, the heart was stained with the M-band marker myomesin. The alternative striation of a- actinin-GFP (green channel) and myomesin (red channel) are clearly visible. Depending on the region taken, the organisation of myofibrils is either criss-crossed (arrowheads) or aligned (arrows). Bar, 10pm. 3.2.2 Tagging of cDNAs and expression vectors The first technical part of the work consists of fusing different catenin proteins with the GFP fluorescent tag. Human aE-catenin, mouse ß-catenin, human plakoglobin and rat connexin 40 cDNAs were generous gifts from Dr D. Rimm (New Haven, USA), respectively Prof. R. Kemler (Freiburg, Germany), Prof. W. Franke (Heidelberg, Germany) and Prof. J.-A. Haefliger (Lausanne, Switzerland). All cDNAs were cloned in pEGFP-Nl vectors (Clontech, Palo Alto, USA) in order to produce fusion protein with a C-terminal tag consisting of the enhanced green fluorescent protein (EGFP, Shimomura et al., 1961; Heim et al., 1995). For this the stop codon of each sequence was removed by PCR and replaced by a restriction site in frame with the EGFP sequence. The other EGFP constructs were received: vinculin-EGFP, from Prof. B. Geiger (Rehovot, Israel), paxillin-EGFP from Prof. A.F. Horwitz (Urbana, USA), zyxin-EGFP from Prof. J. Wheland (Braunschweig, Germany) connexin32-EGFP from Prof. W. Evans (Cardiff, UK), connexin43-EGFP from Dr L. Polontchouk (Zürich, Switzerland) and a-actinin-EGFP previously characterised in our laboratory (S. Meier unpublished data). 3.2.3 Localisation of transfected catenin proteins In an attempt to visualise the formation of adherens junctions in living cardiomyocytes, N- cadherin-GFP was previously used (Eppenberger and Zuppinger, 1999). In our hands, the expression of this particular protein was very weak, probably because its transport through the Golgi apparatus limits the rate of expression (data not shown). As we are interested in adherens junction proteins, we tested different proteins of the catenin family. Three catenins are known to interact with the transmembrane Cadherin in an ß-a-catenin or plakoglobin-a- catenin mutual exclusive manner to build up the internal side of the adherens junction. Moreover, ß-catenin and plakoglobin are part of the Wnt signalling and can be found in the cytoplasm and nucleus upon activation of the cells. The transfection of individual constructs in NRC shows that a-catenin specifically colocalises with Cadherins at cell-cell contacts (Figure 3.10.A-C). For ß-catenin and plakoglobin, most of the transfected cells show a cytoplasmic and nuclear localisation of the construct, with some residual cell-cell contact expression (Figure 3.10.D-F, J-L), whereas a minority of the cells display a localisation of the constructs restricted to the adherens junction (Figure 3.10.G-I, M-O). From this experience we can conclude that a-catenin localisation is more restricted to the ICD than ß-catenin and plakoglobin. This could be explained either by differences in expression level and/or by the dual role of ß-catenin and plakoglobin as signalling molecules and will be discussed later. As 59 Figure 3 10. Transfection of catenin proteins in rat cardiomyocytes. Confocal images of neonatal rat cardiomyocytes transiently transfected with a-catenin-GFP (A-C), ß-catenin- GFP (D-I) and plakoglobin-GFP (J-O). (A,D,G,J,M) are GFP signals. The cells were fixed and stained for a specific sarcomeric marker, (a-actinin in (B,E,H,K) and myomesin in (N)) and for a cell-cell contact marker, pan-cadherin (C,F,J,L,0). Instead of a-catenin- GFP which is efficiently directed to the cell membrane and colocalizes with Cadherins (C), ß-catenin-GFP and plakoglobin-GFP are found very frequently in the cytoplasm and in the nucleus (DJ). In some rare cases, the both proteins locate specifically at the cell- cell contacts (G,M). The bright signal observed (arrows in M,0) corresponds to a cell overexpressing the transfected plasmid. Bar, 10pm. we need a construct, which correctly targets and labels the adherens junction in living cells, a-catenin is from this test experiment the candidate of choice. 3.2.4 Localisation of transfected gap junction proteins Connexins exist in many flavours in the organism and up to three connexins are expressed in ventricular cardiomyocytes: cx40, cx43 and cx45. Unfortunately many antibodies commercially available, which recognise specific isoforms, are poorly reacting in immunoflurescence. In transfection experiments, we decided to test the localisation of three connexin constructs: cx32 (Figure 3.1 l.A-C) which is not expressed in heart cells but in liver cells, cx40 (Figure 3.11.D-F), which is expressed in embryonic cardiomyocytes and cx43 (Figure 3.12.G-I), the ventricular adult connexin isoform. In all cases, connexins are accumulated in perinuclear regions and small cytoplasmic vesicles (Figure 3.11.A,D,G), which is likely representing the trafficking of the protein from the endoplasmic reticulum to the golgi apparatus and finally to the cell-cell contacts. Cx43 has a prominent localisation at cell-cell contacts (Figure 3.11.G), which is almost invisible with cx40 (Figure 3.11.D) and seems inexistent with cx32 (Figure 3.11.A). As many studies on the assembly of intercalated disc in vitro mention that adherens junction formation is a prerequisite for the formation of gap junction, this could mean that there is a cooperative interaction between the cadherin- catenin complex and the formation of connexin 43-based gap junctions. 3.2.5 Localisation of transfected focal adhesion proteins Vinculin, paxillin and zyxin belong to the integrin-based cell-to-matrix contacts. It was therefore interesting to look at the localisation of these different constructs in cardiomyocytes and compare them with the localisation of adherens junction proteins. As expected, the three proteins localise at stress-fibre-like structures, as shown in Figure 3.12.A, and at Z-disc as demonstrated for zyxin in Figure 3.12.G-H but do not concentrate at cadherin-mediated cell- cell contacts (Figure 3.12.D,F). All together, this illustrates well the different adaptations created by the cell cultivation: on one hand, the formation of stress-fibres-like structures creates an extension of the cytoplasm devoid of sarcomeres, which has no counterpart in vivo (arrows in Figure 3.12.A-B). One the second hand the cellular polarity is lost and many of the new cell-cell contacts are spreading on the "lateral" side of myofibrils (arrows in Figure 3.12.D,F), reminiscent of the lateral contacts found earlier in the embryonic cardiomyocytes. As the re-establishment of cell-cell contact in cell culture produces many artefacts, it is wise to use whole mount preparations instead of cultivated cells for experiments regarding the integration of myofibrils. 60 Figure 3.11. Transfection of connexin isoforms in rat cardiomyocytes. Confocal images of neonatal rat cardiomyocytes transient transfected with connexin32-GFP (A-C), connexin40- GFP (D-F) and connexin43-GFP (G-I). (A,D,G) are GFP signals. The cells were fixed and stained for a specific sarcomeric marker, (myomesin in (B) and a-actinin in (E,II)) and for a cell-cell contact marker, pan-cadherin (C,F,I). All three proteins show distinct punctuation inside the cytoplasm which likely represent the trafficking of the proteins in the golgi apparatus and in the sarcoplasmic reticulum (arrowheads in A,D,G). From the three isoforms, only connexin43-GFP is efficiently integrated at sites of cell-cell contact (arrows in G and I). Bar, lOum. Figure 3.12. Transfection of constructs expressing focal adhesion proteins in rat cardiomyocytes. Confocal images of neonatal rat cardiomyocytes transient transfected with vinculin-GFP (A-C), paxillin-GFP (D-F) and zyxin-GFP (G-I). (A,D,G,) are GFP signals. The cells were fixed and stained for a specific sarcomeric marker, (myomesin in (B,H) and a-actinin in (E)) and for a cell-cell contact marker, pan-cadherin (C,F,I). All three proteins concentrate at Z-disc (arrowheads in G and H) and in stress fibre-like structures (arrows in A,B) but not that prominently at cell-cell contacts (arrows in D,F). Bar, 10 urn. 3.2.6 Expression of red fluorescent constructs The purpose of this transfection experiment was to determine the influence of the new red fluorescent tags on the localisation of sarcomeric protein. As previously demonstrated, the GFP tag does not alter the Z-disc targeting of the protein and serves here as reference (S. Meier unpublished data, see as well Figure 3.9 and Figure 3.13.A-B). The test of the red fluorescent tags was made by subcloning a-actinin cDNA in pDsRedl (Clontech, see also Datwyler et al., 2001), and in pmRFPl (Campbell et al, 2002). The transfection in NRCs of the a-actinin-DsRedl was successful in term of fluorescent brightness, even if the maturation of the protein was considerably slower than EGFP (data not shown). A major drawback for us was the discovery of sarcomeric structures modifications, with irregularities in Z-disc spacing and thickness (Figure 3.13.D-E). This aggregation problem with DsRedl fusion proteins was even more dramatic in stable transfected myoblast cells after differentiation (Figure 3.13.D'- E' compared with Figure 3.13.A'-B'). As the problems associated with DsRedl expression are very likely linked to its obligatory tetrameric structure (Baird et al, 2000), we decided to subclone a-actinin in a mutated monomeric version of DsRedl, named RFP (Campbell et al, 2002) to circumvent the aggregation problem. Although the construct is significantly less bright than DsRedl due to a lower quantum fluorescent yield and a faster photobleaching, RFP retains the Z-disc integrity in transfected NRCs (Figure 3.13.G-H) and seems to be a promising candidate for future fusion protein experiments. 3.2.7 Localisation of transfected bicistronic constructs After showing the correct localisation of the a-catenin construct (Figure 3.10.A) for the visualisation of adherens junctions and the correct localisation of both a-actinin-GFP and RFP for the labelling of sarcomeres (Figure 3.13.A,H), the last step is now to express both fusion proteins together in cardiac cells. For this, we tested first the possibility of using bicistronic vectors for the expression of two fusion proteins. The basis of bicistronic expression is the internal ribosome entry site (1RES) element discovered in Picornaviridae but common in many viruses like the Encephalomyocarditis virus (Jang et al, 1990). The 1RES element placed between two coding sequences allows a cap independent translation of the downstream gene. We tested the properties of the 1RES element by insertion of different sequences coding for fusion proteins (EGFP and mRFP tags). In the first cloning, MLC3f- EGFP and a-actinin-RFP were cloned on both sides of 1RES and both protein target to their correct location, i.e A-band for MLC3f and Z-disc for a-actinin (Figure 3.14.A-C). The same is true for a-actinin-EGFP and a-catenin-RFP inserted 5' and 3' of the 1RES. As before, a- 61 Figure 5.13. Transfection of a-actinin cDNA tagged with different fiuorophores in rat cardiomyocytes and mouse C2C12 myoblasts. Confocal images of neonatal rat cardiomyocytes transient transfected with a-actinin-GFP (A-C), a-actinin-DsRedl (D-F), a-aclinin-RFP (G-I) and C2C12 stable transfected with a-actinin-GFP (A',B') and a- actinin-DsRedl (D',E'). (A,A',E,E',H) are GFP/DsRed/RFP signals. The cells were fixed and stained for a specific M-band marker, (myomesin in B,B',D,D',G) and for a cell-cell contact marker, pan-cadherin (C,F,I). All three proteins localize at the Z-disc (arrowheads in A,B,D,E,G,H) but the a-actinin tagged with dsRedl shows irregularities in the Z-disc striation (E) and strong aggregation in stable transfected myoblasts (E'), very likely due to the tetrameric conformation of DsRedl. Different point mutations made in DsRedl lead to the so-called RFP lag. This fluorescent protein has a monomeric conformation which restores the regularities of the sarcomeric striations (H). Bar, 10pm. Figure 3.14. Transfection of bicistronic constructs (1RES) in rat cardiomyocytes. Confocal images of neonatal rat cardiomyocytes transient transfected with MLC3f-GFP-IRES-a-actinin- RFP (A-C), a-actinin-GFP-IRES-a-catenin-RFP (D-F). (A,D) are GFP signals, (B,E) are RFP signals. The cells were fixed and stained for a specific M-band marker, (myomesin in C,F). The protein expressed with the 1RES vectors are correctly located: A-band for MLC3f-GFP (arrowhead in A), Z-disc for a-actinin-RFP and a-actinin-GFP (arrowhead in B,D), cell-cell contact for a-catenin-RFP (arrows in E). We observed that the expression of proteins is usually stronger when cloned upstream of the 1RES element compared with a downstream location. Bar, 10pm. actinin localises at the Z-disc whereas a-catenin is found as expected at cell-cell contacts (Figure 3.14.D-F). From these preliminary experiments, we can see that the use of bicistronic vectors to follow the integration of myofibrils in living cells, like for example in embryoid bodies or in heart expiants, is technically possible, and depends now only on reliable techniques for efficient plasmid transfection. 3.3 Analysis of the ß-catenin conditional knockout 3.3.1 Outline of the project After the study of ICD assembly during embryonic development, the consequences of the deletion of an ICD protein from the heart will be tested in vivo, ß-catenin was chosen because it has on one hand a structural role in the constitution of the ICD but serves also an important function in the signalling network of cells (see paragraphs 1.2.5 and 1.3.1 in the Introduction). 3.3.2 Generation of a heart specific deletion of ß-catenin The general deletion of ß-catenin in the mouse germ line leads to a lethal phenotype during early development (Haegel et al, 1995) and thus prevents any studies of ß-catenin function also in heart. To avoid this limitation we used the Cre-lox technology to inactivate the ß- catenin gene in ventricular cardiomyocytes specifically. Homozygous ß-cateninflox/flox (Brault et al, 2001) mice were mated with the MLC2v-Cre heterozygous strain (Chen et al, 1998a). ° °* Cre positive male offspring (Fl generation) were backcross with ß-catenin females in order to obtain (MLC2vCre/+; ß-cateninflox/flox) male progenitors. Further backcrossing with ß- cateninflox/flox females maintained ß-catenin gene in a homozygous floxed situation and simplified the subsequent genotyping procedures. Offspring of these final breedings were re analyzed at birth and pups exhibited the expected Mendelian frequencies of MLC2v +, referred further as ß-catenin cKO, and MLC2v+/+ "wild-type" (Table 3.2). At this time, the animals did not display any obvious phenotype. The ratio of pups with a deletion of ß-catenin and the corresponding wild type animals was equal at birth but in 3 weeks old animals at weaning there were fewer cKO female while the fraction of the male population with the same genotype remained constant (Table 3.2). After weaning (PD21), there was no further change in the ratios of the different populations, and surprisingly, there was no obvious disadvantage of mice with a deleted ß-catenin gene. 62 Table 3.2. Offsprings of crosses between cfMLC2v Cre/wt ß-catenin flox/flox and 9 ß-catenin flox/flox Wild Type ß-catenin cKO (MLC2v wt/wt) (MLC2v Cre/wt) cf ç d ç PO 21 20 20 19 % of total 26.25 25.00 25.00 23.75 P21 48 45 44 30 % of total 28.74 26.95 26.35 17.96 The genetic background of the mice is C57BI6 3.3.3 Specificity and efficiency of the Cre mediated recombination The MLC2v-Cre knock-in mouse has been analyzed before and a rather specific deletion frequency has been observed in ventricular cardiomyocytes (Chen et al, 1998a; Hirota et al, 1999). The extent and specificity of the Cre-LoxP recombination in adult mice was determined by PCR analysis. We isolated genomic DNA from various organs subjected it to PCR analysis in order to detect the presence or the absence of the functional ß-catenin gene. With the choice of the appropriate pairs of primers the deletion of the gene was easily detected by the presence of a 500 bp fragment as shown in Figure 3.15.A-C (arrows). The most efficient recombination was observed in ventricular tissue, while none could be traced in atrial cardiomyocytes (Figure 3.15.A). Several other organs were tested and only in slow skeletal muscle were traces of gene recombination found, while in all other tested organs no gene rearrangement could be traced. These results demonstrate that the expression of the Cre- recombinase is highly specific for ventricular muscle with the only exception of slow skeletal muscle where the MLC2v promoter is also active as could be expected from published information (Chen et al, 1998a; Shai et al, 2002). In Figure 3.15.B-C, we tested with the same pair of primers the deletion of the ß-catenin floxed allele in neonatal heart. The genomic amplification of deleted ß-catenin (arrows) was performed together with a pair of primers, which amplify a sequence of 366 nucleotides in the connexin gene (arrowheads) to demonstrate the equality of genomic DNA used in the PCR. The amplification, observed at 25 cycles (Figure 3.15.B) or 30 cycles (Figure 3.15.C), is about the same in male and female neonates (columns labelled M Cre F/F compared with F Cre F/F) but much lower than neonates which are bearing one constitutively deleted allele and one submitted to Cre recombinase (columns labelled Cre F/d). Negative controls are also showed in this experiment (columns labelled wt F/F). From this experiment, we conclude that the recombination is detectable at birth but occurs at low levels. 3.3.4 Deletion of ß-catenin through postnatal development In order to verify if gene inactivation by recombination was also accompanied by loss of the corresponding ß-catenin protein in heart tissue, cryosections from hearts obtained from mice at day 3 after birth, 1 month, 2 months, 6 months and 15 months of age were analyzed by fluorescence microscopy. The samples were immunologically stained with antibodies against ß-catenin, against the ECM protein laminin, or with fluorescent phalloidin for myofibrillar F- actin. The confocal micrographs are compiled in Figure 3.16. The three columns in the left 63 Cre/wt Cre Cre Cre wt Cre Cre Cre Cre wt Cre ß-cat F/F F/F F/F F/F F/d F/F F/F F/F F/F F/d Figure 3.15. The deletion of ß-catenin is heart specific and is quite low at neonatal stage. Primers which specifically recognize the deleted allele (arrow in A) were used in PCR reaction on 100 ng of genomic DNA isolated from the indicated tissue of 6 months old mice. The main amplification occurs in the heart ventricle with a slight amplification observed in the slow skeletal muscle extract. In order to show the recombination in neonates, primers which specifically recognize the deleted allele (arrows) and the connexin 43 genomic sequence as control (arrowheads) were used in the PCR reaction with 25 cycles (B) or 30 cycles (C) on 100 ng of genomic DNA isolated from neonatal ventricular tissue of the specified genotype. When compared with pups carrying one constitutively floxed-out gene (Cre; F/d), the recombination in Cre F/F is significantly lower. Figure 3 16. The deletion of ß-catenin at the protein level is observed only in postnatal stages. Confocal images of cryosections of WT (A-O) and ß-catenin cKO (A'-O') immunostained for laminin (A,D,G,J,M,A',D',G',J',M'), ß-catenin (B,E,H,K,N,B\E',H',K',N') and phalloidin (C,F,I,L,0,C',F',r,L',0') at postnatal day 3 (A-C), 1 month (D-F'), 2 months (G-F), 6 months (J-L') and 15 months (M-O').The decrease of ß-catenin signal is clear at 2 months (compare IV with H). From 6 months cKO cells look bigger (compare L' with L and O' with O). Bars, 10pm (A-C) 20pm (D-O'). half of the Figure 3.16 represent sections from wild type mouse hearts, the conditional ß- catenin KO heart samples are compiled in the 3 columns to the right. Quite surprisingly there is no big difference in the ß-catenin pattern in ventricular tissue at day 3 and 1 month as shown in Figure 3.16.B,E and the corresponding mutant in Figure 3.16.B',E'. In both tissues ß-catenin protein is associated with cell borders and is not yet completely confined to the intercalated disc (ICD). In the control samples of 2 months old mice the concentration of ß-catenin to the ICD has progressed and the majority of the antigen is now in stripes perpendicular to the long axis of the cardiomyocytes (Figure 3.16.H') a tendency becoming more pronounced with increasing age (Figure 3.16.K',N') as had been investigated earlier (Angst et al, 1997). For the explanation of the unexpectedly slow disappearance of the ß-catenin protein from the ICD we propose several hypotheses. On one hand, it is possible that the protein is stably incorporated into ICD and is only slowly released from these sites because it may be protected from the proteasomal degradation machinery. On the other hand the excision of the floxed ß- catenin gene may be a sluggish process due to uneven expression of the Cre-recombinase. These possibilities will be discussed below. The ß-catenin protein is lost from samples older than two months as shown in panels H',K',N' of Figure 3.16., leaving occasional traces between cells but which did not co- localise at the ICD. It is very likely that these traces are associated with non-myocytes, which did not undergo the gene rearrangement. With respect to the arrangement of the ECM protein laminin no significant difference was observed between the cKO and wild type cells, but the tendency of increasing space taken up by the ECM is visible. The phalloidin staining in the right column of each series is also similar in both wild type and cKO but there are distinct tendencies in the cKO column with greater cell width and a less homogenous myofibrillar pattern with many more discontinuities (Figure 3.16J',L',0'). The shape of the myocytes is subject to analysis in a paragraph below. 