Article
Reference
Remodeling of gap junctions in mouse hearts hypertrophied by forced retinoic acid signaling
VAN VEEN, Toon A B, et al.
Abstract
Beta-MHC-hRARalpha transgenic mice express a constitutively active (truncated) form of the human retinoic acid receptor which triggers development of dilated cardiomyopathy. In those hearts, we studied expression of gap junction proteins in relation to electrical impulse propagation.
VAN VEEN, Toon A B, et al . Remodeling of gap junctions in mouse hearts hypertrophied by forced retinoic acid signaling. Journal of Molecular and Cellular Cardiology , 2002, vol. 34, no. 10, p. 1411-1423
PMID : 12393001
DOI : 10.1006/jmcc.2002.2102
Available at:
http://archive-ouverte.unige.ch/unige:155571
Disclaimer: layout of this document may differ from the published version.
1 / 1
doi:10.1006/jmcc.2002.2102, available online at http://www.idealibrary.com on
Remodeling of Gap Junctions in Mouse Hearts Hypertrophied by Forced Retinoic Acid Signaling
Toon A. B. van Veen
1, Harold V. M. van Rijen
1, Rob F. Wiegerinck
2, Tobias Opthof
1, Melissa C. Colbert
3, Sophie Clement
4,
Jacques M. T. de Bakker
5,6, Habo J. Jongsma
11
Department of Medical Physiology, University Medical Center Utrecht, The Netherlands;
2
Department of Experimental Cardiology, Academical Medical Center, Amsterdam, The Netherlands;
3
Division of Molecular Cardiovascular Biology, Children's Hospital Medical Center, Cincinnati, USA;
4
Department of Pathology, University of Geneva, Geneva, Switzerland;
5
Interuniversity Cardiology Institute of Netherlands, Utrecht, The Netherlands;
6
Department of Cardiology, University Medical Center Utrecht, The Netherlands
(Received 26 April 2002,accepted for publication 5 August 2002)
T. A. B.VANVEEN, H. V. M.VANRIJEN, R. F. WIEGERINCK, T. OPTHOF, M. C. COLBERT, S. CLEMENT, J. M. T.DEBAKKER AND
H. J. JONGSMA. Remodeling of Gap Junctions in Mouse Hearts Hypertrophied by Forced Retinoic Acid Signaling.
Journal of Molecular and Cellular Cardiology(2002)34, 1411±1423.Background:b-MHC-hRARatransgenic mice express a constitutively active (truncated) form of the human retinoic acid receptor which triggers development of dilated cardiomyopathy. In those hearts, we studied expression of gap junction proteins in relation to electrical impulse propagation.Methods and Results: As compared to wildtype mice, hearts of 4±6 month old mice with 7±12 inserted hRARacopies are marked by an increased heart weight/body weight- and heart weight/tibia length ratio. 3-extremity lead ECGs revealed prolongation of the Q±j interval suggesting delayed ventricular activation.
Mapping of electrical activity of epi- and endocardial left ventricular free wall revealed activation delay, increased heterogeneity in conduction and regional conduction block. Ventricular tachycardia's did not occur spontane- ously nor could be induced by ventricular pacing. Immunohistochemical analysis showed profound and heteroge- neous redistribution and down-regulation of the gap junction protein connexin43 (Cx43) in the left ventricular free wall. Here, hRARaexpression induced re-expression of the hypertrophic markersa-skeletal actin andb-MHC, and in 3 out of 10 severely affected mice, re-expression of Cx40. Concomitant with changes in expression/
distribution of Cx43, changes in expression and distribution ofb-catenin and N-cadherin (two other intercalated disk associated proteins) were observed. Conclusions: b-MHC-hRARa transgenic hearts show heterogeneous re-expression of (early) sarcomeric genes while expression of connexin43, N-cadherin andb-catenin is down- regulated. We postulate that the resulting aberrant ventricular activation does not trigger development of lethal arrhythmias due to the small size of remaining healthy ventricular tissue where the transgene is not expressed.
#2002 Published by Elsevier Science Ltd.
Key Words: Retinoic acid receptor; Gap junction; Connexin; Mouse heart; Conduction; Hypertrophy;
Remodeling.
Please address all correspondence to: Toon A. B. van Veen, PhD, Department of Medical Physiology, University Medical Center Utrecht, PO Box 85060, 3508 AB Utrecht, The Netherlands. Tel:31302538900; Fax:31302539036; E-mail: [email protected] 0022±2828/02/10141113 $35.00/0 #2002 Published by Elsevier Science Ltd.
Introduction
Dilated cardiomyopathy (DCM) is a multifactorial cardiac disease with a poor clinical prognosis and is characterized by cardiac dilatation and myocar- dial dysfunction.1Although the onset of the disease can be diverse, cellular changes in expression/
function of sarcomeric proteins, ion channels and gap junction channels are consistently reported.2±4 Previously, Colbertet al. described the generation of transgenic mice expressing a constitutively active form of the human retinoic acid receptor (hRARa) under control of theb-MHC promoter.5RARs belong to the nuclear receptor superfamily which upon binding of steroid-like molecules (like retinoic acid) act as ligand activated transcription factors. Upon activation, RARs dimerize and regulate expression of genes which contain retinoic acid responsive ele- ments in their proximal promoter.6 Retinoic acid (RA) signaling is important for both cardiac devel- opment and differentiation into adult muscle cells.7,8 Evidence is accumulating showing that changes in RA homeostasis, e.g., induced by RA embryopathy, evoke severe malformations during cardiogenesis.9 Additionally, RA dependent signal transduction appears to preserve the normal differentiated pheno- type of cardiomyocytes by antagonizing the effect of various hypertrophic stimuli.10
Inb-MHC-hRARamice, forced activity of hRARa induced a DCM with pathophysiological characteris- tics as found in classical forms of DCM; severe impair- ment of cardiac function and a reduced contractility while endstage failing hearts showed biventricular chamberdilatationandleftatrialthrombosis.5Before cardiac dysfunction becomes overt, early alterations have been observed in the mRNA levels of genes associated with cardiac hypertrophy.11One of the differentially expressed genes is connexin43 (Cx43), a gap junction protein. Gap junctions are agglo- merates of multiple intercellular channels which in heart facilitate propagation of the electric impulse from cell to cell. They may be composed of Cx40, Cx43 and Cx45 and are primarily located on the longitudinal cell-edges in specialized structures called intercalated disks (IDs), (for review12).
