OUP user menu

Dedifferentiation of atrial myocytes during atrial fibrillation: role of fibroblast proliferation in vitro

Catherine Rücker-Martin, Françoise Pecker, David Godreau, Stéphane N Hatem
DOI: http://dx.doi.org/10.1016/S0008-6363(02)00338-3 38-52 First published online: 1 July 2002


Objectives: Severe myocyte alterations, characterized by enlarged myocytes and myolysis, is observed in fibrillating and dilated atria and contributes to atrial fibrillation. The aim of this study was to determine the nature of this cellular remodeling process and factors involved in its regulation. Methods: In vivo, contractile proteins were studied in 24 human right atrial specimens by means of immunohistochemical techniques. In an attempt to reproduce in vitro the myocyte remodeling and to study its regulation, human atrial myocytes were cultured (n = 27) and analyzed immunocytochemically; intracellular Ca2+ transients (Cai-tr) in response to electrical stimulation were monitored using Fura-2/AM. Results: In diseased specimens, sarcomeres, seen at the periphery of myolytic myocytes, stained positively with antibodies against sarcomeric proteins of the Z-band (α-actinin and titin epitope T12) but not with antibodies against titin epitope T11 (I-band) or desmin (intermediate filament). β-myosin heavy chain (MHC) and smooth muscle α-actin, two proteins of the fetal program, were re-expressed. In culture, diseased myocytes also showed myolysis and glycogen accumulation; their sarcomeres stained positively with anti-α-actinin, anti-T12, anti-β-MHC and anti-smooth muscle α-actin but not with anti-titin T11 or anti-desmin antibodies. At confluence, myocytes regained a normal sarcomeric apparatus and were excitable, as shown by electrical Cai-tr triggering. This redifferentiation process was inhibited by fibroblast proliferation. Conclusion: In diseased atria, myolytic myocytes are in a dedifferentiated state resembling that of immature muscle cells. In vitro, fibroblast proliferation prevents the reversibility of this cellular alteration.

  • Arrhythmia (mechanisms)
  • Atrial function
  • Fibrosis
  • Histo(patho)logy
  • Myocytes
  • Remodeling

Time for primary review 22 days.

This article is refered to in the Editorial by J. Ausma and M. Borgers (pages 9–12) in this issue.

1 Introduction

Severe tissue and cellular alterations are observed in the atrial myocardium of patients with chronic atrial fibrillation (AF), including fibrosis and abundant large dystrophic myocytes showing extensive myolysis [1–5]. These abnormal myocytes, which resemble hibernating ventricular myocytes [6,7], could contribute to the altered functional properties of the atrium and to the initiation and perpetuation of AF. Thus, a better understanding of the significance, role and regulation of these myolytic myocytes in the pathogenesis of AF is of major clinical importance.

The nature of this cellular remodeling is still largely unknown. A degenerative process is ruled out by the lack of lysosomal degradation [3]. Similar cellular remodeling is observed in a goat model of AF induced by electrical pacing and associated with re-emergence of the dedifferentiated phenotype of immature myocardium [3,4]. This observation led to the proposal that myolytic myocytes might reflect adaptation of the atrial myocardium to changes in its working conditions, a process which might be reversible following restoration of normal atrial beating and loading conditions [3,4]. However, in a previous study we found that a number of these abnormal myocytes undergo apoptosis, in the same way as hibernating ventricular myocytes [5]. This indicated that, in some clinical circumstances, the cellular remodeling that occurs in the diseased atrial myocardium may be irreversible.

The present study was undertaken to analyze the nature of the myolytic myocytes observed during human AF, and to identify factors involved in its modulation. We specifically attempted to answer the following questions: (i) are myolytic myocytes during AF dedifferentiated? and (ii) what is the role of fibroblast proliferation in this cellular remodeling process? Using various immunohistological staining methods, we examined the expression and localization of the main contractile proteins that characterized terminally differentiated adult myocytes, including (i) sarcomeric α-actinin and the NH2-terminal part of the giant protein titin (to study Z-band formation) [8,9], (ii) the titin I band region [9,10], (iii) desmin (an intermediate-filament protein) [11] and (iv) smooth muscle α-actin and β-myosin heavy chain (βMHC), both of which belong to the fetal program. To study factors involved in the regulation of these cellular alterations, and especially the role of fibroblast proliferation, we used the ‘dedifferentiation–redifferentiation’ model of atrial myocytes in long-term culture, which has contributed to the identification of mechanisms involved in myocardial remodeling during hypertrophy and heart failure [12–16].