3.3.5 Deletion of ß-catenin at neonatal stage in culture As already suggested by the Figure 3.16.B', the deletion of ß-catenin in neonatal stages at the protein level is expected to be rather low even if the Cre-recombinase activity driven by the MLC2v promoter is supposed to be turned on quite early (Lyons et al, 1995). Nevertheless we wanted to know if the absence of ß-catenin at the adherens junction would lead to any alteration of the cell-cell contacts or defects in the anchoring of myofibrils to the membrane in vitro. For this reason, we isolated cardiomyocytes from both WT and cKO newborn animals 64 and placed them in culture for 2 days then used immunofluorescence for staining nuclei, myofibrilar F-actin, cell-cell contact proteins Cadherins and ß-catenin (Figure 3.17). The observation of the recombined cells, which display no or very weak ß-catenin signal (Figure 3.17.E-H') lead us to the conclusion that the deletion of ß-catenin was not showing any striking difference in term of maintenance of cell-cell contact or integration of the myofibrils to the membrane in neonatal cultured cardiomyocytes. 3.3.6 Regulation of other intercalated disc proteins in the absence of ß-catenin The intercalated disc is an assembly of several proteins and the change in the composition of the adherens junction can be expected to lead to alterations of the ICD structure and/or quantity of the remaining components (Ehler et al, 2001; Ferreira-Cornwell et al, 2002). The integrity of the whole structure was probed on cryosections of adult hearts (from 6 months old mice) by antibodies against critical components of the ICD, i.e. antibodies against Cadherins (adherens junctions), desmoplakin (revealing the desmosomal part of the ICD) and connexin 43 as a marker of gap junctions, in double stains using always antibodies against ß-catenin as control to show the disappearence of the protein in cKO hearts. The staining with the antibody against ß-catenin as shown in Figure 3.18, demonstrates little variability in the WT samples (compare panels of Figure 3.18.B,F,J) and is used as internal control. Likewise, the ß-catenin colocalises with the staining of Cadherins (compare Figure 3.18.A with 3.18.B), desmoplakin (Figure 3.18.E compared with 3.18.F) and connexin 43 (Figure 3.18.1 with 3.18.J). Although these are not very high-resolution micrographs, the co- localisation is obvious. When cryosections of 6 months old cKO were analysed in the same fashion, the expected reduction of ß-catenin was noted in all samples tested (Figure 3.18.D,H,L) but some signal not related to the ICD was observed, possibly due to a non- myocyte expression of the protein. The appearance of the ICD distribution of the other antigens was however not changed significantly, although there appears to be generally more antigen in cKO samples stained for Cadherins (Figure 3.18.C compared to 3.18.A), and desmoplakin (Figure 3.18.G compared to 18.E) while there was no significant increase in the staining of connexin 43 (Figure 3.18.K compared with 3.18.1). The apparent quantitative changes seemed to be related to junction proteins like Cadherins and desmoplakin but possibly not to gap junction components. To verify the hypothesis that ß- catenin deficiency might be compensated by increases of other ICD components were tested by immunoblots of extracts from hearts of siblings WT and cKO mice. To increase the significance of the analysis 3 samples were probed from 3 individual mice at 2 and 6 months 65 Figure 3.17. The deletion of ß-catenin at the protein level at neonatal stage is hardly visible but the cells which do not express ß-catenin have normal cell-cell contacts. Neonatal ventricular cardiomyocytes were isolated from wild type (A-D) or ß-catenin cKO litters (E-H, E'-H' at higher magnification) and stained with DAPI (A,E,E'), phalloidin (B,F,F') ß-catenin (C,G,G') and Cadherins (D,H,H'). Compared to wild-type, cKO cardiomyocytes look similar and cell-cell contacts with deleted ß-catenin are difficult to find (see arrowheads in F,G,H). The anchoring of sarcomeres to the membrane does not seem to be affected by the deletion (arrowheads in F'-H'). Bars, 10pm (A-H) and 6pm (E'-H'). Figure 3.18. The deletion of ß-catenin does not affect intercalated disc organization. Confocal images of cryosections of WT (A,B,E,F,1,J) and cKO (C,D,G,H,K,L) immunostained for Cadherins (A,C), ß-catenin (B,D,F,H,J,L) desmoplakin (E,G) and connexin43 (I,K). Note the pronounced staining of Cadherins in cKO (compare A with C) and some residual intercalated disc staining of ß-catenin (arrow in G,H). In general, the ß-catenin signal not associated with the intercalated disc is more prominent in cKO than in WT (arrowheads in D,H,L compared with B,F) Bar, 20pm. of age of either WT or cKO animals in which the ß-catenin gene is progressively lost and the results are shown in Figure 3.19.A. The deletion of ß-catenin is not complete also at the level of the accumulated protein in 2 months old hearts (lanes labelled 2 months cKO). In the samples taken at 6 months only traces remain (lanes 6 months cKO). These results clearly document that ß-catenin gene deletion may be a sluggish process or that the ß-catenin protein turnover at the ICD is very slow so that the removal from the cadherin/catenin complex takes more than two months. For quantification, the integral of the optical density of the band on immunoblots from the ICD component was standardised against the number derived from the corresponding band generated by a general actin antibody probe. In order to allow direct comparison of the blotted results, the ratio of the two numbers from cKO hearts were compared with the ratio of the same numbers in WT hearts. The values in percent of wild type accumulation were listed in the bar diagram in Figure 3.19.B. The decrease of ß-catenin is remarkable at 2 month and in older hearts (75% of ß-catenin is removed at 6 months) but some protein still remains, very likely expressed from non-myocytes, which are not affected by the MLC2v-driven Cre recombinase. The other antigens which were explored can be clustered into two groups depending on their expression level at 6 months: antigens with increased expression in ß-catenin deleted heart and antigens without any apparent change regardless of the deletion of the ß-catenin gene (cKO) or its presence (WT). In the group of ICD proteins increased in cKO hearts, we find the direct binding partners of ß-catenin, namely ccE-catenin (+37%) and Cadherins (46%), the close-related protein plakoglobin (+27%) and to a lower extent, the desmosomal protein desmoglein-2 (+14%). This increase in relative quantities, which is probably due to a compensatory mechanism like the expression of ß-catenin in the desmosomes of plakoglobin -/- cells (Bierkamp et al, 1999), takes some time, because the values after two months are not significantly increased for three of them. Quite surprisingly aT-catenin and vinculin remain constant at 6 months and the significance of this phenomenon is not clear. The same constancy is true for the gap junction protein connexin 43 and is of great importance because most myocardial diseases are marked by the alterated level of connexin 43 expression, e.g. down-regulation in mouse models of dilated cardiomyopathy (Ehler et al, 2001) and in hypertrophic or ischemic animal models (Wang and Gerdes, 1999) and in human heart (Peters etal, 1993). 66 desmoglein-2 Cadherins vinculin aT-catenin ttE-catenin ß-catenin plakoglobin connexin43 actin 2 months 6 months 200% B J 1 I ^ ^ j——^ ! j ß-cat aE-cat pkg cadhs aT-cat vin dsg-2 cx43 Figure 3.19. The deletion of ß-catenin increases specifically the expression of some other components of adherens junction. (A) Immunoblots of SDS-samples show the decrease of ß- catenin over time which is compensated by the expression of the other members of the cadherin/catenin complex, i.e ocE-catenin, plakoglobin and Cadherins whereas vinculin and aT- catcnin levels are roughly unchanged at 6 months, as desmoglein-2 and connexin43. (B) Densitometric analysis of immunoblots normalized for actin expression. The values are indicated for cKO samples in comparison to WT (100%). In cKO animals, ß-catenin reaches 25% of WT at 6 months whereas aE-catenin, plakoglobin and Cadherins increase ( f37,+26 and +46% of WT). 3.3.7 Possible reasons for the long survival of ß-catenin in ICD. As noted above the long persistence of ß-catenin was astonishing and might be caused by several mechanisms. As speculated before, the expression of the Cre-recombinase was not homogeneous and delayed in certain cells, although it had been shown in other experiments that the promoter MLC2v used in this construct to drive the Cre-recombinase was sufficiently active in cardiomyocytes to induce the genetic rearrangement in most cells. A cell-specific assay for MLC-2v driven Cre activity was therefore necessary to learn more about specific gene inactivation in cardiomyocytes. The activity of the Cre-gene product was tested in combination with the ROSA26 reporter gene, which displays ß-galactosidase activity when an inhibitory segment is deleted by the Cre-recombinase. The ß-galactosidase activity can then be probed by a colorigenic substate forming blue precipitates at the site of active cells demonstrating that the excision has taken place. In cultures from neonatal mice (Figure 3.20.A-A"), it was observed that the size of nuclei in cardiomyocytes was smaller as compared to non-myocytes (arrow in Figure 3.20.A). This characteristic feature was used successfully for the discrimination of the myocytes and non-myocytes with a probability of mistake of a mere 2% (Table 3.3.A), which was further used to calculate the percentage of cardiomyocytes with rearranged genes due to active Cre-recombinase. The genomic recombination was taking place in the order of 47% (Figure 3.20.A'-A" and Table 3.3.B). A similar result was obtained from the analysis of cryosections made from 1 months and 2 months old hearts. In both cases there were significant numbers of rearranged cells visible by the blue stain from the ß-galactosidase activity (Figure 3.20.B-C). Of course, one has to bear in mind that the majority of cells consist of non-myocytes where no rearrangement has taken place and thus the percentage cannot exceed the relative number of cardiomyocytes. These numbers are only partially in agreement with biochemical data discussed in Figure 3.19 and the immunohistochemical data in Figures 3.16 and 3.17, since even in 1 month old heart tissue the ß-catenin content has barely deceased (Figure 3.19.A) while already in neonatal cardiomyocytes about half of the cells contained a rearranged genome. An alternative mechanism might be that the ß-catenin protein once stably incorporated into ICD has an extended half-life, in contrast to the general assumption that it is turned over quite rapidly. The mobilisation of the ß-catenin protein from the ICD might also be dependent on the accessibility of kinases and or phosphatases to the ICD (Nelson and Nusse, 2004). 67 Figure 3.20. The MLC2v-cre-mediated deletion is restricted to cardiomyocytes but the efficiency is low at birth and before 2 month postnatal. MLC2v-Cre mice were breeded with Rosa26R reporter mice which express ß-galactosidase after recombination. At neonatal stages (A,A',A") isolation of cardiac cells where submitted to DAPI (blue) and a-sarcomeric actinin (red) staining (A) or DAPI (blue) and X-Gal (green) staining (A',A").Note that cells with bigger nuclei (arrow in A, arrowheads in A', A") are non-cardiomyocytes (see Table 3.3.A) and are not affected by the cre recombination. The percentage of cardiomyocytes with positive X-gal staining is less than 50% (see Table 3.3.B). The percentage of recombination continues to increase after birth, as shown in X-Gal staining of left ventricle transverse cryosection at 1 month old (B) and 2 months old(C). Bars, 20pm (A-A") and 500pm (B-C). Table 3.3 Recombination at neonatal stage A) Validation of the size of the nucleus* as selection criteria for the distinction between cardiomyocytes and non-cardiomyocytes using cc-sarcomeric actinin as reference. cardiomyocytes cardiQn^cytes Morphology of the nucleus 105 102 a-sarcomeric actinin staining 103 2 Error (%) L90 1.96 B) Estimation of recombination in cardiac cells. Cardiomyocytes are distinguished from non-cardiomyocytes according to the morphology of the nucleus and X-Gal staining recognizes cre- mediated recombined cells. non- _ A cardiomyocytes cardiomyocytes Morphology of the nucleus 310 226 X-Gal staining 146 6 Recombination (%) 47.10 2.65 *c.f figure 3.20 The counting A was made with isolated cells from WT pups and the counting B with 4 MLC2v Cre/wt : Rosa26R +/wt pups. 3.3.8 Is there a hypertrophic response in conditional ß-catenin KO hearts? Deletion of ß-catenin specifically in heart is thought to interfere with the integrity of the ICD and thus also with mechanical coupling. In analogy to the deletion of MLP from mice leading to a dilated cardiomyopathy (Arber et al, 1997), a similar defect was presumed for the ß- catenin KO to lead also to pathological alterations. As a first parameter, the cardiac index (heart weight to body weight ratio) of male mice was determined. As shown in Figure 3.21.A there is no significant increase of the cardiac index visible and even at 15 months of age there is no indication of a hypertrophy at the level of the whole heart tissue. A more detailed analysis was carried out at the level of RNA expression assayed by RT- PCR on several hypertrophy markers and again no significant upregulation was observed as shown in Figure 3.2LB. As a positive control RNA from MLP-KO mice was reverse transcribed and amplified in the presence of primers, which allow the PCR amplification of a set of hypertrophy markers as demonstrated in the MPL-KO lane 7, which shows all these molecules to be increased in this semi-quantitative assay. In the lanes 1-3 three wild type mice with the same genetic background were compared to 3 cKO individuals (lanes 4-6). The cKO mice display clearly the Cre-recombinase RNA (Figure 3.2LB lanes 4-6), but no hypertrophy marker was increased including N-RAP, which was shown to be increased as the earliest protein marker in MLP KO mice (Ehler et al, 2001). In conclusion, the deletion of the intercalated disc protein ß-catenin does not entail the molecular changes normally associated with cardiomyopathies, and there is no apparent response to the lack of ß-catenin in the intercalated disc in the differentiated heart. There were however hints that there might be minor adaptations as indicated by the slightly increased cell width and less continuous staining with phalloidin in the sections shown in Figure 3.16.I',L',0'. The slightly, but not significantly increased heart weight ratio observed in older mice (Figure 3.2LA), prompted a more detailed comparison of single adult cells isolated from hearts of 6 month old wt (Figure 3.22.A,C,E) and cKO (Figure 3.22.B,D,F) male mice. Optical sections of confocal microscopy images are shown in Figure 3.22.A,B stained with the thick filament associated protein MyBP-C, with a-cardiac actin in Figure 3.22.C,D and with desmin in Figure 3.22.E,F. The ß-catenin protein analysed in all specimens showed as expected, the complete absence of the protein in the mutant cells (Figure 3.22.B,D,F). The sections in Figure 3.22.B,D,F but also the three dimensional reconstruction of the cell shape (Figure 3.22.G) show a more branched and frayed structure of the intercalated discs with smaller but more numerous cell-cell contacts but the sarcomeric 68 1 month 2 months 6 months 15 months MLP cKO -RT WT ß-cat KO Figure 3.21. The deletion of ß-catenin does not promote a hypertrophic on at least 6 response. (A) Heart weight to body weight ratio was performed and is not male mice per group (open bars: controls, black bars: cKO) significantly changed in ß-catenin cKO mice. (B) RT-PCR analysis. Total RNA (lug) from 6 months old male mice was reverse-transcribed and cDNA No was exponentially amplified in the presence of gene specific primers. obvious difference can be shown in the level of the classical hypertrophy markers (ANF, BNP, a-skeletal actin, ß-myosin heavy chain) and in the level of N-RAP a gene known to be upregulated in MLP KO. In this experiment, a MLP KO of the matching sex and age was used as positive control, as well as a sample without reverse transcriptase (-RT). WT (3-cat cKO 23050+/-3000 um3 31010+/-5100um3 figure ? 22 cKO cells at 6 months have a change in morphology and volume but the sarcomere organization îemains intact Confocal section of treshly isolated cells from WT (A,C,E) and cKO (B,D,F) were stained with MyBP-C (A,B red chanel) a-cardiac actin (C D red chanel) desmin (r Y red chanel) as well as with ß-catcnm (A-r green chanel) (G) C omputer based 3 D reconstruction on W I (left) and cKO (right) using ß-catenin (green chanel) and phalloidin staining (red channel) shows the morphology ol the cells From this image a volume measurement can be generated (H) Statistical analysis ot the volume measurement of isolated cells performed for 5 Wf and 4 cKO animals At least 30 cells were analysed m each isolation The striated pattern of sarcomeres is not disturbed by the absence of ß-catenin but the morphology (compare F with b) and the volume (see G and II) are modified Bars 20u,m Table 3.4. Cell volume measurement at 6 months Wild Type ß-catenin cKO Length, pm 104.9+/-17.7 109.6+A9.2 Width, pm 26.6+/-2.9 30.0+/-4.4 Length/width 4 3+/1 Q 4.0+/-1.0 ratio Depth, pm 12.0+/-3.2 12.5+/-2.7 Volume pm3 23'050+/-3000 31'010+/-5100* * Values are mean +/-SD. P<0.05 Data were obtained from 5 wild-type and 4 ß-catenin cKO male mice. At least 30 cells were analysed per animal pattern appears unchanged. The cell volumes were determined as indicated in materials in methods and showed an increased mean volume in the mutant cells (Figure 3.22.H). The numerical values are compiled in Table 3.4 and show that the mutant cardiomyocytes have a 34% volume increase, which is due to a expansion in the three dimension (length, width and depth). As mentioned before, the heart weight body weight ratio was not significantly altered and with bigger cardiomyocytes a reduced number of cells is expected so that the volume increase might be a compensation for the loss of cells by apoptosis or the consequence of an earlier stop in cardiomyocyte proliferation. 3.3.9 There is no significant increase of myocyte death in conditional ß-catenin KO hearts. Cryosections of mutant and wild type hearts were analysed by TUNEL assays, and are shown Figure 3.23. In the mutant cells (Figure 3.23.E) no increase of apoptosis was observed when compared to WT cells (Figure 3.23.A) and no apoptotic cells could be observed in a great number of sections (not shown). The positive controls (Figure 3.23.C,G) made by nuclease treatment of control sections clearly show that the TUNEL assay worked. This finding does not indicate that there is absolutely no loss of cells due to apoptosis. The process could be very slow and the number of cells lost per time unit too small to be detected in these heart tissue sections. It is however unlikely that there is a major apoptotic process leading to the loss of cells which has been observed in other cardiac degenerative conditions. Here, we speculate that the hypertrophic response is indeed very mild and the adaptation is possibly restricted to a very early compensation phase not leading to pathological changes in gene activity and cell morphology. 3.3.10 Physiological parameters are not significantly altered in basal conditions. The sex ratio of surviving animals was changed in the population of three weeks old animals, where a deficit of mutant females was visible (Table 3.2) and thus a physiological impact of the mutant was suspected. So far, only static parameters of the heart have been investigated and determinations of physiological measurements had to be made by echocardiography on living animals. The first experiment we made was an hypertension-induced hypertrophy with the one-kidney-one-clip (1K1C) model. Two groups of 6 months old male mice were used and echocardiographic parameters were measured before and after the induction of hypertrophy (Table 3.5). There are no significant alterations of any of the parameters tested between the mutant hearts and wild type hearts under basal conditions (compare columns labelled WT and ß-catenin cKO). These measurements indicate that the loss of ß-catenin can be compensated by other proteins and does not seem to induce alterations that lead to 69 Figure 3.23. 