Alterations in expression level of connexins and subcellular redistribution of gap junctions have been reported in various forms of cardiac disease,13±16 and are associated with an increased propensity to develop cardiac arrhythmias.17 In this study we show that in hypertrophic b-MHC-hRARa mouse hearts, expression/distribution of gap junctions is severely affected which correlates to aberrancies in electrical activation.
Materials and Methods
Animals
bMHC-hRARa-LacZ transgenic mice express a con- stitutively active (by truncation) human retinoic acid receptor under control of a b-MHC promoter.
Mice were generated in a FVB/N background at the Children's Hospital Research Foundation, Cincinnati, USA as described previously.5Breeding of mice originating from line 30 (12 inserted hRAR copies) was performed at the animal facility, Utrecht University, Utrecht, The Netherlands. To exclude unrelated genetic disorders, primary litters of imported transgenic mice were back-crossed with FVB/N wildtype mice. All animals used were 4±6 months of age. The study conformed to the guiding principles of the American Physiological Society.
Experimental design
Mice were genotyped by Southern/Dot blotting of genomic DNA (tail tips). After anesthesia, body weight was determined, body surface ECGs were recorded and tibia length was measured. Next, the heart was excised, prepared and mounted on a Lan- gendorff perfusion system to perform epicardial and endocardial mapping. At the end of the experiment, heart wet weight was assessed and hearts were snap frozen for histochemistry (complete hearts) or for protein isolation (hearts separated into compart- ments). Thin sections were used for immuno¯uor- escence microscopy or ®xed and stained to determine b-galactosidase activity. Separated ven- tricles were minced and protein was isolated for SDS-PAGE and Western blotting.
Genotyping
Genomic DNA was isolated from tail-tips according to Laird et al.,18 overnight digested with BamHI/
Bgl-II and separated (10mg) on a 0.8% agarose gel. By Southern blotting, DNA was transferred to Z-probe membrane (Biorad), ®xed and probed with a 32P labeled 1 kB probe derived from the LacZ insert. Since one speci®c product was produced (Fig. 1A), further characterization of mice was per- formed using Dot-blot analysis. Signals were visua- lized (phospho-imager, STORM 820, Molecular Dynamics, Sunnyvale, CA, USA) and quanti®ed with Image Quant software (Molecular Dynamics).
Copynumbers were calculated by comparing signal
intensities from samples with de®ned amounts of Bgl-II digested vector DNA (modi®ed pKS vector with a 3 kB LacZ fragment cloned in between two Bgl-II sites) representing increasing amounts of inserted copies (samples supplemented with 10mg wildtype mouse DNA). After primary development, blots were stripped and incubated with a probe raised against GAPDH. Calculated copynumbers were corrected for differences in GAPDH.
Preparation of the hearts
Body weight (BW) of mice was measured after which they were anaesthetized by a intraperitoneal injection of urethane (2 g/kg bodyweight), which minimally affects cardiac electrophysiology.19 3-Extremity lead ECGs were recorded and signals
were analyzed using custom-made software. Pre- paration of the hearts and Langendorff perfusion was performed as described before.20Finally, heart (wet) weight (HW) and tibia length (TL) of the left hind limb were measured and hearts were snap frozen in liquid nitrogen.
Recording of electrograms
Extracellular electrograms were recorded with a 247 point multi-terminal electrode as described before.20 Electrode terminals were arranged in a 19 by 13 grid (inter-electrode distance of 0.3 mm). Unipolar electrograms were acquired using a custom built 256-channel data-acquisition system. Epicardial recordings of the ventricles were made in sinus and paced rhythm. To determine spread of electrical Figure 1 (A) Genotyping of the mice with Southern blot analysis. (B) Cartoon of the heart with indicated the affected region in the LVFW. (C±E) bar20mm, LacZ staining (arrow-heads) of thin sections from the LVFW of hearts in group I (C), group II (D) and group III (E). (F, G) Bar plots of HW/BW ratios (F), HW/TL ratios (G), error bars indicate SEM.
*I, *IIsigni®cantly different from group I and group II.
activation during basic and premature stimulation, the epicardium was stimulated from two adjacent central electrodes in the grid at a cycle-length of 100 ms. Every sixteenth stimulus was followed by one premature stimulus. Starting at 90 ms the coupling interval of the extra stimulus was reduced in steps of 5 ms until conduction block occurred. To measure conduction velocity at the endocardial free wall, the left atrium was removed and a custom built grid with eight electrodes was inserted through the mitral valves. The electrode terminals were posi- tioned against the endocardial surface of the left ventricular free wall (LVFW) and unipolar record- ings were made during programmed stimulation from the two most apical electrodes of the grid.