2 Methods

2.1 Cardiac specimens

The investigation conforms with the principles outlined in the Declaration of Helsinki (Cardiovascular Research 1997;35:2–3). With approval from the Ethics Committee of our institution, specimens of right atrial appendage were obtained from 51 patients aged from 4 to 82 years undergoing heart surgery for coronary artery disease (n = 20), valve disease (n = 17) or congenital heart defects (n = 14). Most patients with chronic AF had mitral valve disease associated with left and right atrial dilation and increased systolic pulmonary artery pressure (>45 mmHg), indicating longstanding hemodynamic right atrial overload (Tables 1 and 2). Patients in sinus rhythm (SR) were assigned to the control group and except for two of them, they suffered from coronary artery disease and had no left atria enlargement or pulmonary artery hypertension. Immediately after excision, specimens of right atrial appendage were frozen in isopentane or fixed in 10% buffered formalin overnight before paraffin embedding for in situ histological studies (n = 24) (Table 1) or were washed in calcium-free Krebs–Ringer solution supplemented with 30 mM 2,3-butanedione monoxime (BDM) before myocyte isolation (n = 27) (Table 2).

View this table:
Table 2

Clinical characteristics of the atrial appendage donors for the in vitro study

  • AS, aortic stenosis; CAD, coronary artery disease; CHD, congenital heart defect; EF, ejection fraction of the left ventricle; LA, left atrium; MR, mitral regurgitation; PAP, pulmonary artery pressure; SR, sinus rhythm; NA, not available; I, immunostaining; H, histological analysis; F, functional analysis.

View this table:
Table 1

Clinical characteristics of the atrial appendage donors for the in situ study

  • AS, aortic stenosis; CAD, coronary artery disease; EF, ejection fraction of the left ventricle; HypoRV, hypo right ventricle; LA, left atrium; MR, mitral regurgitation; MS, mitral stenosis; PAP, pulmonary artery pressure; SR, sinus rhythm; TR, tricuspid regurgitation; I, immunostaining; H, histological analysis.

2.2 Histological analysis

Tissue sections were deparaffinized, transferred and rehydrated in decreasing concentrations of alcohol. Cultured myocytes were fixed with 4% formaldehyde at room temperature for 10 min. Preparations were stained with periodic acid-Schiff (PAS) or Masson's trichrome method.

2.3 Atrial myocyte isolation and culture

Myocyte isolation from human right atrial appendage samples and their culture were performed as previously described [14,15]. Briefly, cell dissociation was achieved by several enzymatic steps using collagenase (type IV, Sigma) and protease (type XXIV, Sigma). An average 1 890 000±260 000 (n = 27) myocytes per gram of tissue were obtained, and 66±3% of the isolated myocytes were calcium-tolerant, i.e. rod-shaped and well-striated. Isolated myocytes were resuspended at a final density of 7×104/ml in Dulbecco's modified Eagle's medium (Life Technology, France) supplemented with 10% fetal calf serum (BioWhittaker, France), non-essential amino acids, 1 nM insulin and antibiotics (100 IU/ml penicillin and 0.1 μg/ml streptomycin) (Life Technology, France) then plated (day 0) on laminin-coated dishes (10 μg/ml, Life Technologies). In some cultures, to inhibit the proliferation of non muscle cells, 10 μM cytosine β-d-arabino-furanoside (CA) (Sigma-Aldrich) was added throughout the culture period. Cultures were studied at the first and third weeks.

2.4 Immunological reagents

Mouse monoclonal antibodies against human smooth muscle α-actin FITC conjugate (clone 1A4, 1/25), sarcomeric α-actinin (clone EA-53, 1/200) and titin T11 epitope (clone F149.4B3 or T11, 1/10) were from Sigma. A mouse monoclonal antibody against titin T12 epitope (clone F146.9B9 or T12, 1/10) was from Novo Castra (Tebu). A mouse monoclonal anti-connexin 43 (Cx43) antibody (clone CX 1B1, 1/10) was from Zymed Laboratories Inc (CliniSciences), and a mouse monoclonal antibody directed against β-MHC was a gift from Dr V. Mouly (1/20) [17]. A rabbit polyclonal antibody against fibronectin (1/200) was from Chemicon International (Euromedex), and a rabbit polyclonal antibody against desmin (1/50) was from Cappel (INC). Normal rabbit IgG fractions were from Vector Laboratories, fluorescein- and Texas red-conjugated donkey anti-rabbit IgG (1/30) was from Amersham, and normal isotype-matched mouse IgG and goat biotinylated antimouse IgG secondary antibody (1/30) were from Vector Laboratories. Streptavidin–Texas Red (1/30) was from Amersham.