'I he deletion of ß-catenin does not promote apoptosis. TUNEL assay was performed on WT (A, C) and cK.0 (E,G) cryosection of 2months old mice Positive control (DNAse treatment) is shown (C and G) as well as phase contrast (B,D,F,H). This experiment was repeated on three different heart of WT and ß-catenin cKO mice, examining first the whole section at low magnification. Bar, lOum, Table 3.5 Echocardiographic data with 1K1C experiment WT* ß-catenin WT after ß-catenin cKO cKO* 1K1C** after 1K1C** LVAWD, mm 1.58+/-0.09 1.53+/-0.09 1.90 1.58+/-0.14 LVAWS, mm 1.92+/-0.14 1.88+/-0.12 2.13 1.91+/-0.21 LVD, mm 3.06+/-0.20 3.04+/-0.27 3.00 3.09+/-0.31 LVS, mm 1.64+/-0.16 1.50+/-0.30 1.63 1.40+/-0.16 IVSD, mm 0.34+/-0.05 0.35+/-0.05 0.30 0.36+/-0.04 IVSS, mm 0.44+/-0.07 0.43+/-0.05 0.40 0.46+/-0.05 Fractional 46.4+/-4.0 50.9+/-6.5 46.0 54.7+/-1.8 shortening, % Cardiac index, mg 4.97+/-0.53 5.18+/-0.97 6.58 6.27+/0.73 heart/g body weight LVAWD: left ventricle anterior wall thickness at diastole LVAWS: left ventricle anterior wall thickness at systole LVD: left ventricle diastolic diameter LVS: left ventricle systolic diameter IVSD: intra-ventricular septum thickness at diastole IVSS: intra-ventricular septum thickness at systole Fractional shortening: [(LVD-LVS)/LVD]x100 Values are mean +/-SD. *Data were obtained from 6 wild type and 9 ß-catenin cKO male mice at 6 months of age. **Data obtained from 1 wild type and 5 ß-catenin cKO mice that survived 4 weeks after the 1K1C operation. For this reason, no statistical conclusions can be drawn from this type of hypertrophic induction. decompensation and pathological responses under physiological conditions. One has to notice here that the 1K1C operation is usually performed on 2-3 months old animals but we used mice at 6 months of age, where more ß-catenin is removed in the cKO animals (Figure 3.19.B, columns labelled ß-cat and Figure 3.16.K' compare with 3.16H'). It is very likely that at this age, the operated animals had an increased diameter of the renal artery, which was then too much tightened by the clipping operation and lead to the death of 60% of the animals. Because of the high mortality following the operation in both groups (columns WT after 1K1C and ß-catenin cKO after 1K1C), no conclusion can be drawn for the impact of ß- catenin deletion in hypertrophic conditions induced by the 1K1C procedure. 3.3.11 ß-catenin deletion improves fractional shortening in ß-adrenergic-induced hypertrophy It remains to be seen how the ß-catenin deficient heart copes with additional stresses known to result in cardio-pathological responses. For this, we decided to test the response of ß- catenin deleted heart to ß-adrenergic stimulation. As the implantation of mini osmotic pumps is usually performed at 2-3 months, we decided this time to keep this setting and used 9-10 weeks old animals although the deletion of ß-catenin is not maximal at this time. Half of WT and ß-catenin cKO animals were implanted with the Isoproterenol pumps and the second half was used to test basal conditions. After 2 weeks, contractility parameters were recorded using echocardiography and the results are shown in Figure 3.24 and in details in Table 3.6. The ß- adrenergic stimulation induces an increase in heart rate (compare the frequency of the sinusoidal curves in Figure 3.24.B with 3.24.A) and an increase of the cardiac index (Figure 3.24.C) compared to control groups. The fractional shortening (FS, see Figure 3.24.D) is also increased but quite surprisingly, the heart specific deletion of ß-catenin significantly amplifies this increase of contractility compare to operated WT. This finding is at this moment difficult to interpret because we do not know what this augmentation of contractility means for the heart. It could mean that the stiffness of the wall is reduced because the parameter which is significantly modified in cKO animals after ISO is the left ventricle diameter at systole (Table 3.6, fourth lane) Nevertheless, this experiment definitely confirms that the heart devoid of ventricular ß-catenin can cope quite well with hypertrophic stresses. 70 WTctrl cKOctrl WTISO cKOISO WTctrl cKOctrl WT1S0 cKOISO Figure 3 24 ß-catenin deletion increases the contractility of the heart induced by adrenergic stimulation. Typical echocardiographic measurements of Table 3.6 shown in basal conditions (A) and with ß-adrenergic stimulation by Isoproterenol (ISO) treatment (B). Note the increase heart rate due to ISO treatment (B, increased perio. Refer to 2.9 for the explanation of abbreviations. 3 months old cKO and WT with or without ISO treatment were analysed for heart weight body weight ratio (C) and fractional shortening as measure of cardiac performance (D). See also Table 3.6 for detail of. the cardiac index measured after the sacrifice of the animals is increased, as expected, in both groups treated with ISO compared to control groups. As the fraction shortening (FS) is significantly increased in cKO animals compared to WT (P=0.045) this experiment suggests an inhibitory function of ß-catenin on the catecholamine induced contractility which is removed in the cKO animals. Note also that Table 3.6 Echocardiographic data with Isoproterenol stimulation WT ctrl* ß-catenin cKO WT after ß-catenin cKO Ctrl* ISO* after ISO* LVAWD, mm 1.46+/-0.10 1.45+/-0.13 1.49+/-0.24 1.76+/-0.24" LVAWS, mm 1.68+/-0.06 1.77+/-0.23 2.08+/-0.14" 2.15+/-0.17** LVD, mm 3.07+/-0.11 3.04+/-0.17 3.04+/-0.53 2.76+/-0.18** LVS, mm 1.64+/-0.24 1.60+/-0.25 1.26+/-0.31** 0.92+/-0.21**t IVSD, mm 0.28+/-0.05 0.31+/-0.03 0.34+/-0.05** 0.31+/-0.05 IVSS, mm 0.35+/-0.05 0.39+/-0.07 0.43+/-0.06** 0.36+/-0.07 Fractional 46.8+/-6.3 47.4+/-7.2 58.6+/-7.0** 66.7+/-7.1*** shortening, % Cardiac index, mg 4.90+/-0.34 4.61+/-0.42 5.66+/0.24** 5.70+/0.35** heart/g body weight LVAWD: left ventricle anterior wall thickness at diastole LVAWS: left ventricle anterior wall thickness at systole LVD: left ventricle diastolic diameter LVS: left ventricle systolic diameter IVSD: intra-ventricular septum thickness at diastole IVSS: intra-ventricular septum thickness at systole Fractional shortening: [(LVD-LVS)/LVD]x100 Values are mean +/-SD. *n=8 mice (9-10 weeks old at the beginning of the experiment) in each group with the exception of ß-catenin after ISO where n=7. The echocardiographic parameters are recorded 2 weeks after the implantation (after ISO), together with Ctrl groups. ** P< 0.05 compared with not operated mice t- P< 0.05 compared with WT after ISO 3.4 Importance of ß-catenin in DCM heart 3.4.1 ß-catenin knockout in MLP knockout mice As shown in section 3.3, the heart specific knockout of ß-catenin is tolerated in the adult heart, likely due to the compensation by the other components of the cadherin-catenin complex (Figure 3.19.B). As stated before in the introduction, it is generally believed that heart diseases produced by mutation of cytoskeletal proteins induce dilated cardiomyopathy because of defects in force transmission. It would therefore be interesting to know how cardiomyocytes already depleted of ß-catenin would cope with further supplemental stress due to the deletion of another cytoskeletal component. For this ß-catenin cKO animals were mated with MLP-/- mice (Arber et al., 1997) in order to get, after backcrossing and genotyping analysis, the resulting MLC2vCre/+; ß-cateninflox/flox; MLP"7" mice and their MLC2v+/+; ß-cateninflox/flox; MLP"'" sibilings referred later as MLP*A controls. In this breeding the genetic background of the mice is no longer defined, as ß-catenin cKO are C57B1/6 inbreed and MLP-/- have a B6;129 mixed background referred in a publication as OBF-E52 (Ehler et al., 2001). Because of this genetic mix, we have to be careful in the interpretation of data and keep in mind the possible influence of genetic changes on observed phenotypes. As the MLP locus is on chromosome 7, 38Mbp away from the epistatic gene albino and as MLP- /- mice are indeed albinos, this particular configuration results in the F2 generation in an almost complete co-segregation of the MLP mutation together with the white coat colour, which simplifies the search for double KO animals. 3.4.2.Early postnatal lethality associated with double KO mice Breeding of ß-catenin cKO under MLP -/- background was first analysed at birth. Pups of the male gender exhibited the expected Mendelian frequencies but a small decrease of the female double knockout population was already visible (Figure 3.25.B, columns 1-4). At weaning age, both male and female double knockout were less numerous than expected but quite surprisingly the imbalance was much more severe for the female gender (Figure 3.25.B, columns 5-8). After this critical period, the small number of survival double knockout (around 30% of the expected progeny) progress to adulthood without additional sign of fatigue compared with MLP"A litter mates. When analysed in detail between PO and P5, the double knockout animals were usually smaller than their MLP"/_ control littermates (Figure 3.26.A) and their hearts were considerably enlarged (Figure 3.26.B-C). Thus we speculate that the mortality found in the 71 A) ß-catenin cKO breeding F F M M F F M M ß-cat wt cKO wt cKO wt cKO wt cKO PO P21 ß-catenin cKO-MLP KO breeding 40 36 F F M M F F M M ß-cat : wt cKO wt cKO wt cKO wt cKO PO P21 C) ß-catenin cKO-DRAL KO breeding 13 13 F F M M F F M M cKO ß-cat : wt cKO wt cKO wt cKO wt P21 Figure 3.25. The conditional deletion of ß-catenin in a MLP knock-out background or DRAL knock-out background shows a pronouced perinatal lethality. Compared to ß-catenin cKO alone (A) where we observe a slight decrease in the female population at weaning age, both double KO (B and C) -71% have a reduced population compaired to their control (-69% for MLP, for DRAL). Note that ß-catenin cKO-MLP KO has, like ß-catenin cKO alone, a preferential survival of male mice. hgiire 3 26 The conditional deletion of ß-catenin in a MLP knock-out background shows a perinatal lethality with reduction of body weight, increase in heart volume and myofibril breakdown. Aï P2, a significant number of pups with ß-catenin cKO MLP KO genotype shows growth retardation (A, right) compared with their MLP KO Iittermatcs (A,left) When isolated, the heart of those double KO (C) are enlarged compared to single MLP KO hearts (B). In these litters, the mortality is restricted to the double KO and occurs between PO and P5 (D, black line). Single confocal sections from MLP KO heart (h-G) and double KO (II-J) immunostained with picogreen (E,H), myomesin (F,I) and ß-catenin (G-J) We can observe in double KO a reduction in the number of nuclei (compare H with H) accompanied with a loss of cardiomyocytes (compare f with 1). Note that the lateral border of the damaged region shows a preferential lateral staining of ß-catenin very likely due to the expression of the protein in non-cardiomyoyctes (arrowheads in I,J). Bar, 20iun litters (Figure 3.26.D) is due to heart failure, reminiscent of the early phenotype described earlier when the first MLP-/- mice were generated (Arber et al., 1997). It is important to mention here that the penetrance of this early phenotype in the MLP knockout was incomplete and decreased generation after generation. The MLP-/- mice which are maintained in our animal facility and that we used for our double knockout experiment are devoid of this phenotype. Moreover, a strong confirmation that the early phenotype is not brought back by the mixing of genetic background is the almost complete absence of lethality in the MLP"7" control siblings (1 out of 42 pups observed, see as well Figure 3.26.D gray line). When looking at cryosections triple stained for the nuclear marker Picogreen, the sarcomeric M-band protein myomesin and ß-catenin, we found a profound disorganisation of the myocardium of the double knockout (Figure 3.26.H-J) compared to MLP-/- controls (Figure 3.26.E-G). The immunofluorescence pictures reveal an increase of nuclear size (Figure 3.26.H compared with 3.26.E) together with damaged myofibrillar structures (Figure 3.26.1 compared with 3.26.F) accompanied with a decreased accumulation of ß-catenin at the ICD (Figure 3.26J compared with 3.26.G) and most of the residual protein is found at the lateral membrane (Figure 3.26.J, arrowheads) very likely corresponding to the expression of ß- catenin by non-cardiomyocytes as seen before in adult tissue of ß-catenin cKO mice (Figure 3.16.H',K'). 3.4.3 Hypertrophic response at RNA level In order to get more insights into the molecular mechanism leading to the early lethal phenotype observed we tried to look at the expression of hypertrophy markers at the mRNA levels. For this, hearts from neonatal mice were dissected and total RNA was extracted from ventricles. RT-PCR analyses to detect levels of hypertrophy markers are shown on Figure 3.27.A. Pairs of primer specific for ANF and BNP showed upregulated levels in double knockout males and females (Figure 3.27.A, two first rows, column 3-6 and 10-12) whereas a-skeletal actin seems to be specifically upregulated in female double knockouts (Figure 3.27.A, fifth row, column 3-6). Phospholamban (PLN) and ßMHC were not changed compared to GAPDH, taken as loading control (Figure 3.27.A, row 3-5). Moreover, ANF upregulation was also confirmed by northern blot experiments (Figure 3.27.B, compare column 3-4 with column 1-2). A control experiment (Figure 3.27.C) demonstrates that ANF and a-skeletal actin upregulations depend on MLP deletion because these two markers are not upregulated in the presence of MLP (Figure 3.27.C, column 1 and 3). Once again GAPDH serves as loading control (Figure 3.27.C second row). 72 Figure 3.21. The conditional deletion of ß-catenin in a MLP knock-out background shows an increase of ANF, BNP and a-skeletal actin expression at birth. (A) RNA was extracted from heart ventricle of PO litters, reverse transcribed and cDNA was exponentially amplified in the presence of gene specific primers. Note the upregulation of ANF and BNP for both male and female double KO and the female specific upregulation of a-skeletal actin. (B) RNA was submitted to northern blot analysis with non¬ radioactive probes. The upregulation of ANF is clearly visible for both female and male double KO. (C) control experiment of (A) to show that both ANF and a-skeletal actin upregulation do not occur in ß-catenin cKO at PO. 3.4.4 ANF and g/ß-catenin protein expression Knowing that ANF is upregulated at RNA levels in double knockout neonates, the next step is now to confirm this upregulation at the protein level. In parallel, we also want to know how much of the ß-catenin content is removed at this time from the ventricles. Figure 3.28 shows a compilation of immunoblots. In the first column, a sample of adult atrium is blotted as positive control to detect ANF (Figure 3.28, third row, column 1). The next columns are ventricular samples of double knockout (column 5-7 and 11-13) and MLP-/- samples (column 2-4 and 8-10). ANF shows an increased accumulation in all the double knockouts (third row column 5-7 and 12-13) with one exception (third row, column 11). Note also that ANF accumulation is not changed in all the double knockout animals in the same proportion (third row, compared column 5 with 6 and 7) and the absence of ANF response explains very likely the survival of some "resistant" double knockout animals. In this experiment it is striking to see that the total amount of ß-catenin in ventricular cells is independent of the undergoing genomic deletion (first row, column 2-13). This is really surprising considering that about 50% of cardiomyocytes are already recombined and cannot express ß-catenin any longer (see section 3.3.7, and Table 3.3). As ß-catenin accumulation is not changed, ocE-catenin is not upregulated (Figure 3.28, second row column 2-13) as we observed previously in adult cKO cells (Figure 3.19.A, fifth row). Actin content was used as loading control (Figure 3.28, fourth row) 3.4.5 Sarcomeric organisation As shown in Figure 3.26.E-J, the double knockout induces postnatal myofibrillar disarray in vivo. We wanted to know if this change in structure would also be visible in neonatal cells in culture conditions. Therefore we isolated cardiomyocytes from MLP-/- controls (Figure 3.29.A-D) and double knockout littermates (Figure 3.29.E-H) and looked at nuclear structure and sarcomeric organisation in relation with cell-cell contacts. Quite surprisingly, we did not find any changes in nucleus size, visualised by DAPI staining (Figure 3.29.A,A',E,E') or sarcomeric disarray as demonstrated by phalloidin and a-actinin staining (Figure 3.29.B,B',F,F'). The cell-cell contacts represented by Cadherin staining (Figure 3.29.D.H) were also preserved independent of the presence (Figure 3.29.C,C) or absence of ß-catenin (Figure 3.29.G,G'). The absence of phenotype in cardiomyocytes seems to indicate once again that the role of ß-catenin as intercalated disc protein is not indispensable for the insertion of sarcomeres. If true, this hypothesis means that the signalling activity of ß-catenin 73 At F. ß-cat wt F. ß-cat cKO M. ß-cat wt M. ß-cat cKO Figure 3.28. The conditional deletion of ß-catenin in a MLP knock-out background shows an increase of ANF at the protein level at birth. Immunoblots of SDS extracts from PO ventricles show that the deletion of ß-catenin is not yet visible at the protein level at this stage and the level of another catenin, aE-catenin is not yet upregulated due to the MLP deletion. The increase of ANF, already suggested by the figure 3.27, is confirmed for most of the double KO animals. At=adult atrium sample as positive control. Figure 3.29. Cultured cardiomyocytes from ß-catenin cKO MLP KO have well-organized sarcomeric structures and normal cell-cell contacts. Neonatal ventricular cardiomyocytes were isolated from MLP KO (A-D, A'-C) or ß-catenin cKO MLP KO litter (E-H, E'-Gn) and stained with DAPI (A,A',E,E'), phalloidin (B,F), a-actinin (B',F') ß-catenin (C,C',G,G') and Cadherins (D,H). Compared to MLP KO, double KO cardiomyocytes look similar. The anchoring of sarcomeres to the membrane does not in seem to be affected by the deletion (arrowheads in F,F'G,G' compared with arrows B,B',C,C')- Bar, 10p.m. is most probably the missing factor that leads, in combination with MLP deletion, to postnatal heart failure. 3.5 Importance of ß-catenin in DRAL KO heart 3.5.1 ß-catenin knockout in DRAL knock out mice The experiment with the deletion of ß-catenin in MLP-/- mice suggests that the missing signalling activity of ß-catenin is the determinant factor of the observed lethality. To test this hypothesis further, we decided to repeat the genetic manipulation and introduce this time the conditional deletion of ß-catenin in DRAL-/- mice. As DRAL is a specific interacting co- activator of ß-catenin the rational behind this is to see if the weakening of the signalling pool of ß-catenin can lead to heart failure as well. In the same way, as for the generation of ß- catenin cKO MLP -/- mice, we breed ß-catenin cKO with DRAL -/- animals. As for the MLP experiment, the breeding is performed between strains of different genetic background (C57B1/6 for ß-catenin cKO and B6;129 for DRAL -/-). For this reason, the same warning concerning the influence of genetic background must be repeated here. 3.5.2 Early postnatal lethality associated with double KO mice Breeding of ß-catenin cKO under DRAL -/- background was first analysed at birth. As the experiment is still in progress the number of genotyped animals is rather small. Nevertheless we can detect a slight decrease in the female double knockout population compared to DRAL- /- control littermates (Figure 3.25.C, columns 1-4). The double knockout population shows as well a pronounced postnatal lethality (30% of survival rate) but in this case, compared with the previous experiment, male and female population are evenly affected. This lethality is shown by the genotyping of the breeding at weaning (Figure 3.25.C, columns 5-8 compared with 3.25.B, columns 5-8). After this critical time point, the surviving double knockout animals progress to adulthood and we have never detected so far any signs of fatigue or increased lethality in adult double knockout animals. 3.5.3 Intraventricular septum defect in the double KO mice A careful observation of the litters after birth did not succeed in showing any external phenotypes (Figure 3.30.A). When isolated, the hearts of double knockout animals was grossly undistinguishable in size compared to their DRAL-/- littermates (Figure 3.30.B-C). Therefore, we concluded that the reason for the mortality cannot be caused by an extreme hypertrophic response, like in the previous ß-catenin cKO MLP-/- breeding (Figure 3.26.B- 74 figure 3 30 The conditional deletion of ß-catenin in a DRAL knock-out background shows a perinatal lethality without apparent changes in body weight or heart size and an occasional ventricular septum defect (VSD). At PO, pups from a ß-catenin cKO DRAL KO genotype (A, right) are undistinguishable form ORAL KO littermate (A,left). When isolated, the hearts of those double KO (C) are not enlarged compared to single DRAL KO hearts (B). MRI analysis at L-J 5.5 shows that about half of the double KO embryos (E) have a ventricle septum defect (VSD, red in h) which does not occur in the DRAL KO only (D) LV: left ventricle, RV: right ventricle, IVS: intraventricular septum Bar, lOOiim. C). Because of the absence of a visible phenotype, we decided to investigate a possible deleterious malformation of the heart in response to ß-catenin and DRAL deletion. To do this, we took advantage of the MRI technique which allows the three dimensional visualisation of embryonic mouse heart (Schneider et al., 2003). With the help of this method, it was possible to demonstrate that there was no overt oedema or defects in the great vessels, atria, or veins in the embryos at E15.5 (Prof. S. Battacharya, personal communication). The surprise came with the analysis of the ventricles because about 50% of the double knockout displayed a ventricular septum defect (VSD) (Figure 3.30.E). VSD is by definition the presence of an open persistent connection between right and left ventricle. At the time we analysed the embryos (El 5.5) the two chambers should be separated as shown in (Figure 3.30.D). Although a VSD is not sufficient to induce a lethal phenotype per se at embryonic stage, it has a negative impact after birth on cardiac function because of the mixing of the non- oxygenated and oxygenated blood in the open gap. So far, we cannot explain the link between VSD and the loss of ß-catenin/DRAL animals but we know that DRAL is expressed at high level in the primordium of the ventricular septum (Kong et al., 2001). Because of this, a tempting hypothesis would be that the concomitant deletions of ß-catenin and DRAL lead to an incomplete closure of the septum. 