Data analysis
Activation maps were constructed from the activa- tion times using custom written software based on Matlab (The Mathworks Inc.). Maximal negative dV/dtin the unipolar electrograms was selected as the time of local activation. To quantify zones of conduction delay and/or block, we determined dis- persion of conduction as described by Lammers et al.21In short, for each recording site the largest difference between activation times at the four sur- rounding sites is determined. By binning the data, a histogram of the maximum delay values is gener- ated from which the median (P50), the absolute inhomogeneity (P95ÿP5) and inhomogeneity index ((P95ÿP5)/P50) are determined. Multiple group comparisons were performed using ANOVA with Bonferoni post-hoc analysis. Two group compari- sons were performed using unpairedt-tests. Values are meanSEM.P-values0.05 were considered as statistically signi®cant.
Immunohistochemistry
Frozen hearts were serially sectioned in frontal plane to produce ``four chamber view'' sections of 10mm thickness. They were mounted on aminopro- pyltriethoxysilane (AAS) coated glass slides. For immunolabeling, sections were permeabilizsed in 0.2% Triton X-100 (1 h), blocked in 2% bovine serum albumin (BSA, 30 min) and incubated over- night in presence of 10% normal goat serum (NGS) with primary antibodies. After blocking again with 2% BSA (30 min) secondary labeling was performed with appropriate Texas Red (TR)- or ¯uorescein isothiocyanate (FITC)-conjugated antibodies (2 h in presence of 10% NGS). All chemicals were dissolved in phosphate buffered saline (PBS) which was also
used to wash the sections in between the subsequent incubations. Finally, sections were mounted in Vectashield (Vector Laboratories) and examined with a Nikon Optiphot-2 light microscope equipped for epi¯uorescence.
LacZ staining
Expression of the transgene could be assessed by LacZ staining because of the insertion of b- galactosidase DNA in the transgenic construct.
Thin sections were post-®xed with 2% paraformal- dehyde/0.2% glutaraldehyde (5 min, 4C), rinsed in PBS, and incubated with 1 mg/ml X-Gal, 5 mM
K4FeCN63H2O, 5 mM K3FeCN6, 2 mM MgCl2 6H2O in PBS (overnight, 37C). Finally, sections were dehydrated, embedded in Entelan and analyzed by classical light microscopy.
Protein isolation, SDS-PAGE an Western blotting
Total cellular protein was isolated from four sepa- rated fractions: left ventricular free wall (LVFW), right ventricular free wall (RVFW), interventricular septum (IVS) and pooled atria (A). Tissue was pul- verized in a liquid nitrogen cooled mortar and rapidly transferred to RIPA-buffer (50 mM Tris pH 7.4, 150 mM NaCl, 1% nonidet P40, 1% DOC, 0.1% SDS, 2 mM PMSF and a cocktail of protease inhibitors). Protein content of the supernatant was assessed according to Lowry.22 Equal amounts (50mg/lane) of each sample were separated on 10% SDS-polyacrylamide gels and electrophoreti- cally transferred to nitrocellulose membrane (0.45mm, Biorad). Equality of protein transfer was assessed by Ponceau S staining. Prior to pri- mary antibody incubation (overnight, 4C, in 0.1%
BSA/0.1% Tween20/PBS), the nitrocellulose mem- brane was blocked with 5% dried milk powder (1 h, RT, in 0.1% Tween20/PBS). Next morning, the membrane was washed (35 min, 0.1%
Tween20/PBS), incubated with Horse-Radish-Per- oxidase conjugated secondary antibody (1 h, 4C, 0.1% BSA/0.1% Tween20/PBS), and washed again (65 min, 0.1% Tween20/PBS). Signals were visualized using an Enhanced Chemo Luminiscence reagent (ECL, Amersham) according to the instruc- tions of the manufacturer and exposure to XB-1 ®lm (Kodak).
Antibodies
The following commercially available anti- bodies were used; rabbit polyclonal Cx40 (Alpha
Diagnostics), mouse monoclonal Cx43 and mouse monoclonalb-catenin (Transduction Laboratories), rabbit polyclonal Cx43 (Zymed), mouse monoclonal desmin (Sanbio), rabbit polyclonal N-cadherin and mouse monoclonala-actinin(Sigma), rabbitpolyclo- nal ANF (Biotrend). A mouse monoclonal antibody againstb-MHC was a kind gift of Dr A.F.M Moorman, Dept. Anatomy/Embryology, Academic Medical Center, Amsterdam, The Netherlands. A rabbit poly- clonal antibody against a-skeletal actin was pro- duced at the Department of Pathology, University of Geneva, Switzerland.23 Secondary antibodies (Horse Radish Peroxidase (HRP), Texas red (TR) or Fluorescein-Isothiocyanate (FITC)-conjugated) were purchased from Jackson Laboratories.
Results
Characterization of the mice
In our mouse breeding program, no differences within the population were observed in number of offspring nor in loss of animals after birth that could be associated with expression of high amounts of inserted transgenic construct. Genotyping of the offspring was initially performed with Southern hybridization. In Figure 1A genotyping of 5 mice is shown. A 3.3 kB signal of predicted size,24was found in lanes 1, 3 and 5 while the positive control (pKS vector with LacZ insert) revealed as calculated, a 3 kB band. Since the probe appeared to be speci®c, genotyping was simpli®ed using Dot-blot analysis.
Insertion level of transgenic constructs ranged from 0 to 12 copies.
Mice were categorized in three groups (n10 in each group) based on copynumber of the inserted transgenic construct: group I (control): 0 copies, group II: 3±5 copies and group III: 7±12 copies.
Functional expression of the inserted transgene was assessed by lacZ staining of post ®xed thin cryo- sections. No staining was found in group I hearts (Fig. 1C) nor in atrial tissue of group II and III (not shown). Both group II (Fig. 1D) and III (Fig. 1E) show characteristic blue staining of nuclei (due to
the endogenous nuclear localization signal (NLS), especially in myocytes of the affected intramural area in the left ventricular free wall (LVFW, Fig. 1B).