75 4 Discussion 4.1 Development of the intercalated disc: a long process of maturation 4.1.1 Critical analysis of the method: A major point of discussion concerns the method of investigation for the results obtained in section 3.1. Immunohistochemistry on whole mount heart preparations is limited by the diffusion of antibodies into the layers of cells. Although well adapted to small hearts (E8.5 until El8.5), this technique fails with neonatal and postnatal hearts and we had to perform cryosections before immunohistochemistry for these samples. We cannot rule out that this supplemental step introduces some changes in the preparation, like retraction/dilation of tissues or deletion/unmasking of antigens. Nevertheless a previous study on ICD development in embryonic and neonatal rat heart using cryosections does not show major differences compared with our results concerning adherens junctions and ECM stainings (Wu et al., 2002). Another important criticism concerns the unequal state of maturation of cardiomyocytes within regions of ventricles. People have shown that the development of cardiomyocytes is submitted to regional differences (Meilhac et al., 2003; Meilhac et al., 2004). Because of this, we have to be really careful when interpreting the data obtained from whole mount and cryosections and keep in mind that the morphologic changes observed throughout development in ventricular cells (Table 3.1) could be delayed or accelerated depending on the particular subregions of the ventricle we are looking at. 4.1.2 How do myofibrils grow ? When looking at myofibril growth during fetal development (section 3.1.5 and first row of Table 3.1) we saw that the length of myofibrils changes from 12 um to 42 urn. As this length is correlated with the cell length (compare in Table 3.1 first row with third row), we can predict that the length of myofibrils in adult cells, which have passed through the so-called developmental hypertrophy, reaches about 100 to 120 jam (Leu et al., 2001). Now, how is such an extension of the myofibril possible? Myofibrils must have a mechanism to integrate new sarcomeres to allow extension of the whole structure but so far, almost nothing is known about this integration process. If the end-process in unknown, more information is available on factors involved in the growth induction because the signalling pathways which are involved in developmental hypertrophy are generally also responsible for pathological cardiomyopathies (for review see Nicol et al., 2000). Among them, it seems that activation of one branch of the MAPK pathways (MEK5) is specifically involved in the increase of 76 myofibril length and its constant activation in heart leads to the thinning of the ventricular wall characteristic of DCM (Nicol et al., 2001). In the same field, cardiotrophin-1 (CT-1), a cardiac cytokine expressed early in heart development, is also known as a specific factor promoting the growth of embryonic cardiomyocytes by promoting the assembly of sarcomere in series (Sheng et al., 1996; Wollert et al., 1996). 4.1.3 What is the driving force for the alignment of myofibrils? The progressive alignment of myofibrils inside the cells, as illustrated in Figure 3.1 and quantified in the second row of Table 3.1 is also a process which remains unclear, from a mechanistical point of view. It is known that myofibrils have a certain plasticity, which can be positively influenced by the stretching of the whole tissue. The stretching induces an anisotropic alignment of the myofibrils parallel to the direction of contraction, as observed by others (Zimmermann et al., 2002; Gonen-Wadmany et al, 2004). It is therefore possible to postulate that this alignment of cardiomyocytes is a self-activated mechanism. At the very beginning, a small number of poorly organised cardiomyocytes are responsible for the contraction of the heart. For this reason, this contraction resembles a peristaltic movement (E8.5-E9), which is rather inefficient for circulating blood inside the whole body. The cardiac rhythm is then rapidly established and increases until birth (Ishiwata et al., 2003). Together with the formation of the conduction system (El2 until birth) and the augmentation of myofibrillar content, this results in an increased strength of contraction, which then progressively drives the alignment of the growing myofibrils. This hypothetical "auto- alignment" process is supported by in vitro cultures of neonatal cardiac cells, which display an almost complete alignment of myofibrils when grown either on grooved substrates (Bursac et al., 2002) or on aligned collagen strands (Evans et al., 2003). 4.1.4 Embryonic developmental hypertrophy ? The data provided in the third to the last row of Table 3.1 are demonstrating that the developmental hypertrophy is starting earlier than estimated previously. This embryonic developmental hypertrophy is unidirectional: in about 10 days, cardiomyocyte length is more than 3 times increased (12.2 urn versus 38.2 jam) whereas the width is essentially unchanged (9.4um versus 9.5 urn). This is in clear contrast compared to the postnatal developmental hypertrophy (Leu et al., 2001) where both length and width are augmented from P4 (49.3 urn in length, 14.0 urn in width) to adulthood (128.0 (am in length, 35.8 um in width). We have to mention here that the measurements on freshly isolated postnatal cells gave results slightly greater, but in accordance with our measurements, made on cryosections. 77 4.1.5 Changes in nuclear morphology As cells grow in length and therefore occupy more space inside the tissue, the density of nuclei decreases (Figure 3.1.C,G,K). A more surprising finding concerns the size of cardiomyocyte nuclei. From the data of Table 3.1 (seventh to ninth row), it looks that nuclear size and geometry are changed throughout embryonic development. Although roughly measured with the nuclear length and width, the decrease of nuclear size is as well illustrated in Figure 3.1C,G and supported by the study of Xavier-Vidal and colleagues which show a decrease of nuclear volume in human fetal cardiomyocytes from the second to the third trimester of gestation (Xavier-Vidal and Mandarim-de-Lacerda, 1995). The meaning of this nuclear compaction in not clear yet and we can only hypothesize that this reflects a difference in transcriptional activity and/or cell cycle activity. If the first hypothesis is true, it means that young cardiomyocytes have many more transcriptionally active genes than neonatal and adult cells and the proportion of expanded euchromatin is therefore greater leading to this apparent bigger nuclei, as seen in other cell types (Sato et al., 1994; Frenster, 1974) and in malignant tumour cells {Ikeguchi, 1998 #418}. On the other hand, it is known that DNA synthesis associated with cell cycle increases nuclear size (Ikeguchi et al., 1998). Therefore the bigger nuclei in young cardiomyocytes can also reflect the higher proliferation rate at this stage. 4.1.6 What does the polarisation of AJ and DJ, compared with GJ mean? Immunochemistry and confocal microscopy presented in Figure 3.5 to Figure 3.7 documents the development of the ICD in mouse cardiomyocytes observed at different embryonic stages. Although these data are descriptive, they provide a milestone for the understanding of how cardiomyocytes establish and modify their specific cell-cell contacts in vivo and complete the existing knowledge of the postnatal ICD development (Angst et al., 1997). Compared to the assembly of the sarcomeric apparatus, which takes only a few hours (Ehler et al., 1999), the maturation of the ICD is surprisingly slow and progresses along with the elongation of the cells, and the alignment of myofibrils. Whereas adherens junction and desmosome localisation changes from a surrounding to a longitudinal restricted pattern and accompanies the parallel alignment of myofibrils, (Figure 3.5.A compared with Figure 3.5.0 and Figure 3.6.A compared with Figure 3.6.K), the distribution of gap junctions is not restricted to a bipolar pattern even in postnatal cells (Figure 3.7.K) This particular localisation of gap junctions has an important significance in term of conduction because it has already been shown that lateral gap junctions, as observed in our neonatal and postnatal micrographs, smooth the excitation wave front during longitudinal conduction (Fast and Kleber, 1993) so 78 that the electrical properties of the neonatal cardiomyocytes are quite different compared to adult cells. The late restriction of connexin 43-based gap junction to the ICD seems to be dependent on the pre-existing cell-cell contacts (Hertig et al., 1996a), an hypothesis supported by the interaction of connexin 43 with ß-catenin (Ai et al., 2000) or ZO-1 (Barker et al., 2002; Barker et al., 2001). Nevertheless, the reason for the delay of gap junction restriction to the ICD is still partially unclear. 4.1.7 The sorting out of cell-cell contacts and ECM contacts in postnatal heart The distribution of cell-to-ECM contacts in embryonic heart, as visualised by the anti-laminin staining in Figure 3.2.B and the whole panel of Figure 3.8, is first restricted to non-cardiac or not yet differentiated cells (Figure 3.8.A,C). With the elongation of the cardiomyocytes which creates a longitudinal and a junctional cell border, cell-cell contact made of adherens and desmosomal junctions are coexisting with cell-to-ECM contacts at the lateral sides of the membrane. This coexistence remains even after birth (Figure 3.2.F) and is clearly resolved only in adult cells (Figure 3.2.1). It is therefore not surprising that this mixing of cadherin- mediated and ECM-based contact was already reported, in cultivated cardiomyocytes (Goncharova et al., 1992) and in embryonic rat heart (Wu et al., 2002). How the complete sorting out of lateral costameric contacts and polarised cell-cell contacts is achieved throughout embryonic and postnatal development is of course not understood but it probably involves a differential stability of the cell-cell contacts at the lateral border, which is then degraded and replaced by ECM contacts more adapted to support the stress created at the side by the rhythmical contraction of the cell. 4.2 ICD-myofibril reconstruction in vitro 4.2.1 Criticisms of the method The results presented in the section 3.2 are based on transient transfections of cell-cell contact and sarcomeric proteins in neonatal rat cardiomyocytes. All the constructs are tagged with the green, respectively red fluorescent proteins GFP, DsRed/RFP to allow direct detection in epifluorescent microscopy. Although convenient for studies of protein localisation, the weak point of this method is the lack of expression regulation characterised in some cases by the over expression of the transfected proteins and resulting in uninterpretable data (see arrow in Figure 3.10.M). A second potential problem is the fluorescent tag itself. As illustrated by the Figure 3.13 the fusion of sarcomeric proteins with fluorescent protein can negatively influence the localisation of the fusion partner. The artifact can be due to the oligomeric 79 structure of the fluorescent protein (Baird et al., 2000), as shown in Figure 3.13.E but sometimes, particular domains of the fusion partner, like binding sites or localisation sequences can be disturbed by the fusion. Therefore, the use of GFP/RFP fusion proteins to detect the protein of interest must be tested for each construct individually, with N-terminal and C-terminal fusion to define good candidates for further analysis. 4.2.2 a-catenin as a specific marker of cell-cell contact As shown in Figure 3.10, a-catenin-tagged protein is the most reliable catenin fusion protein that decorates the adherens junction (compare a-catenin localisation in Figure 3.10.A with Cadherins in Figure 3.10.C). Transfection of ß-catenin or plakoglobin (Figure 3.10.D,J) and connexin isoforms (Figure 3.11) were much less efficient in decorating the adherens, respectively gap junction. Although interesting, we did not perform supplementary analysis to find out why ß-catenin and plakoglobin transfection did not reliably localize at the ICD but data from transient transfection of stabilized ß-catenin including our study (F.Schatzmann, diploma work, 2001) strongly suggest that the ectopic expression of either ß-catenin or plakoglobin saturates the degradation machinery resulting in cytoplasmic and nuclear localisation, as seen after Wnt activation (de Melker et al., 2004). The transfection of focal adhesion associated protein shown in Figure 3.12 serves as a negative control for the localisation experiment but illustrates also the difference in the anchoring of myofibrils in cultivated cells, with the formation of a gap between the cellular membrane and the first sarcomere (3.12.A,B arrow) which is not found in vivo. 4.2.3 RFP, a promising complement to GFP in dual labelling As an important part of the research in our lab is dedicated to the assembly of the contractile apparatus and the localisation of sarcomeric proteins, we had many different constructs which could be used for the labelling of specific compartments of the sarcomere. Among them a- actinin constructs have displayed the most robust and constant localisation at the Z-disc (Figure 3.9). For this reason, constructs with red fluorescent proteins were generated with this gene and evaluated in transient transfection (Figure 3.13). Among the two variants of the DsRed protein, the monomeric version, renamed RFP (Campbell et al., 2002) gave satisfactory results and was further successfully subcloned in bicistronic expression system together with GFP fusion proteins (Figure 3.14). The a-actinin/a-catenin dual labelling tested here is therefore a suitable tool to follow the integration of myofibrils in the ICD, at least in transient transfections (Figure 3.14.D,E). 80 4.2.4 Labelling of the ICD-myofibril interface in 3D and during myofibrillogenesis Different models allow the observation of the cardiac tissue in 3D. Among them, expiants of chicken hearts (Shiraishi et al., 1993) is also suitable for transient transfections of sarcomeric proteins (Figure 3.9). Compared with the transfection of isolated neonatal cells, this model is closer to the in vivo situation and our dual labelling could be used there to monitor myofibrillogenesis in time lapse microscopy. The insertion of sarcomeres is another interesting aspect of myofibrillogenesis which could also be studied with our dual labelling system. As mentioned in section 4.1.2 the contractile apparatus is subject to a tremendous elongation during development (see also Table 3.1, first row and Leu et al., 2001) but the mechanism of sarcomere integration is not known. The use of dual fluorescent labeling of sarcomeric proteins (Figure 3.14.A,B) in time lapse microscopy could therefore get insights into the process of myofibril elongation. 4.3 Importance of ß-catenin under physiological and under stress conditions 4.3.1 Reasons for the slow disappearance of ß-catenin Taken together, the data from section 3.3.3 to 3.3.7 document a slow disappearance of ß- catenin mainly in postnatal development. One of the most relevant hypotheses to explain these results is the inefficient process of the Cre recombinase, supported by the ROSA26 reporter experiment (Figure 3.20). The absence of early phenotypes in all characterised conditional knockouts using MLC2v-Cre mice (see section 1.4.2) also substantiates this hypothesis. As example, the conditional deletion of ßi-integrin shows a reduction of the protein content only at 4 weeks after birth (Shai et al., 2002). Connexin 43 deletion is the only reported study where a decrease of the protein is already visible at embryonic stages (Gutstein et al., 2001). A second hypothesis to explain the slow disappearance of ß-catenin is the protein stability, ß-catenin is traditionally considered as a protein with a rapid turnover (Kang et al., 1999) but the half-live of the protein is measured in the cytoplasmic fraction and does not take into account the stabilisation of the protein when inserted in the cell-cell contacts. This "ICD-bound" fraction of ß-catenin with increased stability is in accordance with the ß- catenin immunostainings in our conditional knockout (Figure 3.16.B'and 3.17.G) where the deletion is barely visible at the ICD although about 50% of cardiomyocytes should be recombined at this stage (Table 3.3). For this reason, we speculate that the ß-catenin content in the cells is composed of two pools with different half-lives. Upon Cre-driven genomic deletion, the free cytoplasmic and signalling protein pool is rapidly degraded but the ICD- 81 bound pool is kept away from the degradation machinery and declines only at 2 months after birth (Figure 3.19.B, first two columns). 4.3.2 Adaptation of adherens, desmosomal and gap junction to the ß-catenin deletion Section 3.3.6 documents the changes affecting the cell-cell contacts in the absence of ß- catenin. The specific upregulation of adherens junction proteins (Figure 3.19.A,B) is very likely here a compensation phenomenon because it involves the remaining participant of the cadherin/catenin complex but not the more distant relative aT-catenin (Janssens et al., 2001) and vinculin. Such specific augmentation of adherens proteins has already been documented in DCM (Ehler et al., 2001) and seems therefore to be regulated in a coordinated manner. The constant level of connexin 43 is a good indicator of the harmless effect of ß-catenin removal because this gap junction protein is very sensitive to stress conditions (Severs et al., 2004) and is down or upregulated in many models of cardiomyopathy including DCM (Ehler et al., 1999) and HCM (Wang and Gerdes, 1999). Moreover, the Wnt activated ß-catenin is supposed to be an important factor which positively regulates the expression of connexin 43 (van der Heyden et al., 1998; Ai et al., 2000). From our results, we have to conclude that the transcriptional activity of ß-catenin in adult cardiac cells is low and/or that other regulatory pathways are involved to sustain the expression of this gap junction isoform when ß-catenin is no longer present. 4.3.3 No cardiac hypertrophy but cellular hypertrophy. Is this possible? Data from sections 3.3.8-9 are somehow contradictory in the sense that they show the absence of cardiac hypertrophy and hypertrophic response (Figure 3.12) but in the same time an increased cellular volume of cardiomyocytes (Figure 3.13). One can first argue that the heart weight to body weight ratio (Figure 3.2LA) is not sensitive enough to pick up subtle hypertrophic changes, as mentioned by others (Deschepper et al., 2002) and the expression of the different hypertrophy markers is not activated (3.21.B) because the cellular hypertrophic response is very likely not a pathological process. For this phenomenon, we entertain two different hypotheses. On one hand, the increased cellular volume could be a compensation for the loss of cells occurring at a slow rate and therefore undetectable in our experiment (Figure 3.23) On the other hand, ß-catenin has been firmly established as a potent proliferative agent involved in the majority of colorectal neoplasia (Peifer, 1997; Morin et al., 1997) and in embryonic development including the cardiac outflow tract (Clevers, 2002). We can therefore hypothesize that the deletion of ß-catenin at birth (about 50% of the cardiomyocytes, see 82 Table 3.3) is enough to remove the cytoplasmic pool of the protein leading to a decrease in signalling activity and a premature stop of cellular division in the recombined cells. In order to compensate this lack of cellular proliferation, the hypertrophic growth during postnatal development (Leu et al., 2001) would then increase the mean volume of cardiomyocytes. If this second hypothesis is true, the cellular hypertrophy detected in isolated cells (Figure 3.22) should then be considered as well as a compensative healthy mechanism. The 3D reconstruction pictures (Figure 3.22.G) are also showing changes in cell morphology, resulting in a more "branched" pattern reminiscent of the MLP KO (Ehler et al., 2001) but the volume increase (+35%) cannot be linked to a concentric or eccentric hypertrophy (Nicol et al., 2001; de Simone, 2004), as the change in cell lengh to cell width ratio is not significant (Table 3.4). This is of course another argument in favour of the harmlessness of ß-catenin removal for the heart function. 4.3.4 ß-catenin as a regulator of hypertrophy? The increased cellular volume in ß-catenin cKO cardiomyocytes is in contradiction with published results describing ß-catenin as a necessary and sufficient factor for the induction of hypertrophy in cardiomyocytes, which can be blocked upon phosphorylation by GSK-3ß (Haq et al., 2000; Haq et al., 2003). We can in this case argue that Antos and colleagues also found GSK-3ß as a key player in blocking hypertrophic responses using transgenic mice with a constitutive active GSK-3ß expressed under the control of a heart specific promoter. This phenotype did not seem to involve the blocking of ß-catenin, but NFAT transcription factors (Antos et al., 2002). A definitive answer to this contradiction could be brought by the heart specific stabilisation of ß-catenin using the ß-catenin Exon3 floxed transgenic mouse (Harada et al., 2002). 4.3.5 The sarcomeric structure is preserved As ß-catenin is part of the adherens junction, the cardiac specific deletion of this gene was supposed to weaken the attachment of myofibrils to their anchoring structure. If true, the weakening of this contact would be visible as a disarray of the sarcomeric structure. Figure 3.22 clearly shows that this is not the case and we have to conclude that ß-catenin deletion can be compensated by other molecules, as seen in Figure 3.19 so that the cellular contractility can be preserved. The heart specific deletion of a-catenin, another member of the cadherin/catenin complex is also devoid of obvious heart defects (Ju Chen, personal communication). From this we have to draw the conclusion that the adherens junctions of the 83 ICD are more stable than we naively thought, probably because a certain redundancy exists between the members of the complex: forced E-cadherin expression rescues lethality of the N- cadherin deletion (Ferreira-Cornwell et al., 2002), plakoglobin and ß-catenin are to some extend exchangeable in the cadherin/catenin complex (Bierkamp et al., 1999) and there are at least two a-catenin isoforms expressed in heart cells (Janssens et al., 2001). 4.3.6 Deletion of ß-catenin in hypertrophic conditions: consequences on contractility Echocardiography measurement in basal conditions did not show any difference at 2 months and 6 months between WT and ß-catenin mice (Table 3.5 and 3.6) and the deletion of ß- catenin seems to significantly increase the contractility of the heart in ß-adrenergic induced hypertrophy at 2 months (Table 3.6 last row and Figure 3.24.D). Although not clearly understood in details, these experiments bring the confirmation that the deletion of ß-catenin is innocuous for the adult heart, at least from a functional point of view. In a broader perspective, we can also conclude from these data that the induction of HCM on top of a cytoskeletal deletion has no severe effects. It would then be interesting to see if DCM induction would be worse than HCM induction, as expected from the generally accepted models which describes DCM as a phenotype caused by a force transduction problem and linked to cytoskeletal mutations (Chien, 1999; Perriard et al., 2003). A major drawback for testing this hypothesis is the impossibility so far of inducing DCM with a surgical intervention or pharmacological treatment. The only possible way requires breeding with transgenic or knockout models of DCM like MLP KO mice (Arber et al., 1997) as shown in section 3.4.1 and discussed later in section 4.4.2. 4.3.7 Early lethality of female cKO, dream or reality? The deletion of ß-catenin induces a loss of animals in the female population in the three first postnatal weeks (Table 3.2). This is really intriguing considering the usual advantage of females to survive cardiovascular diseases (Leinwand, 2003) mainly attributed to the cardioprotective effect of oestrogen (Wagner et al., 2002). So far, our attempt to demonstrate at birth a gender-related enhancement of ß-catenin genomic deletion failed (Figure 3.15.B,C) and from these combined data we have to conclude that females are more sensitive than male to overcome the critical postnatal period. If true, the heart specific deletion of ß-catenin would be the second reported case in rodent transgenesis where a heart specific gene modification improves male survival compared to female. The first report concerns the heart specific TNFa 84 overexpression (Kadokami et al., 2000). The gender specific survival is also discussed later in relation with MLP and DRAL double knockout. 4.3.8 Regulation of ß-catenin in the healthy and hypertrophic heart Several models of HCM (van Veen et al, 2002; Tsybouleva et al, 2004; Masuelli et al, 2003) as well as DCM (Ehler et al, 2001) show an increased accumulation of ß-catenin and/or spatial redistribution of the protein consistent with an increased signalling activity. These reports contrast with the relative absence of phenotype in our adult mice after the removal of ß-catenin, as seen by the echocardiographic data in Tables 3.5-6 and the absence of hypertrophic response in Figure 3.21. In a broader perspective, these findings constitute very likely the real take-home message of the experiment: although essential for embryonic development (Haegel et al, 1995), and for early postnatal heart development, ß-catenin is then mostly dispensable for the adult heart and can even be a factor involved in pathological hypertrophy. 4.4 Deletion of ß-catenin on the top of LIM protein deletion 4.4.1 Critics of the mix in genetic backgound Several examples of research in cancer (Hunter et al, 2003), behaviour (Holmes et al, 2002) and cardiovascular diseases (Le Corvoisier et al, 2003a; Le Corvoisier et al, 2003b) have already shown that the results of animal experimentation and clinical trials can be severely influenced by the variability of the genetic background in the tested population. To minimize this problem in animal experimentation, researchers developed inbreed strains by several generation of sibiling's mating leading to an almost identity of the genomic content among the offspring of each particular strain. Unfortunately, each time that a breeding is performed between mice of different inbreed strain, the problems concerning the genetic background variation and the genetic drift come out (Wotjak, 2003). These problems could only be solved by at least six consecutive backcrossing in one of the two inbreed strain used and takes about two to three years. In our case, the MLP-ß-catenin double knockouts as well as DRAL-ß-catenin double knockouts have a mixed genetic background consisting of 129sv and C57B1/6 strains. These mixed genetic background were impossible to avoid considering the time required for the backcrossing to reduce this genetic variability (see above). Nevertheless, when we sum up by gender and genotype the whole population of each double knockout experiment we can already draw some very interesting conclusions (Figure 3.25). First of all, the postnatal 85 lethality approaches about 70% for both double knockouts (68.4% for MLP-ß-catenin KOs respectively 70,8% for DRAL-ß-catenin KOs). Secondly, most of the lethality observed is postnatal, as shown by the comparison of population between birth and weaning age. Thirdly, the increase of lethality observed already in ß-catenin knockout females is increased in the MLP-ß-catenin KO but not obvious in the DRAL-ß-catenin KO. This suggests that the gender sensitivity to the removal of ß-catenin can be further modulated by the deletion of LIM domain proteins. 4.4.2 Are there possible explanations for the early lethality of MLP-ß-catenin KOs? Compared to the HCM induced by a ß-adrenergic agonist in ß-catenin cKO adult animals KO is (Figure 3.24), the same deletion in a DCM background (breeding with MLP mice) much more dramatic, resulting in the death of about 70% of animals during the first postnatal week. Although this comparison is somehow misleading in the sense that the HCM is induced in the adult and the DCM lethality is observed at birth, we think that these observations fit with the axiom of cardiomyopathy development (Chien, 1999): ß-catenin being a cytoskeletal component, its removal has much stronger effects when conjugated with DCM, which is a force transmission problem due to defects in the cytoskeletal organisation. The major characteristics of the MLP-ß-catenin double KO animals are the postnatal enlargement of the heart with the loss of cardiomyocytes in vivo and the upregulation of ANF both at the mRNA and protein level, a situation already found with low penetrance in the MLP knockout (Arber et al, 1997). When isolated in culture, cardiomyocytes have a preserved sarcomeric organisation and well-established cell-cell contacts compared with MLP KO control cells (Figure 3.29) suggesting that the cell death observed in vivo is weakened in culture conditions. DRAL KO? 4.4.3 Is there a possible explanation for the early lethality of ß-catenin cKO In contrast to MLP, DRAL protein interacts directly with ß-catenin. Different publications act as co-activators of the suggest even a tripartite interaction where ß-catenin and DRAL androgen receptor (AR) so that the activity of this receptor can be modulated (Pawlowski et al, 2002; Wei et al, 2003). The AR binds specifically to male hormones such as testosterone and dihydrotestosterone (Zhou et al, 1995). Upon ligand binding, the receptor dissociate from his chaperon heat shock protein and is then rapidly translocated into the nucleus where it binds to androgen receptor elements (ARE) and activates various signalling pathways (MacLean et al, 1997; McEwan, 2004). Knowing that AR is expressed in neonatal heart 86 (Marsh et al, 1998) and that the blood level of masculine hormones is rising in males already before birth (Hughes et al, 2001), so that there is a discrepancy of androgen signalling of DRAL and between males and females, we can hypothesize that the concomitant deletion ß-catenin strongly reduces the AR activity in the myocardium of male mice. As seen in Figure 3.25, the DRAL-ß-catenin KO animals do not have a preferential survival of males (compared link this with ß-catenin cKO alone or MLP-ß-catenin KOs). It is therefore really tempting to conclusion also observation with a hypothetical loss of AR activity in male neonates. This for means that the AR signalling has cardioprotective effects during the postnatal period ß- catenin cKO and MLP-ß-catenin KO male mice but the mechanisms leading to this protection is far from being understood. 4.4.4 Ventricular septum defects as a cause of postnatal lethality? The most striking phenotype of the DRAL-ß-catenin KO experiment is the occasional ventricular septum defect (VSD) observed so far in 50% of the double knockout animal at El5.5 (Figure 3.30.E). As DRAL is normally expressed at high level in the intraventricular of the septum (Kong et al, 2001), it is easy to conceive that ß-catenin removal on the top in the DRAL deletion is interrupting some morphogenic or proliferative signals resulting incomplete closure of the septum. The exact frequency of the VSD phenotype in the DRAL- double ß-catenin KO and the possible occurrence of VSD in ß-catenin single knockout and for the results based on MLP-ß-catenin knockouts are not known because we are still waiting not know as well if the VSDs observed are a larger number of embryos for each group. We do important enough to negatively influence the heart function in postnatal hearts and induce lethality per se. in about 25% of all In human, VSD is known as a very common congenital heart defect found the cases of heart malformation. The transmission is described as autosomal recessive with variable expression and affects equally boys and girls. Larger holes may interfere with a and child's feeding and growth and may cause rapid breathing, excessive sweating, poor and back weight gain. The vessels which carry blood from the heart to the lungs again may heart failure. This become congested, or overloaded with blood, resulting in congestive of VSD has been usually occurs when the child is 6 to 8 weeks old. In mice, the occurrence et reported in many different transgenic animals (Clark et al, 1999; Lee et al, 2000; Feng al, and 2002; Sakata et al, 2002) but there is so far no obvious link between any of these models either DRAL or ß-catenin. 87 4.4.5 Postnatal lethality in relationship with the initiation of developmental hypertrophy? It is remarkable that the critical period where most of the lethality is observed in our single or double knockout experiments (Figure 3.25) corresponds to the onset of postnatal Leu et developmental hypertrophy as described before (Armstrong et al, 2000; al, 2001). Therefore, it is possible that the switch between hyperplastic proliferation and hypertrophic in growth (Li et al, 1996) is affected in our knockout model considering the role of ß-catenin hyperplastic proliferation (Rottbauer et al, 2002). This idea is reinforced by the increased cellular hypertrophy (Figure 3.22) that we attribute to a premature inhibition of cardiomyocyte proliferation in the absence of ß-catenin (section 3.3.8) and emphasizes once of again the key role of ß-catenin as a signalling molecule for the regulation proliferative the processes. This remark also highlights one crucial point for future experiments: necessity in order to a full and to measure individually the content of the two pools of ß-catenin get comprehensive representation of the role of this protein in embryonic and postnatal heart development. 88 i 5 References Aberle, H., Butz, S., Stappert, J., Weissig, H., Kemler, R. and Hoschuetzky, H. (1994). Assembly of the cadherin-catenin complex in vitro with recombinant proteins. J Cell Sei 107 ( Pt 12), 3655-63. Aberle, H., Bauer, A., Stappert, J., Kispert, A. and Kemler, R. (1997). beta-catenin is a target for the ubiquitin-proteasome pathway. Embo J16, 3797-804. Agah, R., Frenkel, P. A., French, B. A., Michael, L. H., Overbeek, P. A. and Schneider, M. D. (1997). Gene recombination in postmitotic cells. Targeted expression of Cre recombinase provokes cardiac-restricted, site-specific rearrangement in adult ventricular muscle in vivo. J Clin Invest 100,169-79. Ahuja, P., Perriard, E., Perriard, J. C. and Ehler, E. (2004). Sequential myofibrillar breakdown accompanies mitotic division of mammalian cardiomyocytes. J Cell Sei 117, 3295-306. Ai, Z., Fischer, A., Spray, D. C, Brown, A. M. and Fishman, G. I. (2000). Wntl regulation of connexin43 in cardiac myocytes. J Clin Invest 105,161-71. Alberts, B., Bray, D., Lewis, J., Raff, M., Roberts, K. and Waston, J. D. (1994). Molecular Biology of the Cell. New York: Garland Publishing. Alcolea, S., Theveniau-Ruissy, M., Jarry-Guichard, T., Maries, I., Tzouanacou, E., Chauvin, J. P., Briand, J. P., Moorman, A. F., Lamers, W. H. and Gros, D. B. (1999). Circ Res Downregulation of connexin 45 gene products during mouse heart development. 84, 1365-79. Allen, E., Yu, Q. C. and Fuchs, E. (1996). Mice expressing a mutant desmosomal Cadherin exhibit abnormalities in desmosomes, proliferation, and epidermal differentiation. J Cell Biol 133,1367-82. Angst, B. D., Nilles, L. A. and Green, K. J. (1990). Desmoplakin II expression is not restricted to stratified epithelia. J Cell Sei 97 ( Pt 2), 247-57. Angst, B. D., Buxton, R. S. and Magee, A. I. (1995). Characterization of human cardiac desmosomal Cadherins. Ann N YAcad Sei 752, 101-4. Angst, B. D., Khan, L. U., Severs, N. J., Whitely, K., Rothery, S., Thompson, R. P., Magee, A. I. and Gourdie, R. G. (1997). Dissociated spatial patterning of gap junctions and cell adhesion junctions during postnatal differentiation of ventricular myocardium. Circ Res 80, 88-94. 89 Antos, C. L., McKinsey, T. A., Frey, N., Kutschke, W., McAnally, J., Shelton, J. M., Richardson, J. A., Hill, J. A. and Oison, E. N. (2002). Activated glycogen synthase-3 beta 907-12. suppresses cardiac hypertrophy in vivo. Proc Natl Acad Sei USA99, Arber, S., Halder, G. and Caroni, P. (1994). Muscle LIM protein, a novel essential regulator of myogenesis, promotes myogenic differentiation. Cell 79, 221-31. Arber, S., Hunter, J. J., Ross, J., Jr., Hongo, M., Sansig, G., Borg, J., Perriard, J. C, Chien, K. R. and Caroni, P. (1997). MLP-deficient mice exhibit a disruption of cardiac cytoarchitectural organization, dilated cardiomyopathy, and heart failure. Cell 88, 393-403. Armstrong, M. T., Lee, D. Y. and Armstrong, P. B. (2000). Regulation of proliferation of the fetal myocardium. Dev Dyn 219, 226-36. Auerbach, D., Rothen-Ruthishauser, B., Bantle, S., Leu, M., Ehler, E., Helfman, D. and Perriard, J. C. (1997). Molecular mechanisms of myofibril assembly in heart. Cell Struct Fund 22,139-46. Auerbach, D., Bantle, S., Keller, S., Hinderung, V., Leu, M., Ehler, E. and Perriard, J. C. (1999). Different domains of the M-band protein myomesin are involved in myosin binding and M-band targeting. Mol Biol Cell 10, 1297-308. Baird, G. S., Zacharias, D. A. and Tsien, R. Y. (2000). Biochemistry, mutagenesis, and Natl Acad Sei oligomerization of DsRed, a red fluorescent protein from coral. Proc USA97, 11984-9. Barker, R. J., Price, R. L. and Gourdie, R. G. (2001). Increased co-localization of connexin43 and ZO-1 in dissociated adult myocytes. Cell Commun Adhes 8,205-8. Barker, R. J., Price, R. L. and Gourdie, R. G. (2002). Increased association ofZO-1 with connexin43 during remodeling of cardiac gap junctions. Circ Res 90, 317-24. Basson, C. T., Bachinsky, D. R., Lin, R. C, Levi, T., Elkins, J. A., Soults, J., Grayzel, D., Kroumpouzou, E., Traill, T. A., Leblanc-Straceski, J. et al. (1997). Mutations in human Nat Genet TBX5 [corrected] cause limb and cardiac malformation in Holt-Oram syndrome. 15, 30-5. Beardslee, M. A., Laing, J. G., Beyer, E. C. and Saffitz, J. E. (1998). Rapid turnover of connexin43 in the adult rat heart [see comments]. Circ Res 83, 629-35. Behrens, J., Jerchow, B. A., Wurtele, M., Grimm, J., Asbrand, C, Wirtz, R., Kühl, M., Wedlich, D. and Birchmeier, W. (1998). Functional interaction of an axin homolog, conductin, with beta-catenin, APC, and GSK3beta. Science 280, 596-9. Ben-Ze'ev, A. and Geiger, B. (1998). Differential molecular interactions of beta-catenin and plakoglobin in adhesion, signaling and cancer. Curr Opin Cell Biol 10, 629-39. 90 Berken, A., Abel, J. and Unfried, K. (2003). beta 1-integrin mediates asbestos-induced phosphorylation of AKT and ERK1/2 in a rat pleural mesothelial cell line. Oncogene 22, 8524-8. Beyer, E. C, Paul, D. L. and Goodenough, D. A. (1987). Connexin43: a protein from rat J Cell Biol 2621-9. heart homologous to a gap junction protein from liver. 105, Bierkamp, C, McLaughlin, K. J., Schwarz, H., Huber, O. and Kemler, R. (1996). Embryonic heart and skin defects in mice lacking plakoglobin. Dev Biol 180, 780-5. of Bierkamp, C, Schwarz, H., Huber, O. and Kemler, R. (1999). Desmosomal localization beta-catenin in the skin of plakoglobin null-mutant mice. Development 126, 371-81. Borg, T. K. and Caulfield, J. B. (1981). The collagen matrix of the heart. Fed Proc 40, 2037-41. A. M. Borg, T. K., Goldsmith, E. C, Price, R., Carver, W., Terracio, L. and Samarel, (2000). Specialization at the Z line of cardiac myocytes. Cardiovasc Res 46, 277-85. adhesion. Braga, V. M. (1999). Small GTPases and regulation of Cadherin dependent cell-cell Mol Pathol 52,197-202. Brault, V., Moore, R., Kutsch, S., Ishibashi, M., Rowitch, D. H., McMahon, A. P., Sommer, L., Boussadia, O. and Kemler, R. (2001). Inactivation of the beta-catenin gene by Wntl-Cre-mediated deletion results in dramatic brain malformation and failure of craniofacial development. Development 128,1253-64. Bruning, J. C, Michael, M. D., Winnay, J. N., Hayashi, T., Horsch, D., Accili, D., Goodyear, L. J. and Kahn, C. R. (1998). A muscle-specific insulin receptor knockout exhibits features of the metabolic syndrome ofNIDDM without altering glucose tolerance. Mol Cell 2, 559-69. cultures with Bursac, N., Parker, K. K., Iravanian, S. and Tung, L. (2002). Cardiomyocyte studies of controlled macroscopic anisotropy: a model for functional electrophysiological cardiac muscle. Circ Res 91, e45-54. Campbell, R. E., Tour, O., Palmer, A. E., Steinbach, P. A., Baird, G. S., Zacharias, D. A. and Tsien, R. Y. (2002). A monomeric red fluorescent protein. Proc Natl Acad Sei USA99, 7877-82. Capetanaki, Y. and Milner, D. J. (1998). Desmin cytoskeleton in muscle integrity and function. Subcell Biochem 31, 463-95. Cattelino, A., Liebner, S., Gallini, R., Zanetti, A., Balconi, G., Corsi, A., Bianco, P., beta- Wolburg, H., Moore, R., Oreda, B. et al. (2003). The conditional inactivation of the vascular and increased vascular catenin gene in endothelial cells causes a defective pattern fragility. J Cell Biol 162,1111-22. 