The number of stained nuclei in group III was much larger than in group II. In areas of group III LVFW which were heavily stained with LacZ, nuclear staining was accompanied by diffuse perinuclear staining due to excessive expression of the trans- gene. In those animals, a few myocytes of the right ventricular free wall (RVFW) stained positive as well.
Heart weight/body weight ratio's (HW/BW) as well as heart weight/tibia length ratio's (HW/TL) of the three groups are shown in Figure 1F, 1G and summarized in Table 1. HW/BW ratio's of group I (5.10.210ÿ3) and group II (6.10.310ÿ3) were not statistically different (P0.15) but the ratio of group III (7.30.510ÿ3) was signi®- cantly larger than ratio's of both group I (P0.0007) and group II (P0.045). HW/TL ratios of group I (6.90.110ÿ2) and II (6.80.110ÿ2) were not different (P1.0) but the ratio of group III (10.6 1.010ÿ2) was sig- ni®cantly larger than the ratios of group I (P0.021) and II (P0.017). Two not included hearts in group III weighted more than 400 mg (average HW group I; 155 mg, group III; 232 mg) and showed severe biventricular dilatation and left atrial thrombosis (not shown).
Extremity leads (ECG) and epicardial/endocardial mapping
After anesthesia, 3-extremity lead ECGs were recorded. Figure 2A shows typical examples of group I, II and III ECGs. Statistical analysis is sum- marized in Table I. Duration of P±Q interval (Fig. 2B) was signi®cantly longer in group III (39.73.0) compared to group I (33.90.3,P0.01) and II (33.61.0, P0.011). Duration of R±R interval (Fig. 2C) was not signi®cantly different between the three groups (I; 106.82.8, II; 102.73.0, III; 101.12.7, P0.39). Q±j interval (Fig. 2D) was signi®cantly prolonged in group III animals (19.52.3) as compared to group I (10.80.3, P,0.0001)) and group II (10.60.2,P,0.0001).
Table 1 Statistics of ECG analysis and HW/BW, HW/TL ratios. Group Icontrol, Group IImildly affected, Group IIIseverely affected. Data are meanSEM
Group I Group II Group III
HW/BW (10ÿ3) 5.10.2 6.10.3 7.30.5
HW/TL (10ÿ2) 6.90.1 6.80.1 10.61.0
R±R (ms) 106.82.8 102.73.0 101.12.7
P±Q (ms) 33.90.3 33.61.0 39.73.0
Q±j (ms) 10.80.3 10.60.2 19.53.0
In 3 out of 10 group III mice, a fractionated QRS complex was found (Fig. 2A, inset).
Epicardial maps of the LVFW are shown in Figure 3A, B. In line with their absence during ECG recordings, spontaneous sustained ventricular tachy-arrhythmias or single ectopic complexes could not be detected either in hearts of group I, II or III. LVFW epicardial maps of group I and III were comparable in sinus rhythm (SR, Fig. 3A, left panel), and during programmed stimulation from the center of the electrode grid at basic cycle length (BCL, Fig.3A, mid-panel). However, profound con- duction delay and regional block were detected in group III after premature stimulation (Fig. 3A, right panel). Regional block upon premature stimulation was observed in 7 out of 10 group III hearts but never in group I or II hearts. This difference was not due to differences in premature coupling intervals
(effective refractory period5 ms). The premature coupling interval in group III was 75.76.5 ms, which is not signi®cantly different from group I controls (72.01.2 ms, P0.65). The increased heterogeneity of LV-conduction is quanti®ed in Figure 3B. There is no difference for the median between the groups I and III. The absolute inhomo- geneity reveals a trend towards increased dispersion for group III, but did not reach signi®cance neither for BCL (P0.07) nor premature stimuli (P
0.06). The homogeneity index is, however, signi®- cantly larger for premature stimuli in group III mice (P0.02). Programmed premature stimula- tion (3 extra stimuli) of the hearts to asses vulner- ability for ventricular arrhythmias resulted in single ectopic beats in 3 out of 10 group III mice (not shown), but never in mice of groups I and II. RVFW epicardial maps (sinus rhythm) of group I, II and III Figure 2 (A) ECG complexes of Einthoven-2 recordings from groups I, II and III. Inset: example of a fractionated QRS complex as found in 3 out of 10 group III mice. (B±D) Bar plots of P±Q interval-length (B), R±R interval-length (C) and Q±j interval-length (D) of groups I, II and III. Average is given in Table 1, error bars indicate SEM. *I,*IIsigni®cantly different from group I and group II.
were comparable but showed propagation of the electric impulse with earliest activation found at variable sites. As for the LV there was a trend towards increased heterogeneity of conduction for mice in group III, but neither the absolute value nor the index reached signi®cance (data not shown).
After epicardial mapping, the left atrium was removed and a 8-point electrode grid was inserted through the mitral valves and placed against the endocardial surface of the LVFW. As with the epicardial recordings, no ventricular arrhythmias were detected and signi®cant (P0.042) con- duction slowing of an extra stimulus was found
in group III hearts compared to group I. When paced from the apical electrodes in the linear grid, conduction velocity in group I was 43.04.2 and 36.83.5 cm/s during basic and premature stimulation, respectively, and 34.33.7 and 24.43.4 cm/s in group III.