91 Chen, J., Kubalak, S. W. and Chien, K. R. (1998a). Ventricular muscle-restricted targeting in cardiac chamber of the RXRalpha gene reveals a non-cell-autonomous requirement morphogenesis. Development 125, 1943-9. Chen, J., Kubalak, S. W., Minamisawa, S., Price, R. L., Becker, K. D., Hickey, R., Ross, J., Jr. and Chien, K. R. (1998b). Selective requirement of myosin light chain 2v in embryonic heart function. J Biol Chem 273,1252-6. Chien, K. R. (1999). Stress pathways and heart failure. Cell 98, 555-8. Chitaev, N. A., Leube, R. E., Troyanovsky, R. B., Eshkind, L. G., Franke, W. W. and Troyanovsky, S. M. (1996). The binding of plakoglobin to desmosomal Cadherins: patterns of binding sites and topogenic potential. J Cell Biol 133, 359-69. Chu, G., Haghighi, K. and Kranias, E. G. (2002). From mouse to man: understanding heart failure through genetically altered mouse models. J Card Fail 8, S432-49. Chu, P. H., Bardwell, W. M., Gu, Y., Ross, J., Jr. and Chen, J. (2000). FHL2 (SLIM3) is not essential for cardiac development and function. Mol Cell Biol 20, 7460-2. Clark, T. G., Conway, S. J., Scott, I. C, Labosky, P. A., Winnier, G., Bundy, J., Hogan, B. L. and Greenspan, D. S. (1999). The mammalian Tolloid-like 1 gene, Till, is necessary for normal septation and positioning of the heart. Development 126, 2631-42. Clevers, H. (2002). Inflating cell numbers by Wnt. Mol Cell 10,1260-1. of Codd, M. B., Sugrue, D. D., Gersh, B. J. and Melton, L. J., 3rd. (1989). Epidemiology idiopathic dilated and hypertrophic cardiomyopathy. A population-based study in Olmsted County, Minnesota, 1975-1984. Circulation 80, 564-72. Cottrell, G. T., Lin, R., Warn-Cramer, B. J., Lau, A. F. and Burt, J. M. (2003). Mechanism of v-Src- and mitogen-activated protein kinase-induced reduction of gap junction communication. Am J Physiol Cell Physiol 284, C511-20. Dabiri, G. A., Turnacioglu, K. K., Sanger, J. M. and Sanger, J. W. (1997). Myofibrillogenesis visualized in living embryonic cardiomyocytes. Proc Natl Acad Sei USA 94, 9493-8. Daleau, P., Boudriau, S., Michaud, M., Jolicoeur, C. and Kingma, J. G., Jr. (2001). Cx43 Preconditioning in the absence or presence of sustained ischemia modulates myocardial Pharmacol 371-8. protein levels and gap junction distribution. Can J Physiol 79, Daniel, J. M. and Reynolds, A. B. (1997). Tyrosine phosphorylation and cadherin/catenin function. Bioessays 19, 883-91. 92 Danielian, P. S., Muccino, D., Rowitch, D. H., Michael, S. K. and McMahon, A. P. a tamoxifen-inducible (1998). Modification of gene activity in mouse embryos in utero by form of Cre recombinase. Curr Biol 8,1323-6. Datwyler, D. A., Magyar, J. P., Busceti, V., Hirschy, A., Perriard, J. C, Bailey, J. E. and Eppenberger, H. M. (2001). Recombinant Sindbis virus allows expression and precise targeting of proteins of the contractile apparatus in cultured cardiomyocytes. Basic Res Cardiol 96, 630-5. Davis, M. A., Ireton, R. C. and Reynolds, A. B. (2003). A core function for pl20-catenin in Cadherin turnover. J Cell Biol 163, 525-34. de Melker, A. A., Desban, N. and Duband, J. L. (2004). Cellular localization and signaling activity of beta-catenin in migrating neural crest cells. Dev Dyn 230, 708-26. relevant is the de Simone, G. (2004). Concentric or eccentric hypertrophy: how clinically difference? Hypertension 43, 714-5. Delorme, B., Dahl, E., Jarry-Guichard, T., Maries, I., Briand, J. P., Willecke, K., Gros, D. and Theveniau-Ruissy, M. (1995). Developmental regulation of connexin 40 gene of the conduction Dev expression in mouse heart correlates with the differentiation system. Dyn 204, 358-71. D. and Delorme, B., Dahl, E., Jarry-Guichard, T., Briand, J. P., Willecke, K., Gros, Theveniau-Ruissy, M. (1997). Expression pattern of connexin gene products at the early developmental stages of the mouse cardiovascular system. Circ Res 81,423-37. Deschepper, C. F., Picard, S., Thibault, G., Touyz, R. and Rouleau, J. L. (2002). Characterization of myocardium, isolated cardiomyocytes, and blood pressure in WKHA and WKY rats. Am J Physiol Heart Circ Physiol 282, H149-55. Dihlmann, S., Klein, S. and Doeberitz, M. v. K. (2003). Reduction of {beta}-Catenin/T-Cell Transcription Factor Signaling by Aspirin and Indomethacin Is Caused by an Increased Stabilization of Phosphorylated {beta}-Catenin. Mol Cancer Ther 2, 509-516. C is Doble, B. W., Ping, P. and Kardami, E. (2000). The epsilon subtype of protein kinase required for cardiomyocyte connexin-43 phosphorylation. Circ Res 86, 293-301. H. C. Doyle, D. D., Goings, G. E., Upshaw-Earley, J., Page, E., Ranscht, B. and Palfrey, associated with (1998). T-cadherin is a major glycophosphoinositol-anchored protein noncaveolar detergent-insoluble domains of the cardiac sarcolemma. J Biol Chem 273, 6937- 43. Dupont, E., Matsushita, T., Kaba, R. A., Vozzi, C, Coppen, S. R., Khan, N., Kaprielian, in human R., Yacoub, M. H. and Severs, N. J. (2001). Altered connexin expression congestive heart failure. J Mol Cell Cardiol 33, 359-71. 93 Eghbali, M., Czaja, M. J., Zeydel, M., Weiner, F. R., Zern, M. A., Seifter, S. and Blumenfeld, O. O. (1988). Collagen chain mRNAs in isolated heart cells from young and adult rats. J Mol Cell Cardiol 20, 267-76. Ehler, E., Rothen, B. M., Hammerle, S. P., Komiyama, M. and Perriard, J. C. (1999). Myofibrillogenesis in the developing chicken heart: assembly of Z-disk, M-line and the thick filaments. J Cell Sei 112,1529-39. Ehler, E., Horowits, R., Zuppinger, C, Price, R. L., Perriard, E., Leu, M., Caroni, P., Sussman, M., Eppenberger, H. M. and Perriard, J. C. (2001). Alterations at the intercalated disk associated with the absence of muscle LIM protein. J Cell Biol 153,763-72. Ehler, E., Fowler, V. M. and Perriard, J. C. (2004). Myofibrillogenesis in the developing chicken heart: role of actin isoforms and of the pointed end actin capping protein tropomodulin during thin filament assembly. Dev Dyn 229, 745-55. Eisenberg, L. M. and Markwald, R. R. (1995). Molecular regulation of atrioventricular valvuloseptal morphogenesis. Circ Res 11, 1-6. Eppenberger, H. M. and Zuppinger, C. (1999). In vitro reestablishment of cell-cell contacts in adult rat cardiomyocytes. Functional role oftransmembrane components in the formation of new intercalated disk-like cell contacts. Faseb J13 Suppl, S83-9. Novel 3D Evans, H. J., Sweet, J. K., Price, R. L., Yost, M. and Goodwin, R. L. (2003). Heart Circ culture system for study of cardiac myocyte development. Am J Physiol Physiol 285, H570-8. Fässler, R. and Meyer, M. (1995). Consequences of lack of beta 1 integrin gene expression in mice. Genes Dev 9,1896-908. Fässler, R., Rohwedel, J., Maltsev, V., Bloch, W., Lentini, S., Guan, K., Gullberg, D., Hescheler, J., Addicks, K. and Wobus, A. M. (1996). Differentiation and integrity of 1 J Cell Sei 109 Pt cardiac muscle cells are impaired in the absence of beta integrin. ( 13), 2989-99. of Fast, V. G. and Kleber, A. G. (1993). Microscopic conduction in cultured strands neonatal rat heart cells measured with voltage-sensitive dyes. Circ Res 73, 914-25. Feng, Q., Song, W., Lu, X., Hamilton, J. A., Lei, M., Peng, T. and Yee, S. P. (2002). Development of heart failure and congenital septal defects in mice lacking endothelial nitric oxide synthase. Circulation 106, 873-9. Ferreira-Cornwell, M. C, Luo, Y., Narula, N., Lenox, J. M., Lieberman, M. and Radice, G. L. (2002). Remodeling the intercalated disc leads to cardiomyopathy in mice misexpressing Cadherins in the heart. J Cell Sei 115,1623-34. 94 Firulli, A. B., McFadden, D. G., Lin, Q., Srivastava, D. and Olson, E. N. (1998). Heart and extra-embryonic mesodermal defects in mouse embryos lacking the bHLH transcription factor Handl, Nat Genet 18,266-70. Fishman, M. C. and Chien, K. R. (1997). Fashioning the vertebrate heart: earliest embryonic decisions. Development 124, 2099-117. Fishman, M. C. and Olson, E. N, (1997). Parsing the heart: genetic modules for organ assembly. Cell 91, 153-6. MLP Flick, M. J. and Konieczny, S. F. (2000). The muscle regulatory and structural protein Cell Sei 113 Pt 1553-64. is a cytoskeletal binding partner of betal-spectrin. J ( 9), Foley, A. and Mercola, M. (2004). Heart induction: embryology to cardiomyocyte regeneration. Trends Cardiovasc Med 14,121-5. H. Freeman, K., Colon-Rivera, C, Olsson, M. C, Moore, R. L., Weinberger, D., Grupp, I. L., Vikstrom, K. L., Iaccarino, G., Koch, W. J. and Leinwand, L. A. (2001). Progression from hypertrophic to dilated cardiomyopathy in mice that express a mutant myosin transgene. Am J Physiol Heart Circ Physiol 280, HI 51-9. In Frenster, J. H. (1974). Ultrastructure and Function of Heterochromatin and Euchromatin. Academic Press. The Cell Nucleus, vol. Volume 1 (ed. D. Busch), pp. 565-580. New-York: Fromaget, C, el Aoumari, A. and Gros, D. (1992). Distribution pattern of connexin 43, a differentiation of mouse heart 51, gap junctional protein, during the myocytes. Differentiation 9-20. L. and Gallicano, G. I., Kouklis, P., Bauer, C, Yin, M., Vasioukhin, V., Degenstein, Fuchs, E. (1998). Desmoplakin is required early in development for assembly of desmosomes and cytoskeletal linkage. J Cell Biol 143,2009-22. G. Gassmann, M., Casagranda, F., Orioli, D., Simon, H., Lai, C, Klein, R. and Lemke, (1995). Aberrant neural and cardiac development in mice lacking the ErbB4 neuregulin receptor. Nature 378, 390-4. Goncharova, E. J., Kam, Z. and Geiger, B. (1992). The involvement of adherens junction components in myofibrillogenesis in cultured cardiac myocytes. Development 114,173-83. Gonen-Wadmany, M., Gepstein, L. and Seliktar, D. (2004). Controlling the Cellular Organization of Tissue-Engineered Cardiac Constructs. Ann NY Acad Sei 1015,299-311. Gourdie, R. G., Green, C. R., Severs, N. J. and Thompson, R. P. (1992). Immunolabelling and mature rat heart. Anat patterns of gap junction connexins in the developing Embryol (Berl) 185, 363-78. 95 J Granzier, H. and Labeit, S. (2002). Cardiac titin: an adjustable multi-functional spring. Physiol 541, 335-42. Gutstein, D. E., Morley, G. E., Tamaddon, H., Vaidya, D., Schneider, M. D., Chen, J., Chien, K. R., Stuhlmann, H. and Fishman, G. I. (2001). Conduction slowing and sudden arrhythmic death in mice with cardiac-restricted inactivation of connexin43. Circ Res 88, 333- 9. Gutstein, D. E., Liu, F. Y., Meyers, M. B., Choo, A. and Fishman, G. I. (2003). The organization of adherens junctions and desmosomes at the cardiac intercalated disc is independent of gap junctions. J Cell Sei 116, 875-85. Haegel, H., Larue, L., Ohsugi, M., Fedorov, L., Herrenknecht, K. and Kemler, R. (1995). 3529-37. Lack of beta-catenin affects mouse development at gastrulation. Development 121, Hall, D. G., Morley, G. E., Vaidya, D., Ard, M., Kimball, T. R., Witt, S. A. and Colbert, M. C. (2000). Early onset heart failure in transgenic mice with dilated cardiomyopathy. Pediatr Res 48, 36-42. Haq, S., Choukroun, G., Kang, Z. B., Ranu, H., Matsui, T., Rosenzweig, A., Molkentin, kinase- J. D., Alessandrini, A., Woodgett, J., Hajjar, R. et al. (2000). Glycogen synthase J Cell Biol 3beta is a negative regulator of cardiomyocyte hypertrophy. 151,117-30. Haq, S., Michael, A., Andreucci, M., Bhattacharya, K., Dotto, P., Walters, B., Woodgett, J., Kilter, H. and Force, T. (2003). Stabilization of beta-catenin by a Wnt-independent mechanism regulates cardiomyocyte growth. Proc Natl Acad Sei USA 100,4610-5. M. Harada, N., Miyoshi, H., Murai, N., Oshima, H., Tamai, Y., Oshima, M. and Taketo, of M. (2002). Lack of tumorigenesis in the mouse liver after adenovirus-mediated expression a dominant stable mutant of beta-catenin. Cancer Res 62,1971-7, Hari, L., Brault, V., Kleber, M., Lee, H. Y., Me, F., Leimeroth, R., Paratore, C, Suter, in U., Kemler, R. and Sommer, L. (2002). Lineage-specific requirements of beta-catenin neural crest development. J Cell Biol 159, 867-80. Hart, M., Concordet, J. P., Lassot, L, Albert, I., del los Santos, R., Durand, H., Perret, C, Rubinfeld, B., Margottin, F., Benarous, R. et al. (1999). The F-box protein beta-TrCP associates with phosphorylated beta-catenin and regulates its activity in the cell Curr Biol 9, 207-10. Dev Biol 203-16. Harvey, R. P. (1996). NK-2 homeobox genes and heart development. 178, Hatzfeld, M. and Nachtsheim, C. (1996). Cloning and characterization of a new armadillo of family member, p0071, associated with the junctional plaque: evidence for a subfamily closely related proteins. J Cell Sei 109 ( Pt 11), 2767-78. 96 Heim, R., Cubitt, A. B. and Tsien, R. Y. (1995). Improved green fluorescence. Nature 373, 663-4. Hemler, M. E. (1998). Integrin associated proteins. Curr Opin Cell Biol 10, 578-85. Hertig, C. M., Butz, S., Koch, S., Eppenberger-Eberhardt, M., Kemler, R. and IL Eppenberger, H. M. (1996a). N-cadherin in adult rat cardiomyocytes in culture. Spatio¬ involved in cell-cell contact and communication. Formation temporal appearance of proteins oftwo distinct N-cadherin/catenin complexes. J Cell Sei 109,11-20. M. Hertig, C. M., Eppenberger-Eberhardt, M., Koch, S. and Eppenberger, H. (1996b). N-cadherin in adult rat cardiomyocytes in culture. I. Functional role of N-cadherin and impairment of cell-cell contact by a truncated N-cadherin mutant. J Cell Sei 109,1-10. roles of Hertig, C. M., Kubalak, S. W., Wang, Y. and Chien, K. R. (1999). Synergistic 3-kinase neuregulin-1 and insulin-like growth factor-I in activation of the phosphatidylinositol pathway and cardiac chamber morphogenesis. J Biol Chem 214, 37362-9. W. and Hirota, H., Chen, J., Betz, U. A., Rajewsky, K., Gu, Y., Ross, J., Jr., Muller, is a critical event Chien, K. R. (1999). Loss of a gpl30 cardiac muscle cell survival pathway in the onset of heart failure during biomechanical stress. Cell 97,189-98. of the PI Hoess, R., Abremski, K. and Sternberg, N. (1984). The nature of the interaction Biol 761- recombinase Cre with the recombining site loxP. Cold Spring Harb Symp Quant 49, 8. N. Holmes, A., Wrenn, C. C, Harris, A. P., Thayer, K. E. and Crawley, J. (2002). tests for Behavioral profiles of inbred strains on novel olfactory, spatial and emotional 55-69. reference memory in mice. Genes Brain Behav 1, F. and Hornberger, L. K., Singhroy, S., Cavalle-Garrido, T., Tsang, W., Keeley, Rabinovitch, M. (2000). Synthesis of extracellular matrix and adhesion through beta(l) Circ Res 508-15. integrins are critical for fetal ventricular myocyte proliferation. 87, S. and Huang, C, Sheikh, F., Hollander, M., Cai, C, Becker, D., Chu, P. H., Evans, cardiac Chen, J. (2003). Embryonic atrial function is essential for mouse embryogenesis, morphogenesis and angiogenesis. Development 130, 6111-9. F. and Hughes, I. A., Lim, H. N., Martin, H., Mongan, N. P., Dovey, L., Ahmed, S. Endocrinol Hawkins, J. R. (2001). Developmental aspects of androgen action. Mol Cell 185, 33-41. Hunter, K., Welch, D. R. and Liu, E. T. (2003). Genetic background is an important determinant of metastatic potential. Nat Genet 34, 23-4; author reply 25. 97 Huristone, A. F., Haramis, A. P., Wienholds, E., Begthel, H., Korving, J., Van Eeden, F., Cuppen, E., Zivkovic, D., Plasterk, R. H. and Clevers, H. (2003). The Wnt/beta-catenin pathway regulates cardiac valve formation. Nature 425, 633-7. Hyer, J., Johansen, M., Prasad, A., Wessels, A., Kirby, M. L., Gourdie, R. G. and Mikawa, T. (1999). Induction of Purkinje fiber differentiation by coronary arterialization. Proc Natl Acad Sei USA 96,13214-8. Ikeguchi, M., Sato, N., Hirooka, Y. and Kaibara, N. (1998). Computerized nuclear morphometry of hepatocellular carcinoma and its relation to proliferative activity. J Surg Oncol 68,225-30. Imanaka-Yoshida, K. (1997). Myofibrillogenesis in precardiac mesoderm expiant culture. Cell Struct Fund 22, 45-9. Imanaka-Yoshida, K., Knudsen, K. A. and Linask, K. K. (1998). N-cadherin is required for the differentiation and initial myofibrillogenesis of chick cardiomyocytes. Cell Motil Cytoskeleton 39, 52-62. Inoue, H., Nojima, H. and Okayama, H. (1990). High efficiency transformation of Escherichia coli with plasmids. Gene 96,23-8. M. Ireton, R. C, Davis, M. A., van Hengel, J., Mariner, D. J., Barnes, K., Thoreson, A., Anastasiadis, P. Z., Matrisian, L., Bundy, L. M., Sealy, L. et al. (2002). A novel role for pi20 catenin in E-cadherin function. J Cell Biol 159,465-76. I. Isac, C. M., Ruiz, P., Pfitzmaier, B., Haase, H., Birchmeier, W. and Morano, (1999). Plakoglobin is essential for myocardial compliance but dispensable for myofibril insertion into adherens junctions. J Cell Biochem 72, 8-15. Ishiwata, T., Nakazawa, M., Pu, W. T., Tevosian, S. G. and Izumo, S. (2003). Developmental changes in ventricular diastolic function correlate with changes in ventricular myoarchitecture in normal mouse embryos. Circ Res 93, 857-65. Jang, S. K., Pestova, T. V., Hellen, C. U., Witherell, G. W. and Wimmer, E. (1990). Cap- independent translation of Picornavirus RNAs: structure and function of the internal ribosomal entry site. Enzyme 44,292-309. Janssens, B., Goossens, S., Staes, K., Gilbert, B., van Hengel, J., Colpaert, C, Bruyneel, beta-catenin- E., Mareel, M. and van Roy, F. (2001). alphaT-catenin: a novel tissue-specific binding protein mediating strong cell-cell adhesion. J Cell Sei 114, 3177-88. F. Julius, M. A., Schelbert, B., Hsu, W., Fitzpatrick, E., Jho, E., Fagotto, F., Costantini, and and Kitajewski, J. (2000). Domains of axin and disheveled required for interaction function in wnt signaling. Biochem Biophys Res Commun 276, 1162-9. 98 Kadokami, T., McTiernan, C. F., Kubota, T., Frye, C. S. and Feldman, A. M. (2000). Sex-related survival differences in murine cardiomyopathy are associated with differences in TNF-receptor expression. J Clin Invest 106, 589-97. Kang, D. E., Soriano, S., Frosch, M. P., Collins, T., Naruse, S., Sisodia, S. S., Leibowitz, G., Levine, F. and Koo, E. H. (1999). Presenilin 1 facilitates the constitutive turnover of beta-catenin: differential activity of Alzheimer's disease-linked PS1 mutants in the beta- catenin-signaling pathway. JNeurosci 19, 4229-37. Kaufmann, U., Zuppinger, C, Waibler, Z., Rüdiger, M., Urbich, C, Martin, B., Jockusch, B. M., Eppenberger, H. and Starzinski-Powitz, A. (2000). The armadillo repeat 4121-35. region targets ARVCF to cadherin-based cellular junctions. J Cell Sei 113 ( Pt 22), Keller, R. S., Shai, S. Y., Babbitt, C. J., Pham, C. G., Solaro, R. J., Valencik, M. L., Loftus, J. C. and Ross, R. S. (2001). Disruption of integrin function in the murine J myocardium leads to perinatal lethality, fibrosis, and abnormal cardiac performance. Am Pathol 158, 1079-90. Kim, H., Yoon, C. S. and Rah, B. (1999). Expression of extracellular matrix components fibronectin and laminin in the human fetal heart. Cell Struct Funct 24,19-26. Kirby, M. L. and Waldo, K. L. (1995). Neural crest and cardiovascular patterning. Circ Res 77,211-5. Kirchhoff, S., Nelles, E., Hagendorff, A., Kruger, O., Traub, O. and Willecke, K. (1998). Reduced cardiac conduction velocity and predisposition to arrhythmias in connexin40- deficient mice. Curr Biol 8, 299-302. Kisanuki, Y. Y., Hammer, R. E., Miyazaki, J., Williams, S. C, Richardson, J. A. and Yanagisawa, M. (2001). Tie2-Cre transgenic mice: a new model for endothelial cell-lineage analysis in vivo. Dev Biol 230, 230-42. Knoll, R., Hoshijima, M., Hoffman, H. M., Person, V., Lorenzen-Schmidt, I., Bang, M. L., Hayashi, T., Shiga, N., Yasukawa, H., Schaper, W. et al. (2002). The cardiac defective in a subset of mechanical stretch sensor machinery involves a Z disc complex that is human dilated cardiomyopathy. Cell 111, 943-55. Knoll, R., Hoshijima, M. and Chien, K. (2003). Cardiac mechanotransduction and implications for heart disease. JMol Med 81, 750-6. Koch, P. J. and Franke, W. W. (1994). Desmosomal Cadherins: another growing multigene family of adhesion molecules. Curr Opin Cell Biol 6, 682-7. N. Koike, M., Kose, S., Furuta, M., Taniguchi, N., Yokoya, F., Yoneda, Y. and Imamoto, but distinct molecular (2004). beta -catenin shows an overlapping sequence requirement interactions for its bi-directional passage through nuclear pores. J Biol Chem. 99 Komiyama, M., Soldati, T., von Arx, P. and Perriard, J. C. (1996). The intracompartmental sorting of myosin alkali light chain isoproteins reflects the sequence of J Cell Sei developmental expression as determined by double epitope-tagging competition. 109 ( Pt 8), 2089-99. Kong, Y., Flick, M. J., Kudla, A. J. and Konieczny, S. F. (1997). Muscle LIM protein 4750-60. promotes myogenesis by enhancing the activity of MyoD. Mol Cell Biol 17, Kong, Y., Shelton, J. M., Rothermel, B., Li, X., Richardson, J. A., Bassel-Duby, R. and Williams, R. S. (2001). Cardiac-specific LIM protein FHL2 modifies the hypertrophic 2731-8. response to beta-adrenergic stimulation. Circulation 103, Kostin, S., Hein, S., Bauer, E. P. and Schaper, J. (1999). Spatiotemporal development and distribution of intercellular junctions in adult rat cardiomyocytes in culture. Circ Res 85,154- 67. E. Kostin, S., Rieger, M., Dammer, S., Hein, S., Richter, M., Klovekorn, W. P., Bauer, P. and Schaper, J. (2003). Gap junction remodeling and altered connexin43 expression in the failing human heart. Mol Cell Biochem 242,135-44. Kruger, O., Plum, A., Kim, J. S., Winterhager, E., Maxeiner, S., Hallas, G., Kirchhoff, in S., Traub, O., Lamers, W. H. and Willecke, K. (2000). Defective vascular development connexin 45-deficient mice. Development 127, 4179-93. Kubalak, S. W., Miller-Hance, W. C, O'Brien, T. X., Dyson, E. and Chien, K. R. (1994). Chamber specification of atrial myosin light chain-2 expression precedes septation during murine cardiogenesis. J Biol Chem 269, 16961-70. Kuhbandner, S., Brummer, S., Metzger, D., Chambon, P., Hofmann, F. and Feil, R. (2000). Temporally controlled somatic mutagenesis in smooth muscle. Genesis 28, 15-22. Kumai, M., Nishii, K., Nakamura, K., Takeda, N., Suzuki, M. and Shibata, Y. (2000). 