Expression of hypertrophic markers and down regulation of ID associated proteins
To characterize the degree of hypertrophy in the transgenic hearts, cryosections consecutive to those stained with LacZ, were incubated with antibodies raised against the hypertrophic markers Atrial Natriuretic Peptide (ANP), b-MHC and a-skeletal- actin. In group I, II and III, ANP was found in the atria as small punctate labeling while no ventricular labeling could be detected (not shown). Both in the RVFW and LVFW of group I, a small number of myocytes, primarily located near the apex and near the out¯ow tract, labeled positive for b-MHC anda-skeletal-actin. In group III, no aberrant label- ing was found in the RVFW and the IVS but abundant labeling fora-skeletal-actin (Fig. 4B) and b-MHC (Fig. 4D) was detected intramurally in the LVFW. Here, nearly all cells labeled positive while at a comparable location in group I (Fig. 4A and C, respectively) this was below 10%. The regions intensely positive for b-MHC and a-skeletal-actin coincided with the intense LacZ staining in consecu- tive sections.
In group I hearts, labeling with a monoclonal antibody raised against the gap junction protein Cx43 revealed a homogeneous characteristic pat- tern in the LVFW, the IVS and the RVFW. Labeling is primarily situated on the IDs at the longitudinal cell-cell contacts (Fig. 4E). A similar labeling pattern was obtained using antibodies against N-cadherin (Fig. 4G) andb-catenin (Fig. 4I); two other ID asso- ciated proteins. N-cadherin is a transmembrane protein in the adherens junction which mechanic- ally connects adjacent cells by anchoring the cyto- skeleton via b-catenin to the sarcolemma. In the intramural region of group III LVFW where the highest intensity of LacZ staining and marked upregulation of b-MHC and a-skeletal-actin was found, cells were almost devoid of Cx43 labeling (Fig. 4F). Depending on the transgene copynumber and the functional expression level of the transgene, the size of this area ranges from small clusters of 5±10 cells to40% of the LVFW. Down-regulation of N-cadherin (Fig. 4H) andb-catenin (Fig. 4J) was also found in the LVFW of group III. In those sec- tions, immunonegative areas could be observed Figure 3 (A) LVFW epicardial activation maps of a
group I (upper three panels) and a group III (lower three panels) hearts. Earliest activation is found in red, latest in blue, numbers indicate activation time in ms. SRsinus rhythm, BCLbasis cycle length. B: Bar plot showing the inhomogeneity in LV conduction in groups I and III hearts.
which varied in size but were smaller compared to the Cx43 negative areas.
Western blotting experiments supported the results obtained with immunohistochemical analysis. As compared to group I, no differences in expression levels of Cx43, N-cadherin, b-catenin anda-skeletal actin were found in the pooled atria, the RVFW and the IVS of group III (data not
shown). However, Figure 5 shows that in samples from the LVFW of group III mice, expression of Cx45, N-cadherin and b-catenin are strongly reduced compared to LVFW samples of group I hearts. In contrast, a-skeletal actin is upregulated in the group III mice compared to group I. Anti- bodies against b-MHC proved inappropriate to use for Western blotting.
Figure 4 Immunohistochemical staining of the transmural region of the LVFW in group I (left panel) and group III (right panel) hearts. Bar (shown in A) for A±J is l00mm.
Heterogeneity in remodeling
In group I LVFW (and group I, III RVFW and IVS), the IDs are marked by a strong co-localization of Cx43 and b-catenin (Fig. 6A) and Cx43 and N- cadherin (Fig. 6B). Down-regulation of Cx43 in group III LVFW appeared very heterogeneous since cells in the very apical, basal and epicardial layers of the LVFW are less affected or even show a fairly normal expression pattern. Labeling with a second, but polyclonal, anti-Cx43 antibody con®rmed the prior observations although the level of diffuse resi- dual Cx43 labeling in the severely affected intra- mural regions was slightly higher compared to labeling with the monoclonal anti-Cx43 antibody (data not shown). Next to down-regulation, cells were found which showed a redistribution of Cx43 both in group II hearts (small affected areas) and at the borders of the large Cx43 negative areas in group III hearts. In those cells, diffuse labeling marked the complete sarcolemma (Fig. 6C, arrow- head). Furthermore, co-expression of Cx43 with bothb-catenin and N-cadherin (Fig. 6D) appeared to be 1ost. In transitional cells between the Cx43 positive and Cx43 negative area,b-catenin and N- cadherin expression and distribution seemed fairly unaffected while a substantial amount of Cx43 is found as lateral labeling (Fig. 6D, arrowhead).
Changes in expression and distribution of N- cadherin and b-catenin seem to lag behind
remodeling of Cx43 expression since in Cx43 nega- tive areas, remnant labeling of b-catenin and N- cadherin was observed which mainly localized at lateral cell-cell contacts (Fig. 6E, arrowheads). Dou- ble-labeling for Cx43 and b-MHC (Fig. 6F) or for Cx43 and a-skeletal-actin (Fig. 6G) revealed that cells which labeled positive for b-MHC or a-skel- etal-actin are marked by the absence of Cx43 while surrounding b-MHC and a-skeletal-actin negative cells still express Cx43.
In 3 out of 10 group III animals, a very low amount of Cx40 re-expression was found as lateral labeling within small clusters of myocytes in the LVFW (3±5 Cx40 positive clusters per section ana- lyzed, Fig. 6H, arrowheads). No Cx40 could be detected both in group I hearts and group III work- ing ventricular myocytes of the IVS and RVFW.
Finally, no labeling of Cx45 gap junctions in the LVFW was found (data not shown).
Discussion
This study shows that: (1) Inb-MHC-hRARatrans- genic mice, functional expression of the transgenic construct results in an increase in heart weight as re¯ected by elevated HW/BW and HW/TL ratios.
The degree of phenotypical changes depends on the number of inserted copies. (2) In affected animals, ECG analysis shows a prolongation of P±Q and Q±j interval-length while R±R interval-length remains unchanged. Mapping of electrical activity revealed a delayed ventricular activation with enhanced spatial heterogeneity. Upon premature stimulation, regional conduction block and ectopic activity were observed. Severe spontaneous or pacing-induced ventricular arrhythmia's were absent. (3) In the LVFW, functional expression of the transgene is marked and coincides with re-expression of the (early) sarcomeric proteins b-MHC and a-skeletal- actin while ID-associated proteins like Cx43, N-cadherin and b-catenin are subcellularly redis- tributed prior to heterogeneous down-regulation.