3501- Loss of connexin45 causes a cushion defect in early cardiogenesis. Development 127, 12. their Kuroda, S., Fukata, M., Nakagawa, M. and Kaibuchi, K. (1999). Cdc42, Racl, and cell-cell adhesion. Biochem effector IQGAP1 as molecular switches for cadherin-mediated Biophys Res Commun 262,1-6. Laemmli, U. K. (1970). Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227, 680-5. J. C. Lange, S., Auerbach, D., McLoughlin, P., Perriard, E., Schäfer, B. W., Perriard, muscle and Ehler, E. (2002). Subcellular targeting of metabolic enzymes to titin in heart may be mediated by DRAL/FHL-2. J Cell Sei 115,4925-36. 100 Le Corvoisier, P., Park, H. Y., Carlson, K. M., Donahue, M. P., Marchuk, D. A. and Arch Rockman, H. A. (2003a). Impact of genetic polymorphisms on heart failure prognosis. Mal Coeur Vaiss 96,197-206. Le Corvoisier, P., Park, H. Y. and Rockman, H. A. (2003b). Modifier genes and heart failure. Minerva Cardioangiol 51,107-20. R. and Lee, H. Y., Kleber, M., Hari, L., Brault, V., Suter, U., Taketo, M. M., Kemler, in Sommer, L. (2004). Instructive role of Wnt/beta-catenin in sensory fate specification neural crest stem cells. Science 303,1020-3. Lee, K. F., Simon, H., Chen, H., Bates, B., Hung, M. C. and Hauser, C. (1995). Requirement for neuregulin receptor erbB2 in neural and cardiac development. Nature 378, 394-8. Lee, Y., Song, A. J., Baker, R., Micales, B., Conway, S. J. and Lyons, G. E. (2000). for normal heart Circ Res Jumonji, a nuclear protein that is necessary development. 86,932-8. J. Clin. Leinwand, L. A. (2003). Sex is a potent modifier of the cardiovascular system. Invest. 112, 302-307. of the Leu, M., Ehler, E. and Perriard, J. C. (2001). Characterisation of postnatal growth murine heart. Anat Embryol (Berl) 204, 217-24. Nat Rev Genet Lewandoski, M. (2001). Conditional control of gene expression in the mouse. 2, 743-55. Li, F., Wang, X., Capasso, J. M. and Gerdes, A. M. (1996). Rapid transition of cardiac JMol Cell Cardiol myocytes from hyperplasia to hypertrophy during postnatal development. 28, 1737-46. recombinase Li, J., Chen, F. and Epstein, J. A. (2000). Neural crest expression of Cre directed by the proximal Pax3 promoter in transgenic mice. Genesis 26,162-4. C. Li, Q. Y., Newbury-Ecob, R. A., Terrett, J. A., Wilson, D. I., Curtis, A. R., Yi, H., is Gebühr, T., Bullen, P. J., Robson, S. C, Strachan, T. et al. (1997). Holt-Oram syndrome Nat Genet caused by mutations in TBX5, a member of the Brachyury (T) gene family. 15, 21-9. Lickert, H., Kutsch, S., Kanzler, B., Tamai, Y., Taketo, M. M. and Kemler, R. (2002). Formation of multiple hearts in mice following deletion of beta-catenin in the embryonic endoderm. Dev Cell 3,171-81. Liou, Y. C, Ryo, A., Huang, H. K., Lu, P. J., Bronson, R., Fujimori, F., Uchida, T., Hunter, T. and Lu, K. P. (2002). Loss of Pinl function in the mouse causes phenotypes resembling cyclinDl-null phenotypes. Proc Natl Acad Sei USA99,1335-40. 101 A. Liu, P., Wakamiya, M., Shea, M. J., Albrecht, U., Behringer, R. R. and Bradley, (1999). Requirement for Wnt3 in vertebrate axis formation. Nat Genet 22, 361-5. C. Lo, C. W., Cohen, M. F., Huang, G. Y., Lazatin, B. O., Patel, N., Sullivan, R., Pauken, and Park, S. M. (1997). Cx43 gap junction gene expression and gap junctional communication in mouse neural crest cells. Dev Genet 20,119-32. in LoRusso, S. M., Rhee, D., Sanger, J. M. and Sanger, J. W. (1997). Premyofibrils spreading adult cardiomyocytes in tissue culture: evidence for reexpression of the embryonic cells. Cell Motil program for myofibrillogenesis in adult Cytoskeleton 37,183-98. Louis, H. A., Pino, J. D., Schmeichel, K. L., Pomies, P. and Beckerle, M. C. (1997). Comparison of three members of the cysteine-rich protein family reveals functional 27484-91. conservation and divergent patterns of gene expression. J Biol Chem 272, M. Luo, Y., Ferreira-Cornwell, M., Baldwin, H., Kostetskii, I., Lenox, J., Lieberman, of and Radice, G. (2001). Rescuing the N-cadherin knockout by cardiac-specific expression N- or E-cadherin. Development 128, 459-69. for Luo, Y. and Radice, G. L. (2003). Cadherin-mediated adhesion is essential myofibril of the contractile J continuity across the plasma membrane but not for assembly apparatus. Cell Sei 116,1471-9. L. and R. P. Lyons, I., Parsons, L. M., Hartley, L., Li, R., Andrews, J. E., Robb, Harvey, (1995). Myogenic and morphogenetic defects in the heart tubes of murine embryos lacking the homeo box gene Nkx2-5. Genes Dev 9,1654-66. functional domains MacLean, H. E., Warne, G. L. and Zajac, J. D. (1997). Localization of in the androgen receptor. J Steroid Biochem Mol Biol 62, 233-42. and beta Maitra, N., Flink, I. L., Bahl, J. J. and Morkin, E. (2000). Expression of alpha integrins during terminal differentiation of cardiomyocytes. Cardiovasc Res 47, 715-25. A. Maretto, S., Cordenonsi, M., Dupont, S., Braghetta, P., Broccoli, V., Hassan, B., Volpin, D., Bressan, G. M. and Piccolo, S. (2003). Mapping Wnt/beta-catenin signaling Proc Natl Acad Sei USA 100, 3299- during mouse development and in colorectal tumors. 304. Markwald, R., Eisenberg, C, Eisenberg, L., Trusk, T. and Sugi, Y. (1996). Epithelial¬ mesenchymal transformations in early avian heart development. ActaAnat (Basel) 156,173- 86. G. E. and Marsh, J. D., Lehmann, M. H., Ritchie, R. H., Gwathmey, J. K., Green, Schiebinger, R. J. (1998). Androgen receptors mediate hypertrophy in cardiac myocytes. Circulation 98, 256-61. 102 Martin, B., Schneider, R., Janetzky, S., Waibler, Z., Pandur, P., Kuhl, M., Behrens, J., V. The von der Mark, K., Starzinski-Powitz, A. and Wixler, (2002). LIM-only protein FHL2 interacts with beta-catenin and promotes differentiation of mouse myoblasts. J Cell 5/0/159,113-22. Marvin, M. J., Di Rocco, G., Gardiner, A., Bush, S. M. and Lassar, A. B. (2001). Inhibition of Wnt activity induces heart formation from posterior mesoderm. Genes Dev 15, 316-27. Masuelli, L., Bei, R., Sacchetti, P., Scappaticci, I., Francalanci, P., Albonici, L., Coletti, in A., Palumbo, C, Minieri, M., Fiaccavento, R. et al. (2003). Beta-catenin accumulates intercalated disks of hypertrophic cardiomyopathic hearts. Cardiovasc Res 60, 376-87. Matsushita, T., Oyamada, M., Fujimoto, K., Yasuda, Y., Masuda, S., Wada, Y., Oka, T. and Takamatsu, T. (1999). Remodeling of cell-cell and cell-extracellular matrix interactions at the border zone of rat myocardial infarcts. Circ Res 85,1046-55. McCrea, P. D., Brieher, W. M. and Gumbiner, B. M. (1993). Induction of a secondary body axis in Xenopus by antibodies to beta-catenin. J Cell Biol 123, 477-84. McEwan, I, J. (2004). Molecular mechanisms of androgen receptor-mediated gene regulation: structure-function analysis of the AF-1 domain. Endocr Relat Cancer 11, 281-93. Meilhac, S. M., Kelly, R. G., Rocancourt, D., Eloy-Trinquet, S., Nicolas, J. F. and Buckingham, M. E. (2003). A retrospective clonal analysis of the myocardium reveals two phases of clonal growth in the developing mouse heart. Development 130, 3877-89. The Meilhac, S. M., Esner, M., Kelly, R. G., Nicolas, J. F. and Buckingham, M. E. (2004). clonal origin of myocardial cells in different regions of the embryonic mouse heart. Dev Cell 6, 685-98. Messerli, J. M., Eppenberger-Eberhardt, M. E., Rutishauser, B. M., Schwarb, P., von Arx, P., Koch-Schneidemann, S., Eppenberger, H. M. and Perriard, J. C. (1993). Remodelling of cardiomyocyte cytoarchitecture visualized by three-dimensional (3D) confocal microscopy. Histochemistry 100, 193-202. A. M. Milner, D. J., Taffet, G. E., Wang, X., Pham, T., Tamura, T., Hartley, C, Gerdes, and and Capetanaki, Y. (1999). The absence of desmin leads to cardiomyocyte hypertrophy cardiac dilation with compromised systolic function. J Mol Cell Cardiol 31,2063-76. Minamisawa, S., Gu, Y., Ross, J., Jr., Chien, K. R. and Chen, J. (1999a). A post- 2- transcriptional compensatory pathway in heterozygous ventricular myosin light chain deficient mice results in lack of gene dosage effect during normal cardiac growth or hypertrophy. J Biol Chem 274, 10066-70. 103 M. Minamisawa, S., Hoshijima, M., Chu, G., Ward, C. A., Frank, K., Gu, Y., Martone, E., Wang, Y., Ross, J., Jr., Kranias, E. G. et al. (1999b). Chrome phospholamban- sarcoplasmic reticulum calcium ATPase interaction is the critical calcium cycling defect in dilated cardiomyopathy. Cell 99, 313-22. Mohapatra, B., Jimenez, S., Lin, J. H., Bowles, K. R., Coveler, K. J., Marx, J. G., in the Chrisco, M. A., Murphy, R. T., Lurie, P. R., Schwartz, R. J. et al. (2003). Mutations and endocardial muscle LIM protein and alpha-actinin-2 genes in dilated cardiomyopathy fibroelastosis. Mol Genet Metab 80,207-15. Mohun, T. and Sparrow, D. (1997). Early steps in vertebrate cardiogenesis. Curr Opin Genet Dev 1, 628-33. Morin, P. J., Sparks, A. B., Korinek, V., Barker, N., Clevers, H., Vogelstein, B. and mutations Kinzler, K. W. (1997). Activation of beta-catenin-Tcf signaling in colon cancer by in beta-catenin or APC. Science 275,1787-90. Morkel, M., Huelsken, J., Wakamiya, M., Ding, J., van de Wetering, M., Clevers, H., Beta-catenin Taketo, M. M., Behringer, R. R., Shen, M. M. and Birchmeier, W. (2003). axis and regulates Cripto- and Wnt3-dependent gene expression programs in mouse mesoderm formation. Development 130, 6283-94. The CS5 Mould, A. P., Komoriya, A., Yamada, K. M. and Humphries, M. J. (1991). of fibronectin the 4 peptide is a second site in the IIICS region recognized by integrin alpha beta 1. Inhibition of alpha 4 beta 1 function by RGD peptide homologues. J Biol Chem 266, 3579-85. Mounier, N. and Sparrow, J. C. (1997). Structural comparisons of muscle and nonmuscle 89-97. actins give insights into the evolution of their functional differences. J Mol Evol 44, The Mulholland, D. J., Cheng, H., Reid, K., Rennie, P. S. and Nelson, C. C. (2002). translocation of androgen receptor can promote beta-catenin nuclear independently adenomatous polyposis coll J Biol Chem 277,17933-43. Müller, J. M., Isele, U., Metzger, E., Rempel, A., Moser, M., Pscherer, A., Breyer, T., Holubarsch, C, Buettner, R. and Schule, R. (2000). FHL2, a novel tissue-specific coactivator of the androgen receptor. Embo J19, 359-69. and beta - Nakamura, T., Sano, M., Songyang, Z. and Schneider, M. D. (2003). A Wnt- catenin-dependent pathway for mammalian cardiac myogenesis. Proc Natl Acad Sei USA 100, 5834-9. Nastasi, T., Bongiovanni, A., Campos, Y., Mann, L., Toy, J. N., Bostrom, J., Rottier, R., Hahn, C, Conaway, J. W., Harris, A. J. et al. (2004). Ozz-E3, a muscle-specific ubiquitin ligase, regulates beta-catenin degradation during myogenesis. Dev Cell 6, 269-82. 104 Nelson, W. J. and Nusse, R. (2004). Convergence of Wnt, beta-catenin, and Cadherin pathways. Science 303, 1483-7. Nicol, R. L., Frey, N. and Oison, E. N. (2000). From the sarcomere to the nucleus: role of genetics and signaling in structural heart disease. Annu Rev Genomics Hum Genet 1, 179-223. Nicol, R. L., Frey, N., Pearson, G., Cobb, M., Richardson, J. and Olson, E. N. (2001). Activated MEK5 induces serial assembly of sarcomeres and eccentric cardiac hypertrophy. Embo J 20, 2757-67. Nieset, J. E., Redfield, A. R., Jin, F., Knudsen, K. A., Johnson, K. R. and Wheelock, M. J. (1997). Characterization of the interactions of alpha-catenin with alpha-actinin and beta- catenin/plakoglobin. J Cell Sei 110 ( Pt 8), 1013-22. North, A. J., Bardsley, W. G., Hyam, J., Bornslaeger, E. A., Cordingley, H. C, Trinnaman, B., Hatzfeld, M., Green, K. J., Magee, A. I. and Garrod, D. R. (1999). 112 Pt 4325-36. Molecular map of the desmosomal plaque. J Cell Sei ( 23), O'Brien, T. X., Lee, K. J. and Chien, K. R. (1993). Positional specification of ventricular myosin light chain 2 expression in the primitive murine heart tube. Proc Natl Acad Sei USA 90,5157-61. Olson, E. N. and Srivastava, D. (1996). Molecular pathways controlling heart development. Science 272, 671-6. Orford, K., Crockett, C, Jensen, J. P., Weissman, A. M. and Byers, S. W. (1997). Serine phosphorylation-regulated ubiquitination and degradation of beta-catenin. J Biol Chem 212, 24735-8. of the Otey, C. A., Vasquez, G. B., Burridge, K. and Erickson, B. W. (1993). Mapping alpha-actinin binding site within the beta 1 integrin cytoplasmic domain. J Biol Chem 268, 21193-7. to the Ozawa, M. and Kemler, R. (1998). Altered cell adhesion activity by pervanadate due dissociation of alpha-catenin from the E-cadherin.catenin complex. J Biol Chem 273, 6166- 70. K. Ozcelik, C, Erdmann, B., Pilz, B., Wettschureck, N., Britsch, S., Hubner, N., Chien, R., Birchmeier, C. and Garratt, A. N. (2002). Conditional mutation of the ErbB2 (HER2) Acad Sei receptor in cardiomyocytes leads to dilated cardiomyopathy. Proc Natl USA99, 8880-5. lattice Pardo, J. V., Siliciano, J. D. and Craig, S. W. (1983). A vinculin-containing cortical in skeletal muscle: transverse lattice elements ("costameres") mark sites of attachment between myofibrils and sarcolemma. Proc Natl Acad Sei US A 80,1008-12. 105 Pashmforoush, M., Lu, J. T., Chen, H., Amand, T. S., Kondo, R., Pradervand, S., Evans, S. M., Clark, B., Feramisco, J. R., Giles, W. et al. (2004). Nkx2-5 pathways and congenital heart disease; loss of ventricular myocyte lineage specification leads to progressive cardiomyopathy and complete heart block. Cell 117, 373-86. Paul, D. L., Ebihara, L., Takemoto, L. J., Swenson, K. I. and Goodenough, D. A. (1991). currents in Connexin46, a novel lens gap junction protein, induces voltage-gated nonjunctional plasma membrane of Xenopus oocytes. J Cell Biol 115,1077-89. Pawlowski, J. E., Ertel, J. R., Allen, M. P., Xu, M., Butler, C, Wilson, E. M. and Wierman, M. E. (2002). Liganded androgen receptor interaction with beta-catenin: nuclear co-localization and modulation of transcriptional activity in neuronal cells. J Biol Chem 277, 20702-10. Peifer, M., Rauskolb, C, Williams, M., Riggleman, B. and Wieschaus, E. (1991). The in both segment polarity gene armadillo interacts with the wingless signaling pathway embryonic and adult pattern formation. Development 111, 1029-43. Peifer, M. (1996). Regulating cell proliferation: as easy as APC. Science 272, 974-5. Peifer, M. (1997). Beta-catenin as oncogene: the smoking gun. Science 275,1752-3. the Perriard, J. C, Hirschy, A. and Ehler, E. (2003). Dilated cardiomyopathy: a disease of intercalated disc? Trends Cardiovasc Med 13, 30-8. Peters, N. S., Green, C. R., Poole-Wilson, P. A. and Severs, N. J. (1993). Reduced content from and ischemic of connexin43 gap junctions in ventricular myocardium hypertrophied human hearts. Circulation 88, 864-75. Petrich, B. G., Gong, X., Lerner, D. L., Wang, X., Brown, J. H., Saffitz, J. E. and Wang, Y. (2002). c-Jun N-terminal kinase activation mediates downregulation of connexin43 in cardiomyocytes. Circ Res 91, 640-7. Piedra, J., Miravet, S., Castano, J., Palmer, H. G., Heisterkamp, N., Garcia de Herreros, A. and Dunach, M. (2003). pi20 Catenin-associated Fer and Fyn tyrosine kinases regulate beta-catenin Tyr-142 phosphorylation and beta-catenin-alpha-catenin Interaction. Mol Cell Biol 23,2287-97. in Polakis, P., Hart, M. and Rubinfeld, B. (1999). Defects in the regulation of beta-catenin colorectal cancer. Adv Exp Med Biol 470,23-32. Pulkkinen, L., Choi, Y. W., Simpson, A., Montagutelli, X., Sundberg, J., Uitto, J. and Mahoney, M. G. (2002). Loss of cell adhesion in Dsg3bal-Pas mice with homozygous deletion mutation (2079dell4) in the desmoglein 3 gene. J Invest Dermatol 119,1237-43. 106 R. Radice, G. L., Rayburn, H., Matsunami, H., Knudsen, K. A., Takeichi, M. and Hynes, Dev Biol O. (1997). Developmental defects in mouse embryos lacking N-cadherin. 181, 64-78. Rajewsky, K., Gu, H., Kuhn, R., Betz, U. A., Müller, W., Roes, J. and Schwenk, F. (1996). Conditional gene targeting. J Clin Invest 98, 600-3. T. Reaume, A. G., de Sousa, P. A., Kulkarni, S., Langille, B. L., Zhu, D., Davies, C, mice Juneja, S. C, Kidder, G. M. and Rossant, J. (1995). Cardiac malformation in neonatal lacking connexin43. Science 267,1831-4. Reed, K. E., Westphale, E. M., Larson, D. M., Wang, H. Z., Veenstra, R. D. and Beyer, E. C. (1993). Molecular cloning and functional expression of human connexin37, an 997-1004. endothelial cell gap junction protein. J Clin Invest 91, in Rezvani, M. and Liew, C. C. (2000). Role of the adenomatous polyposis coli gene product human cardiac development and disease. J Biol Chem 275,18470-5. Riggs, J. E., Bodensteiner, J. B. and Schochet, S. S., Jr. (2003). Congenital myopathies/dystrophies. Neurol Clin 21, 779-94; v-vi. Riley, P., Anson-Cartwright, L. and Cross, J. C. (1998). The Handl bHLH transcription factor is essential for placentation and cardiac morphogenesis. Nat Genet 18,271-5. Rimm, D. L., Koslov, E. R., Kebriaei, P., Cianci, C. D. and Morrow, J. S. (1995). Alpha the attachment of F-actin to l(E)-catenin is an actin-binding and -bundling protein mediating the membrane adhesion complex. Proc Natl Acad Sei USA 92, 8813-7. C. Ross, R. S., Pham, C, Shai, S. Y., Goldhaber, J. L, Fenczik, C, Glembotski, C, Ginsberg, M. H. and Loftus, J. C. (1998). Betal integrins participate in the hypertrophic Res response of rat ventricular myocytes. Circ 82,1160-72. Rothen-Rutishauser, B. M., Ehler, E., Perriard, E., Messerli, J. M. and Perriard, J. C. (1998). Different behaviour of the non-sarcomeric cytoskeleton in neonatal and adult rat cardiomyocytes. J Mol Cell Cardiol 30,19-31. Rottbauer, W., Saurin, A. J., Lickert, H., Shen, X., Burns, C. G., Wo, Z. G., Kemler, R., Kingston, R., Wu, C. and Fishman, M. (2002). Reptin and pontin antagonistically regulate heart growth in zebrafish embryos. Cell 111, 661-72. Roura, S., Miravet, S., Piedra, J., Garcia de Herreros, A. and Dunach, M. (1999). Regulation of E-cadherin/Catenin association by tyrosine phosphorylation. J Biol Chem 274, 36734-40. Rubinfeld, B., Robbins, P., El-Gamil, M., Albert, I., Porfiri, E. and Polakis, P. (1997). Stabilization of beta-catenin by genetic defects in melanoma cell lines. Science 275, 1790-2. 107 Ruiz, P., Brinkmann, V., Ledermann, B., Behrend, M., Grund, C, Thalhammer, C, of Vogel, F., Birchmeier, C, Gunthert, U., Franke, W. W. et al. (1996). Targeted mutation plakoglobin in mice reveals essential functions of desmosomes in the embryonic heart. J Cell Biol 135,215-25. Ruiz-Lozano, P. and Chien, K. R. (2003). Cre-constructing the heart. Nat Genet 33, 8-9. Ryo, A., Nakamura, M., Wulf, G., Liou, Y. C. and Lu, K. P. (2001). Pinl regulates turnover and subcellular localization of beta-catenin by inhibiting its interaction with APC. Nat Cell Biol 3, 793-801. T- Sacristan, M. P., Vestal, D. J., Dours-Zimmermann, M. T. and Ranscht, B. (1993). cadherin 2: molecular characterization, function in cell adhesion, and coexpression with T- cadherin and N-cadherin. JNeurosci Res 34,664-80. Sakanaka, C, Leong, P., Xu, L„ Harrison, S. D. and Williams, L. T. (1999). Casein kinase iepsilon in the wnt pathway: regulation of beta-catenin function. Proc Natl Acad Sei U SA 96,12548-52. Sakata, Y., Kamei, C. N., Nakagami, H., Bronson, R., Liao, J. K. and Chin, M. T. (2002). Ventricular septal defect and cardiomyopathy in mice lacking the transcription factor CHF1/Hey2. Proc Natl Acad Sei USA 99, 16197-202. Sato, S., Burgess, S. B. and Mcllwain, D. L. (1994). Transcription and motoneuron size. J Neurochem 63, 1609-15. Cre/lox Methods 14, Sauer, B. (1998). Inducible gene targeting in mice using the system. 381-92. human Schäfer, S., Koch, P. J. and Franke, W. W. (1994). Identification of the ubiquitous desmoglein, Dsg2, and the expression catalogue of the desmoglein subfamily of desmosomal Cadherins. Exp Cell Res 211, 391-9. Schmidt, M. M., Guan, K. and Wobus, A. M. (2001). Lithium influences differentiation cells in vitro. IntJDev and tissue-specific gene expression of mouse embryonic stem (ES) 5/0/45,421-9. Schneider, J. E., Bamforth, S. D., Farthing, C. R., Clarke, K., Neubauer, S. and Bhattacharya, S. (2003). Rapid identification and 3D reconstruction of complex cardiac malformations in transgenic mouse embryos using fast gradient echo sequence magnetic resonance imaging. JMol Cell Cardiol 35,217-22. Scholl, F. A., McLoughlin, P., Ehler, E., de Giovanni, C. and Schäfer, B. W. (2000). a half LIM domain induces DRAL is a p53-responsive gene whose four and protein product apoptosis. J Cell Biol 151, 495-506. 108 Seidman, J. G. and Seidman, C. (2001). The genetic basis for cardiomyopathy: from mutation identification to mechanistic paradigms. Cell 104, 557-67. Sen, A., Dunnmon, P., Henderson, S. A., Gerard, R. D. and Chien, K. R. (1988). Terminally differentiated neonatal rat myocardial cells proliferate and maintain specific differentiated functions following expression of SV40 large T antigen. J Biol Chem 263, 19132-6. Severs, N. J. (2000). The cardiac muscle cell. Bioessays 22,188-99. Severs, N. J., Coppen, S. R., Dupont, E., Yeh, H. I., Ko, Y. S. and Matsushita, T. (2004). Gap junction alterations in human cardiac disease. Cardiovasc Res 62, 368-77. Shai, S. Y., Harpf, A. E., Babbitt, C. J., Jordan, M. C, Fishbein, M. C, Chen, J., Omura, M., Leil, T. A., Becker, K. D., Jiang, M. et al. (2002). Cardiac myocyte-specific cardiac failure. Circ Res excision of the betal integrin gene results in myocardial fibrosis and 90,458-64. Sheng, Z., Pennica, D., Wood, W. I. and Chien, K. R. (1996). Cardiotrophin-1 displays early expression in the murine heart tube and promotes cardiac myocyte survival Development 122,419-28. Shimomura, O., Johnson, F. H. and Saiga, Y. (1961). Purification and properties of Cypridina luciferase. J Cell Comp Physiol 58,113-23. and Shimoyama, Y., Nagafuchi, A., Fujita, S., Gotoh, M., Takeichi, M., Tsukita, S. involvement Hirohashi, S. (1992). Cadherin dysfunction in a human cancer cell line: possible of loss of alpha-catenin expression in reduced cell-cell adhesiveness. Cancer Res 52, 5770-4. Shiraishi, I., Takamatsu, T. and Fujita, S. (1993). 