Development of dilated cardiomyopathy
Transgenic mice overexpressing a constitutively active form of the human retinoic acid receptor develop a classical form of dilated cardiomyopathy.
The severity of pathophysiological and structural alterations depends on the number of copies inserted that give rise to functional expression level of the transgenic construct.5 Compared to group I, the control group with no inserted copies, only animals Figure 5 Representative Western blots of LVFW
samples derived from groups I and III hearts. In group III tissue, hypertrophic status is shown by a strong elevation of the a-skeletal-actin signal while ID associated proteins as Cx43, N-cadherin andb-catenin are down-regulated.
with 7±12 inserted copies (group III) are marked by an increased HW/BW ratio and HW/TL ratio. In group II (3±5 inserted copies), both ratios were com- parable to control mice. To exclude false positive/
negative results caused by ¯uctuating bodyweight, heart weight was also related to tibia length to assess severity of the myopathy.25Theb-MHC promoter, driving hRARa expression, caused expression primarily in the left ventricle (this study). As the pathophysiological processproceeds,orwhenalarge number of copies are functionally inserted, expres- sion of the transgene is found also in the right ventricles, though at a signi®cantly lower level.
Consequently, primarily left ventricular myocytes exhibit increased expression of hypertrophic mar- kers likeb-MHC anda-skeletal actin. Integration of these proteins in the contractile architecture of the
myocytes will likely affect the contractile properties of the cells adversely as shown previously.5,11
Mechanical and electrical dysfunction
In the LVFW, where expression of the transgene evokes expression of above mentioned hypertrophic markers, structural remodeling of the IDs has been identi®ed. At the level of the IDs, which mediate the electrical and mechanical coupling of myocytes, a severe and heterogeneous downregulation of Cx43, and to a lower degree of N-cadherin andb-catenin has been found. This implies that both electrical (Cx43) and mechanical coupling (N-cadherin and b-catenin) are affected. Mechanical dysfunction of these hearts has been reported before and was Figure 6 Immunohistochemical staining of the transmural region of the LVFW of group I (A and B), and group III hearts (C±H). Bars represent 20mm in C, E and H, 40mm in A, B, D, F, G.
characterized by a marked reduction in shortening fraction, velocity of circumferential ®ber shortening and systolic performance.5At the subcellular level, impairment of the intracellular connection between the contractile ®bers and adherens junctions (caused byb-catenin depletion) and impairment of intercellular mechanical coupling (caused by N- cadherin depletion) both contribute to development of mechanical dysfunction of affected hearts. In this context, it was shown that a dystrophin missense mutation,26 deletion of muscle LIM protein,27 or overexpression of cadherin isoforms,28 triggers development of DCM accompanied with structural reorganization of the IDs.
Next to mechanical dysfunction, electrical func- tion is also impaired. When compared to control (group I), ECG analysis revealed a prolongation of the P±Q interval in group III but not in group II.
This prolongation is mainly caused by an increased P-wave duration (10 ms in groups I and II,15 ms in group III) which might result from a conduction delay in the atria caused by slowed conduction or atrial enlargement.29±31 The QRS complex (mea- sured as Q±j interval) was considerably prolonged in group III animals with in 3 mice (out of 10) a fractionated QRS complex suggesting an asynchro- nous activation of the left and right ventricle or an asynchronous activation of affected and unaffected tissue within the left ventricle. Prolongation of the QRS complex due to impairment of propagation through the conduction system seemed unlikely since immunohistochemical analysis revealed a similar connexin expression pattern in the conduc- tion system of groups I and III hearts. Consequently, to explain the prolonged QRS duration, an impaired ventricular conduction was expected. Both epicar- dial mapping and mapping of the left ventricular endocardial free wall con®rmed this hypothesis (although more outspoken during premature stimu- lation) as in group III hearts a conduction delay of the electrical impulse was shown. Furthermore, LV conduct on appeared to be non-uniform as shown by a signi®cant increase in heterogeneity. During premature stimulation, regional conduction block was found in 7 out of 10 affected hearts, and in 3 out of 10, ectopic activity was recorded. In spite of these aberrancies, severe ventricular arrhythmias did not occur spontaneously during either ECG measure- ments or in Langendorff perfused hearts, and could not be induced even by aggressive programmed stimulation. Since ventricular arrhythmias can be life threatening, the fact that we did not observe any copynumber related spontaneous death of mice in our breeding colony indicates that probablyin vivo ventricular tachy-arrhythmias were absent too
thereby stressing the relative electrical stability of the hearts.
Redistribution and down-regulation of Cx43, N-cadherin andb-catenin
Immunohistochemical analysis of the LVFW of group III hearts showed a profound and heteroge- neous decrease in expression of Cx43, the main connexin isoform expressed between ventricular working cardiomyocytes. Similar to changes in N-cadherin and b-catenin expression, alterations were observed at spots where the transgene was functionally expressed. In all hearts (groups I, II and III), expression of Cx43, N-cadherin andb-catenin was unaffected in the working myocardium of the IVS and the RVFW which con®rms dependence of alterations on expression of the transgene. Depletion of Cx43 in the LVFW seems to precede that of N-cadherin and b-catenin since expression of N-cadherin and b-catenin generally is still present at the borderzone of an affected area while Cx43 is already lost. More to the center of an affected area all markers are lost. A second characteristic of the structural changes is redistribution of Cx43, and to a minor extent of N-cadherin and b-catenin, from the IDs to lateral cell borders. This phenomenon is mainly observed in the borderzone between nor- mal and affected tissue and thus seems to precede total loss.
Consequences of gap junction remodeling for electrical activation
Alterations in expression level and/or redistribution of Cx43 are consistently described observations in various models of ventricular hypertrophy and are proposed to increase the propensity to induce arrhythmias.13,15±17In this respect, heterogeneity in expression might be as potent.32 The b-MHC- hRARa transgenic hearts are marked both by a redistribution and a strong heterogeneous reduction in Cx43 expression. Remarkably, these profound changes (assumed to affect electrical coupling) did not result in severe spontaneous or inducible ven- tricular arrhythmias. Several possible explanations for this phenomenon may be suggested:
(1) Re-expression of other connexin isoforms might compensate for the loss of Cx43. Of the other two tested isoform known to be expressed in cardiac muscle (Cx40 and Cx45), Cx45 was not detected in the LVFW and even
though in 3 out of 10 severely affected hearts patchy re-expression of Cx40 was detected, the low level excludes compensation by this con- nexin isoform.
(2) Immunostaining with a polyclonal anti-Cx43 antibody revealed a higher degree of residual diffuse Cx43 labeling as compared to the label- ing detected with the monoclonal anti-Cx43 antibody. This residual diffuse presence of gap junctions in the affected area of the LVFW might cause a delay in propagation without inducing reentry. If alterations in Cx43 phos- phorylation were involved and the monoclonal and polyclonal Cx43 antibodies possessed different preferences to bind phosphorylated and unphosphorylated forms of Cx43, this would result in inappropriate detection of the actual residual Cx43 expression. However, Cx43 speci®city of both antibodies was con-
®rmed on Western blots prepared from mouse Cx43 transfected Hela cells which in addition revealed that both antibodies possessed equal preferences for the (un)phosphorylated states of Cx43 (data not shown).
(3) In two different types of cardiac-restricted conditional Cx43 knock-out mice, ventricular arrhythmias and sudden death were observed.33,34However, in these models, aver- age reduction of Cx43 expression in the ven- tricles was either robust (about 90%),33 or highly heterogeneous throughout the complete myocardium.34These changes in Cx43 expres- sion are substantially larger and more uniform as compared to ourb-MHC-hRARatransgenic mice in which size of Cx43 depleted area ranged from small clusters of about 5±10 cells in mildly affected hearts to about 40% of total LVFW area in severely affected hearts leaving the interventricular septum and the right ventricle unaffected.
We propose that the small size of mouse ventricles and the even smaller size of normal area in the affected hearts could be the limiting factor exclud- ing the occurrence of reentry circuits while local heterogeneity in Cx43 depletion might contribute to the regional character of conduction block.
Acknowledgements
A. A. B. van Veen and H. V. M. van Rijen were
®nancially supported by a project grant no.97.184 from the Netherlands Heart Foundation (to H. J. J.).
The authors wish to thank Dr A. F. M. Moorman for providing theb-MHC antibody.
References
1. KONSTAM MA, REMME WJ. Treatment guidelines in heart failure.Prog Cardiovasc Dis1998;41: 65±72.
2. MORGAN HE, BAKER KM. Cardiac hypertrophy.
Mechanical, neural, and endocrine dependence.
Circulation1991;83: 13±25.
3. HEINS, KOSTINS, HELINGA, MAENOY, SCHAPERJ. The role of the cytoskeleton in heart failure. Cardiovasc Res2000;45: 273±278.
4. JONGSMA HJ, WILDERS R. Gap junctions in cardio- vascular disease.Circ Res2000;86: 1193±1197.
5. COLBERT MC, HALL DG, KIMBALL TR, WITT SA, LORENZ JN, KIRBY ML, HEWETT TE, KLEVITSKY R, ROBBINSJ. Cardiac compartment-speci®c overexpres- sion of a modi®ed retinoic acid receptor produces dilated cardiomyopathy and congestive heart failure in transgenic mice. J Clin Invest 1997; 100:
1958±1968.
6. MANGELSDORF DJ, EVANS RM. The RXR hetero- dimers and orphan receptors. Cell 1995; 83:
841±850.
7. KASTNER P, MESSADDEQ N, MARK M, WENDLING O, GRONDONA JM, WARD S, GHYSELINCK N, CHAMBON P.
Vitamin A de®ciency and mutations of RXRalpha, RXRbeta and RARalpha lead to early differentiation of embryonic ventricular cardiomyocytes. Develop- ment1997;124: 4749±4758.
8. SUBBARAYAN V, MARK M, MESSADEQ N, RUSTIN P, CHAMBON P, KASTNER P. RXRalpha overexpression in cardiomyocytes causes dilated cardiomyopathy but fails to rescue myocardial hypoplasia in RXRalpha-null fetuses. J Clin Invest 2000; 105:
387±394.
9. LAMMEREJ, CHENDT, HOARRM, AGNISHND, BENKEPJ, BRAUN JT, CURRY CJ, FERNHOFF PM, GRIX AW JR, LOTT IT et al. Retinoic acid embryopathy. N Engl J Med1985;313: 837±841.
10. ZHOUMD, SUCOVHM, EVANSRM, CHIENKR. Retinoid- dependent pathways suppress myocardial cell hypertrophy. Proc Natl Acad Sci USA 1995; 92:
7391±7395.
11. HALLDG, MORLEYGE, VAIDYAD, ARDM, KIMBALLTR, WITT SA, COLBERT MC. Early onset hear failure in transgenic mice with dilated cardiomyopathy.
Pediatr Res2000;48: 36±42.
12. GROSD, JONGSMAHJ. Connexins in mammalian heart function.BioEssays1996;18: 719±730.
13. SEPP R, SEVERS NJ, GOURDIE RG. Altered patterns of intercellular junction distribution in human patients with hypertrophic cardiomyopathy.Heart1996;76:
412±417.