3-D observation of N-cadherin expression during cardiac myofibrillogenesis of the chick embryo using a confocal laser scanning microscope. Anat Embryol (Berl) 187, 115-20. Simon, A. M. and Goodenough, D. A. (1998). Diverse functions of vertebrate gap junctions. Trends Cell Biol 8, 477-83. have Simon, A. M., Goodenough, D. A. and Paul, D. L. (1998). Mice lacking connexin40 cardiac conduction abnormalities characteristic of atrioventricular block and bundle branch block. Curr Biol 8, 295-8. Altered Smith, J. H., Green, C. R., Peters, N. S., Rothery, S. and Severs, N. J. (1991). in ischemic heart disease. An immunohistochemical patterns of gap junction distribution study of human myocardium using laser scanning confocal microscopy. Am J Pathol 139, 801-21. 109 K. Sohal, D. S., Nghiem, M., Crackower, M. A., Witt, S. A., Kimball, T. R., Tymitz, M., Penninger, J. M. and Molkentin, J. D. (2001). Temporally Regulated and Tissue-Specific Gene Manipulations in the Adult and Embryonic Heart Using a Tamoxifen-Inducible Cre Protein. Circ Res 89, 20-25. Cell Srivastava, D. and Olson, E. N. (1997). Knowing in your heart what's right. Trends in Biology 7,447-453. Srivastava, D., Thomas, T., Lin, Q., Kirby, M. L., Brown, D. and Olson, E. N. (1997). Regulation of cardiac mesodermal and neural crest development by the bHLH transcription factor, dHAND. Nat Genet 16,154-60. Srivastava, D. and Olson, E. N. (2000). A genetic blueprint for cardiac development. Nature 407,221-6. of Stefanou, D., Goussia, A. C, Arkoumani, E. and Agnantis, N. J. (2003). Expression vascular endothelial growth factor and the adhesion molecule E-cadherin in non-small cell lung cancer. Anticancer Res 23, 4715-20. Takeichi, M. (1995). Morphogenetic roles of classic Cadherins. Curr Opin Cell Biol 7, 619- 27. The cardiac Tanaka, M., Chen, Z., Bartunkova, S., Yamasaki, N. and Izumo, S. (1999). of essential for heart homeobox gene Csx/Nkx2.5 lies genetically upstream multiple genes development. Development 126,1269-80. Tarnavski, O., McMullen, J. R., Schinke, M., Nie, Q., Kong, S. and Izumo, S. (2004). of mouse models of Mouse cardiac surgery: comprehensive techniques for the generation human diseases and their application for genomic studies. Physiol Genomics 16, 349-60. adhesion kinase in Taylor, J. M., Rovin, J. D. and Parsons, J. T. (2000). A role for focal phenylephrine-induced hypertrophy of rat ventricular cardiomyocytes. J Biol Chem 275, 19250-7. of a Telo, P., Breviario, F., Huber, P., Panzeri, C. and Dejana, E. (1998). Identification novel Cadherin (vascular endothelial cadherin-2) located at intercellular junctions in endothelial cells. J Biol Chem 273, 17565-72. Fertilization to Theiler, K. (1972). The House Mouse: Development and Normal Stages from 4 weeks of Age. Berlin. 1 and Theis, D. G., Koch, P. J. and Franke, W. W. (1993). Differential synthesis of type Int JDev Biol type 2 desmocollin mRNAs in human stratified epithelia. 37,101-10. of cardiac Tokuyasu, K. T. and Maher, P. A. (1987). Immunocytochemical studies within titin myofibrillogenesis in early chick embryos. II. Generation of alpha-actinin dots 2795-801. spots at the time of the first myofibril formation. J Cell Biol 105, 110 Torres, M., Stoykova, A., Huber, O., Chowdhury, K., Bonaldo, P., Mansouri, A., Butz, its S., Kemler, R. and Gruss, P. (1997). An alpha-E-catenin gene trap mutation defines function in preimplantation development. Proc Natl Acad Sei USA94, 901-6. Torsoni, A. S., Constancio, S. S., Nadruz, W., Jr., Hanks, S. K. and Franchini, K. G. (2003). Focal adhesion kinase is activated and mediates the early hypertrophic response to stretch in cardiac myocytes. Circ Res 93,140-7. Tsukatani, Y., Suzuki, K. and Takahashi, K. (1997). Loss of density-dependent growth inhibition and dissociation of alpha-catenin from E-cadherin. J Cell Physiol 173, 54-63. Tsukazaki, K., Nikami, H., Shimizu, Y., Kawada, T., Yoshida, T. and Saito, M. (1995). of cold Chronic administration of beta-adrenergic agonists can mimic the stimulative effect tissue. J Biochem 96-100. exposure on protein synthesis in rat brown adipose (Tokyo) 117, Tsybouleva, N., Zhang, L., Chen, S., Patel, R., Lutucuta, S., Nemoto, S., DeFreitas, G., Entman, M., Carabello, B. A., Roberts, R. et al. (2004). Aldosterone, through novel the defect and the signaling proteins, is a fundamental molecular bridge between genetic cardiac phenotype of hypertrophic cardiomyopathy. Circulation 109, 1284-91. membrane Unwin, P. N. and Ennis, P. D. (1984). Two configurations of a channel-forming protein. Nature 307, 609-13. cell Utomo, A. R., Nikitin, A. Y. and Lee, W. H. (1999). Temporal, spatial, and type- specific control of Cre-mediated DNA recombination in transgenic mice. Nat Biotechnol 17, 1091-6. van der Flier, A., Gaspar, A. C, Thorsteinsdottir, S., Baudoin, C, Groeneveld, E., of the Mummery, C. L. and Sonnenberg, A. (1997). Spatial and temporal expression betalD integrin during mouse development. Dev Dyn 210,472-86. M. Boonstra, J., Defize, van der Heyden, M. A., Rook, M. B., Hermans, M., Rijksen, G., for Wnt L. H. and Destree, O. H. (1998). Identification of connexin43 as a functional target signalling. J Cell Sei 111 ( Pt 12), 1741-9. H. and D. O. A van der Ven, P. F., Bartsch, J. W., Gautel, M., Jockusch, Fürst, (2000). functional knock-out of titin results in defective myofibril assembly. J Cell Sei 113 ( Pt 8), 1405-14. T. Cardiac channels: van Veen, A. A., van Rijen, H. V. and Opthof, (2001). gap junction modulation of expression and channel properties. Cardiovasc Res 51,217-29. R. M. C, van Veen, T. A., van Rijen, H. V., Wiegerinck, F., Opthof, T., Colbert, in Clement, S., de Bakker, J. M. and Jongsma, H. J. (2002). Remodeling of gap junctions acid J Mol Cell Cardiol 34, 1411-23. mouse hearts hypertrophied by forced retinoic signaling. Ill and Wagner, J. D., Kaplan, J. R. and Burkman, R. T. (2002). Reproductive hormones cardiovascular disease mechanism of action and clinical implications. Obstet Gynecol Clin North Am 29,475-93. and Wang, X. and Gerdes, A. M. (1999). Chronic pressure overload cardiac hypertrophy failure in guinea pigs: III. Intercalated disc remodeling. JMol Cell Cardiol 31, 333-43. Watabe-Uchida, M., Uchida, N., Imamura, Y., Nagafuchi, A., Fujimoto, K., Uemura, T., Vermeulen, S., van Roy, F., Adamson, E. D. and Takeichi, M. (1998). alpha-Catenin- vinculin interaction functions to organize the apical junctional complex in epithelial cells. J Cell Biol 142, 847-57. Wei, Y., Renard, C. A., Labalette, C, Wu, Y., Levy, L„ Neuveut, C, Prieur, X., Flajolet, M., Prigent, S. and Buendia, M. A. (2003). Identification of the LIM protein FHL2 as a coactivator of beta-catenin. J Biol Chem 278, 5188-94. and Wheater, P. R., Burkitt, H. G. and Daniels, V. G. (1987). Functional histology: Text colour atlas. New-York: Elsevier Science Health Science div. Wiesel, P., Mazzolai, L., Nussberger, J. and Pedrazzini, T. (1997). Two-kidney, one clip and one-kidney, one clip hypertension in mice. Hypertension 29,1025-30. Wixler, V., Geerts, D., Laplantine, E., Westhoff, D., Smyth, N., Aumailley, M„ the Sonnenberg, A. and Paulsson, M. (2000). The LIM-only protein DRAL/FHL2 binds to cytoplasmic domain of several alpha and beta integrin chains and is recruited to adhesion complexes. J Biol Chem 275, 33669-78. Wollert, K. C, Taga, T., Saito, M., Narazaki, M., Kishimoto, T., Glembotski, C. C, Vernallis, A. B., Heath, J. K., Pennica, D., Wood, W. I. et al. (1996). Cardiotrophin-1 of sarcomeric units in activates a distinct form of cardiac muscle cell hypertrophy. Assembly series VIA gpl30/leukemia inhibitory factor receptor-dependent pathways. J Biol Chem 271, 9535-45. nomenclature. Wotjak, C. T. (2003). C57BLack/BOX? The importance of exact mouse strain Trends Genet 19, 183-4. N-cadherin- and Wu, J. C, Sung, H. C, Chung, T. H. and DePhilip, R. M. (2002). Role of 717- integrin-based costameres in the development of rat cardiomyocytes. J Cell Biochem 84, 24. of Wu, J. C, Tsai, R. Y. and Chung, T. H. (2003). Role of catenins in the development gap junctions in rat cardiomyocytes. J Cell Biochem 88, 823-35. Xavier-Vidal, R. and Mandarim-de-Lacerda, C. A. (1995). Cardiomyocyte proliferation and hypertrophy in the human fetus: quantitative study of the myocyte nuclei. Bull Assoc Anat (Nancy) 79,27-31. 112 in heart and Xu, W., Baribault, H. and Adamson, E. D. (1998). Vinculin knockout results brain defects during embryonic development. Development 125, 327-37. Yang, F., Li, X., Sharma, M., Sasaki, C. Y., Longo, D. L„ Lim, B. and Sun, Z. (2002). Linking beta-catenin to androgen-signaling pathway. J Biol Chem 277, 11336-44. Zhou, Z. X., Lane, M. V., Kemppainen, J. A., French, F. S. and Wilson, E. M. (1995). Specificity of ligand-dependent androgen receptor stabilization: receptor domain interactions influence ligand dissociation and receptor stability. Mol Endocrinol 9, 208-18. Zhurinsky, J., Shtutman, M. and Ben-Ze'ev, A. (2000). Differential mechanisms of LEF/TCF family-dependent transcriptional activation by beta-catenin and plakoglobin. Mol Cell Biol 20,4238-52. Zimmer, D. B., Green, C. R., Evans, W. H. and Gilula, N. B. (1987). Topological analysis of the major protein in isolated intact rat liver gap junctions and gap junction-derived single membrane structures. J Biol Chem 262, 7751-63. Zimmermann, W. H., Schneiderbanger, K., Schubert, P., Didie, M., Munzel, F., Heubach, J. F., Kostin, S., Neuhuber, W. L. and Eschenhagen, T. (2002). Tissue Circ Res engineering of a differentiated cardiac muscle construct. 90,223-30. 113 6 Appendix 6.1 Database of plasmids expressing Intercalated disc proteins and ES cells selection system ELhTTH N-cadherin-DsRed-N3 pDsRed-N3 weak 3.13.E ah2.5 sarc. alpha-actinin DsRed-N3 pDsRed-N3 aggregates, see Figure USA ah3r a-(human El) catenin pCDNA3.0 J. Morrow, New Haven, ah4 oc-catenin-Agel ah3r ah5.1 a-catenin-EGFP-Nl pEGFP-Nl works fine, see Figure 3.10.A ah6.1 a-catenin-DsRed-N 1 pDsRed-Nl aggregates ! ah7r aE(-mouse) catenin in pBS pBS R. Kemler, Freiburg, Germany ah8r ß-(mouse)catenin (pSKßTot) pBS R. Kemler, Freiburg, Germany ah9 ß-catenin-Agel ah8r 3.10.D ahlO.l ß-catenin-EGFP-Nl pEGFP-Nl nuclear localisation, see Figure ahll.l ß-catenin-DsRed-Nl pDsRed-Nl aggregates in nucleus ahl2r ß-(xenopus)catenin-EGFP-Nl R. Kypta, London, UK (ß-x-catenin) ahl3r plakoglobin (human) in pBS pBS W. Franke, Heidelberg, Germany (HPGCa2.1) ahl4 plakoglobin-Agel ahl3r ahl5.1 plakoglobin-EGFP-Nl pEGFP-Nl nuclear localisation, see Figure 3.10.J nucleus ahl6.1 plakoglobin-DsRed-N 1 p DsRed-Nl aggregates in ahl7r plakoglobin (xenopus) ? W. Franke, Heidelberg, Germany (PGXL6.3) ahl8r Vinculin-EGFP pGZ21 B. Geiger, Rehovot, Israel see Figure 3.11.A ah!9r US9-EGFP A. Beavis, Princeton, USA ah20r Connexin 32-GFP W.Evans, Cardiff UK no localisation at ICD, see Figure 3.12.A 114 ah21r Connexin 43-GFP L. Polontchouk, Zurich, Switzerland localisation at ICD, see Figure 3.12.G ah22r ß-actin promotor-ßötubulin A. Mattus, Basel, Switzerland EGFP ah23r CMV-human plakophilin-GFP pCS2mt M Klymkowsky, Boulder, USA ah24r pEGVP22 pEGFP-Cl C. Cardoso, Berlin, Germany ah25r pCS2+NLS pCS2 D. Turner, Seattle, USA Canada ah26r pGSK3ß-wt J. Woodgett, Toronto, Canada ah27r pGSK3ß Ala9 J. Woodgett, Toronto, ah28 EGFP-Nl+BglII EGFP-N1 intermediate construct ah29 pIRES kan pIRES modification of the pIRES plasmid from Clontech to change Ampr to Kanr ah30 pIRES modified ah29 change of the MCSI ah31.1 pIRES modified 2 ah30 introduction of Nrul site 5' of the (+Nru I) CMV promoter ah32.1 plRES-a-actinin-GFP ah31 works fine ah33.1 pIRES-a-actinin-DsRed ah31 aggregates ah34.1 pIRES-MLC3f-GFP ah30 MLC3f expressed 3' of the 1RES does not localize ! does ah35.1 pIRES-GFP-MLC3f ah32 MLC3f expressed 3' of the 1RES not localize ! ah36.1 p-a-catenin-DsRed-IRES-a- ah32.1 actinin-GFP ah37.1 p-a-catenin-GFP-IRES-a- ah33.1 actinin-DsRed with ah38 pNSCL 1RES ah33 exhange of the CMV promoter the cTnl minimal promoter ah39 pNSCL a-catenin-DsRed-IRES- ah40 very weak expression a-actinin-GFP ah40r N19RhoA pRK5 Myc A. Hall, London, UK ah41r N17Racl pRK5 Myc A. Hall, London, UK ah42r N17Cdc42 pRK5 Myc A. Hall, London, UK ah43r pBJ-Rat Axin-Myc pBJ-c-myc A. Kikuchi, Hiroshima, Japan ah44r APCwt pNeoBam B. Vogelstein, Baltimore, USA ah45r APC 1309C pNeoBam B. Vogelstein, Baltimore, USA 115 ah46 HA-APC A600C ? ah47.1 ß-catenin-Flag ahlO useless because the Flag antibody gives a high pericellular background ah48.1 ß-catenin A146N-Flag ah47 cfah47.1 ah49.1 ß-catenin M site-Flag ahlO cfah47.1 ah50 HA-APC in pCDNA3.1 pCDNA3.1 ah51 HA-APC 1309C inpCDNA3.1 pCDNA3.1 ah52.8 NLS-GFP pCS2+NLS strong nuclear localisation ah53.4 ß-catenin-HA pMCS-HA works... ah55 a-catenin-EGFP-Nl with Sal I pEGFP-Nl site ah56.1 p-a-catenin-GFP-IRES-a- ah36.1 double GFP signals actinin-GFP ah57.4 pFlins with insert pFlins pBS with 2x 2insulator sequences ah59 p-ot-catenin-GFP-IRES-a- ah33 actinin-GFP + Nru I ah60 p-oc-catenin-DsRed-IRES-a- ah33 actinin-GFP + NruI ah61 p-oc-catenin-GFP-IRES-a- ah33 actinin-DsRed + Nru I ah62 pIRES-a-actinin-GFP + Nru I ah33 ah63.1 Ins- pNSCL a-catenin-DsRed- ah57.4 huge plasmid ~ 20kbp IRES-a-actinin-GFP ah65 pNSCL a-catenin-GFP-IRES-a- ah33 actinin GFP ~ ah66 Ins-p-a-caten in-GFP-IRES-cc- ah57.4 huge plasmid 20kbp actinin-GFP ah67.3 Ins-p-a-catenin-DsRed-IRES-ct- ah57.4 huge plasmid ~ 20kbp actinin-GFP ah68.6 Ins-pIRES-a-actinin-GFP ah57.4 ah69.6 CMVenh-cTNI-IRES ah40 ah70 CMV-enh-cTNI-ct-catenin-GFP- ah69.6 IRES-a-actinin-GFP ah71 Vinculin-EGFP-Nl pEGFP-Nl ahl8r is better ah72,5 PGK puro in EGFP-N1 EGFP-N1 116 ah73.3 pCMVenh-cTNI-IRES-blast ah69.6 ah74.1 EGFP-C2+ Myr-Palm signal EGFP-C2 ah75.2 CMVenh-cTNI-a-actinin-GFP ah73.3 IRES-blast ah76 ECFP-MLC3f ECFP-C1 not detectable in green channel ah77 EYFP-MLC3f EYFP-C1 EYFP detected in green channel, yellowish ah80.7 Insulator + pGK puro ah57.4 ah81 pGEM + Cx43 chicken pGEM-T ah82r Mouse-axin-c-Myc F. Costantini, New York, USA ah83.7 Ins-CMVenh-cTNI-a-actinin- ah57.4 GFP-IRES-blast ah84.3 Ins-CMV-enh-cTNI-oc-catenin- ah57.4 GFP-IRES-a-actinin-GFP ah88.1 TOP-Luc pGL2 basic imitation of the famous pTOPFLASH plasmid ah89.1 FOP-Luc imitation of the famous pFOPFLASH plasmid ah90.1 TOP-LacZ pUTlll ß-galactosidase version of pTOPFLASH ah91.1 FOP-LacZ pUTl 11 ß-galactosidase version of pFOPFLASH ah92 Palm-GFP-Farn pEGFP-F does not work (FS4.2) ah97.1 pBCAT-HA 654F ah53.4 ah98.1 pBCAT-HA 142E ah53.4 ahl01.2 pACAT-HA pMCS-HA ah 102 cx40r-EGFP-Nl pEGFP-Nl see Figure 3.12.D ahl03 cx40r-HA pMCS-HA ah 104.1 cx45-EGFP-Nl pEGFP-Nl ahl05.1 cx45-HA pMCS-HA ahl06.1 pACAT 143st-EGFP ah5.1 ahl07.1 pACAT 632stop-EGFP ah5.1 ahl08r pCAG-EGFP-GPI ? G. Kondoh, Osaka, Japan ahl09r Dvl-1 pBS D. Sussman, Baltimore, USA 117 ahllOr Dvl-1 pVSK D. Sussman, Baltimore, USA ahlllr zyxin-EGFP ? R. Wehland, Braunschweig, Germany see Figure 3.11.G ahll2r paxillin-EGFP ? A. Horwitz, Urbana, USA see Figure 3.1 l.D ahll3r Wntl mouse cDNA pBS? R. Nusse, Stanford, USA ahll4r Wnt3a mouse cDNA pBS? R. Nusse, Stanford, USA ahllS Wntl-EGFP pEGFP-Nl ahllö.l Wntl-HA pMCS-HA ahll7.1 Wnt3a-EGFP pEGFP-Nl ahll8.1 Wnt3a-HA pMCS-HA ahll9r Dsg-cmyc ? K. Green, Chicago, USA ahl20r Dsc-cmyc ? K Green, Chicago,USA ahl21r pANCre R. Kemler, Freiburg, Germany ahl22r ß-catenin Probe A (for southern) pBS R. Kemler, Freiburg, Germany ahl23r 3x.MLCpLuc A.Cattini, Winnipeg, Canada ahl24r ANFpGL3 A. Zeiher, Frankfurt, Germany ahl25 dN-cadherin-GFP dN-cadherin weak expression VSV ah 126 dNdC-cadherin-GFP dNdC-cadherin weak expression VSV ahl28 Dsc2a-HA pMCS-HA weak expression ahl31r cx40r in pGEM-T pGEM-T ahl32.5 Dsg2-HA pMCS-HA weak expression ahl36.1 pMCS-HA-Nl to allow direct subcloning from pEGFP-Nl ahl38r cx45 cDNA J.-A. Haefliger, Lausanne, Switzerland ahl41 MLC2v 3x promoter pEGFP-Nl EGFP ahl42 Myomesin promoter EGFP pEGFP-Nl ok ah143 cTnl promoter EGFP pEGFP-Nl ok ahl53 PGK promoter EGFP pEGFP-Nl ok ah 145 Dsg2-EGFP pEGFP-Nl weak expression ah 146 Dsc2a-EGFP pEGFP-Nl weak expression ah 148 Shuttle vector for promoter cloning 118 ahl49.4 My-aact-GFP ah 142 ahl50.2 cTNI aact-GFP ahl43 ahlS1.3 MLC3x aact-GFP ahl41 ahl54 MA2.3-DsRedl pDsRedl-Nl sarcomeres are killed ! ahl55r pHTNC pTriEx-1 F. Edenhofer, Cologne, Germany The transduction of the Cre protein tested in ROSA26R isolated NMC was very weak ahl56r pREP4 D. Auerbach, Zurich, Switzerland ahl61 selection plasmid intermediate ah 168 selection plasmid cf Thesis of R. Bugorsky ahl69r pRFP-Nl RTsien, San Diego, USA monomeric DsRed protein ah 170.6 pRFP-N3 ahl69r ahl71.5 a.actinin-RFP ahl70.6 localisation is fine but RFP is weaker than DsRed or EGFP see Figure 3.13.H ahl72 a-catenin-RFP ahl69r same remark as for ah 171.5 ah173 pMLC3f-IRES-a-actinin-RFP ah31.1 see Figure 3.14.A,B ah 174 p-a-catenin-RFP-IRES-oc- ah32.1 see Figure 4.12.D,E actinin-GFP 119 Acknowledgements I would like to take here the opportunity to thank all the people involved in this thesis: Professor Jean-Claude Perriard for giving me the opportunity to work in his lab, his support and helpful comments on my work during my PhD. The Scientific assistant, Postdoctoral and PhD students of the Perriard group past and present: Evelyne Perriard, Elisabeth Ehler, Irina Agarkova, Preeti Ahuja, Mohamed Nemir, Jaya Krishnan, Dany Auerbach, Martin Leu, Pierre Giro, Stephan Lange, Ruslan Bugorsky, Roman Schbnauer, as well as Franziska Schatzmann who contributed to this work during her diploma thesis. The external helpers: Adrien Croquelois and Thierry Pedrazzini from the University Hospital of Lausanne Nicolas Lindegger and Ernst Niggli from the University of Bern Shoumo Bhattacharya and co-workers from the University of Oxford and Zurich I would like also to thank my girl friend, my family, my friends from Neuchâtel for their constant support during the "PhD years". whole for I will do my best to show to all of them, in a more individual way, my gratitude their precious love and friendship. With all my heart, thank you ! ! ! ALain 120 Curriculum Vitae Alain Hirschy Schwandenacker 44 Home +41 -44-303-1518 8052 Zurich Cell+41-78-647-2614 Switzerland [email protected] Personal Information Nationality: Swiss Date of birth: April 15,1975 Marital status: Single Career Objective A challenging position in an innovating company where I can apply and develop my scientific skills to the benefit of my employer, my colleagues and myself. Professional Experience March 1999-August 1999 Research Trainee Novartis Animal Health, Fribourg, Switzerland for the • Participated in the implementation of high throughput screening (HTS) assays development of new drugs against sheep endoparasites. Education September 1999-September 2004 Ph.D. in Cell Biology, Group of Prof. Dr J.C. Perriard Swiss Institute of Technology ETH Zurich, Switzerland Dissertation: Integration of myofibrils in the developing heart and challenges on the intercalated disc stability. October 1994-November 1998 Master of Science in Biochemistry and Parasitology passed with Honors University of Neuchâtel, Switzerland Thesis: Development of specific polyclonal antibodies for the identification of the histone HI in Trypanosmoma b. brucei. Additional Coursework April 2004 Introductory Course in Laboratory Animal Science University of Zurich, Switzerland June 2003 Course: "Recent Developments in Transgenic Research" University Hospital, Lausanne, Switzerland 121 Technical Skills Molecular Biology techniques: DNA and RNA isolation, PCR, RT-PCR, cloning, sequencing, non-radioactive Northernblotting. Protein Biochemistry techniques: bacterial protein expression, SDS-PAGE, immunoblotting. Cell culture techniques: isolation of cardiac cells, transfection of cell lines and primary cells, Adenovirus-mediated gene transfer, culture and differentiation of ES cells. Strong experience in immunofluorescence microscopy including confocal microscopy. and 3 years of practice with laboratory mice including breeding, time-mating genotyping. IT Skills Familiar with PC and Macintosh computer systems, wide use of Microsoft Office (Word, Excel, Powerpoint) and Adobe (Photoshop, Acrobat) programs. Administration of Filemaker databases. Expertise in Molecular Biology software, especially programs for DNA analysis (DNA Star package), plasmid construction map (Vector NTI) and GenBank searches (BLASTN). Language Skills French: mother tongue German: advanced level (5 years in the German-speaking part of Switzerland) English: fluent Italian and Spanish: basic level Publications Datwyler, D. A., Magyar, J. P., Busceti, V., Hirschy, A., Perriard, J. C, Bailey, J. E. and Eppenberger, H. M. (2001). Recombinant Sindbis virus allows expression and precise targeting of proteins of the contractile apparatus in cultured cardiomyocytes. Basic Res Cardiol 96, 630-5. Perriard, J. C, Hirschy, A. and Ehler, E. (2003). Dilated cardiomyopathy: a disease of the intercalated disc? Trends Cardiovasc Med 13, 30-8. Key Presentations "The heart specific deletion of beta-catenin" oral presentation at the 4th International Workshop on Cardiomyocyte Cell Biology: Stem Cells and Cellular Mechanisms, Ascona, Switzerland, April 2003. Poster presentations at the Annual Meeting of the Swiss Cardiology Network, Interlaken, Switzerland, June 2002 and September 2004. 122