14. PETERSNS, COROMILASJ, SEVERSNJ, WITAL. Disturbed connexin43 gap junction distribution correlates with the location of re-entrant circuits in the epicar- dial border zone of healing canine infarcts that cause ventricular tachycardia. Circulation 1997;
95: 988±996.
15. SEVERSNJ. The cardiac muscle cell.Bioessays2000;
22: 188±199.
16. UZZAMAN M, HONJO H, TAKAGISHI Y, EMDAD L, MAGEE AI, SEVERS NJ, KODAMA I. Remodeling of gap junctional coupling in hypertrophied right ventricles of rats with monocrotaline-induced
pulmonary hypertension. Circ Res 2000; 36:
871±878.
17. TOMASELLI GF, MARBAÂN E. Electrophysiological remodeling in hypertrophy and heart failure.
Cardiovasc Res1999;42: 270±283.
18. LAIRD PW, ZIJDERVELD A, LINDERS K, RUDNICKI MA, JAENISCH R, BERNS A. Simpli®ed mammalian DNA isolation procedure. Nucleic Acids Res 1991; 19:
4293.
19. KASS DA, HAREJM, GEORGAKOPOULOSD. Murine car- diac function: a cautionary tail.Circ Res1998;82:
519±522.
20. vANRIJENHV,VANVEENTA,VANKEMPENMJ, WILMS- SCHOPMAN FJ, POTSE M, KRUEGER O, WILLECKE K, OPTHOFT, JONGSMAHJ,DEBAKKERJM. Impaired con- duction in the bundle branches of mouse hearts lacking the gap junction protein connexin40.Circu- lation2001;103: 1591±1598.
21. LAMMERSWJ, SCHALIJMJ, KIRCHHOFCJ, ALLESSIEMA.
Quanti®cation of spatial inhomogeneity in conduc- tion and initiation of re-entrant atrial arrhythmias.
Am J Physiol1990;259: 1254±1263.
22. LOWRY OH, ROSEBROUGH NJ, FARR AL, RANDALL RJ.
Protein measurement with the folin phenol reagent.
J Biol Chem1951;193: 265±275.
23. CLEMENTS, CHAPONNIERC, GABBIANIG. A subpopula- tion of cardiomyocytes expressing alpha-skeletal actin is identi®ed by a speci®c polyclonal antibody.
Circ Res1999;85: e51±e58.
24. CASH DE, BOCK CB, SCHUGHART K, LINNEY E, UNDERHILLTM. Retinoic acid receptor alpha function in vertebrate limb skeletogenesis: a modulator of chondrogenesis.J Cell Biol1997;136: 445±457.
25. VINCENT M, BOUSSAIRI EH, CARTIER R, LO M, SASSOLAS A, CERUTTI C, BARRES C, GUSTIN MP, CUISINAUD G, SAMANI NJ et al. High blood pressure and metabolic disorders are associated in the Lyon hypertensive rat.J Hypertens1993;11: 1179±1185.
26. ORTIZ-LOPEZ R, LI H, SU J, GOYTIA V, TOWBIN JA.
Evidence for a dystrophin missense mutation as a cause of X-linked dilated cardiomyopathy.Circulation 1997;95: 2434±2440.
27. EHLER E, HOROWITS R, ZUPPINGER C, PRICE RL, PERRIARD E, LEU M, CARONI P, SUSSMAN M, EPPENBERGER HM, PERRIARD JC. Alterations at the intercalated disk associated with the absence of muscle lim protein. J Cell Biol 2001; 153:
763±772.
28. FERREIRA-CORNWELLMC, LUOY, NARULAN, LENOXJM, LIEBERMANM, RADICEGL. Remodeling the intercalated disc leads to cardiomyopathy in mice misexpressing cadherins in the heart. J Cell Sci 2002; 115:
1623±1634.
29. CHENYJ, CHENSA, TAICT, YUWC, FENGAN, DINGYA, CHANG MS. Electrophysiologic characteristics of a dilatedatriuminpatientswithparoxysmalatrial®bril- lation and atrial ¯utter. J Interv Card Electrophysiol 1998;2: 181±186.
30. VERHEULE S, VAN BATENBURG CA, COENJAERTS FE, KIRCHHOFFS, WILLECKEK, JONGSMAHJ. Cardiac conduc- tion abnormalities in mice lacking the gap junction protein connexin40. J Cardiovasc Electrophysiol 1999;10: 1380±1389.
31. HAGENDORFF A, SCHUMACHER B, KIRCHHOFF S, LUÈDERITZ B, WILLECKE K. Conduction disturbances and increased atrial vulnerability in Connexin40 de®cient mice analyzed by transesophagal stimula- tion.Circulation1999;99: 1508±1515.
32. FAST VG, DARROWBJ, SAFFITZJE, KLEBERAG. Aniso- tropic activation spread in heart cell monolayers assessed by high-resolution optical mapping.
Role of tissue discontinuities. Circ Res 1996; 79:
115±127.
33. GUTSTEIN DE, MORLEY GE, TAMADDON H, VAIDYA D, SCHNEIDER MD, CHEN J, CHIEN KR, STUHLMANN H, FISHMAN GI. Conduction slowing and sudden arrhythmic death in mice with cardiac-restricted inactivation of connexin43. Circ Res 2001; 88:
333±339.
34. GUTSTEINDE, MORLEYGE, VAIDYAD, LIUF, CHEN FL, STUHLMANNH, FISHMANGI. Heterogeneous expression of gap junction channels in the heart leads to conduction defects and ventricular dysfunction.
Circulation2001;104: 1194±1199.