2.5 Immunostaining

Indirect immunofluorescence was performed on frozen sections (5 μm) of atrial appendages and on cultured myocytes fixed with 4% formaldehyde at room temperature for 10 min as previously described [14,15]. Briefly, sections and myocytes were incubated in PBS containing 5% BSA for 30 min to block nonspecific binding sites, followed by overnight incubation with mouse monoclonal antibodies at 4 °C. Then, sections and cultured myocytes were incubated with goat biotinylated antimouse IgG secondary antibodies. After washing in PBS, sections and myocytes were incubated with streptavidin–Texas red. In a second step, preparations were incubated with rabbit polyclonal antibodies and then with fluorescein- or Texas red-conjugated donkey anti-rabbit IgG as described above. Coverslips were mounted in mounting medium (Fluoprep, bioMerieux). In control experiments the incubation steps with primary antibodies were omitted. Slides were examined with an MRC-1024 (Bio-Rad, UK) confocal scanning laser with microscope (Nikon Optiphot Fluorescence) using Lasersharp software (Bio-Rad). The resulting images were printed on an Epson Stylus Photo color printer.

2.6 Morphometric study

To quantify the amount and the percentage of myocytes with myolysis, two Masson's trichrome-stained sections per atrial sample were analyzed, and at least 400 cells per section in which the nucleus was present in the plane of the section were analyzed. Cells were scored morphometrically as severely myolytic if >25% of the sarcomere was absent [3].

The surface area of sarcomeric α-actinin, titin T12 and titin T11 immunostaining was measured and analyzed using Cyberview 4.0 software (Cervus International). This software allows structures to be selected on the basis of pixel intensity values in a given color channel. The image was transformed into gray luminance values ranging from 0 (black) to 255 (white). The image analyzer software automatically evaluated positive labeling by using a threshold method, i.e. selecting pixels whose intensity level was greater than a threshold value (background). Results were given as the percentage of the threshold area relative to the total area.

A semi-quantitative approach was used to compare cultures for glycogen, sarcomeric α-actinin, titin T12 and T11 staining. This procedure was applied to four different cultures using a 20× objective: a score of 0 indicates the absence of labeling; positive staining was graded ±, +, ++ or +++. The reproducibility of the quantification was assessed with two independent examiners.

2.7 Fura-2 loading and [Ca2+]iimaging

Myocytes attached to laminin were bathed in 2 ml of saline buffer (10 mM glucose, 130 mM NaCl, 5 mM KCl, 10 mM HEPES buffered at pH 7.4 with Tris base, 1 mM MgCl2, 2 mM CaCl2) and incubated for 20 min at 25 °C with 1.5 μM Fura-2/AM (3 μl of 1 mM Fura-2/AM in DMSO) in the presence of 1 mg/ml bovine serum albumin as previously described [18]. Cells were then washed and incubated in the same buffer for 15 min at 25 °C to facilitate hydrolysis of intracellular Fura-2/AM. Field electrical stimulation (square waves, 10 ms duration, 0.1, 0.2, 0.5 or 1 Hz) was delivered through a pair of platinum electrodes connected to a HAMEG stimulator (Paris, France).

2.8 Statistical analysis

Results are shown as means±S.E.M., and were analyzed using Student's t-test. Significance was accepted when P<0.05.

3 Results

3.1 Myolytic myocytes show characteristics of dedifferentiated cells

As illustrated Fig. 1A, in samples from patients of the control group (Table 1), the vast majority of myocytes were of uniform size and regularly organized within the trabeculae whereas less than 15±2% (n = 8 patients) of them showed myolysis. In contrast, myocardial tissue from patients with AF (Table 1) showed large areas of myocytes with extensive myolysis; these cells represented 66±2% (n = 8 patients) of total myocytes (Fig. 1B); the myofibrils had been replaced by glycogen granules (not shown) [5]. Moreover, bundles of myofibers were separated by thick layer of fibrosis that also accumulated between myocytes to form collagenous septa. Most of the patients with AF had a chronic arrhythmia and mitral valve diseases however, there was no clear relationships between the degree of the tissue remodeling and the duration of AF or the severity of the cardiopathy (i.e. atrial dilatation, pulmonary hypertension). Of note, some degree of extracellular fibrosis was also seen in control patients above 62 years old [19,20].

Fig. 1

Masson's trichrome staining of control (A, patient 22 in Table 1) and fibrillating atrial specimens (patient 8 in Table 1, 10 years since first AF), showing severe myolysis (B, arrowhead) and marked fibrosis. Bar=25 μm.

In order to determine which sarcomere protein was altered in myolytic myocytes, and to evaluate the severity of this process, immunostaining was performed in six patients in AF and eight patients in SR (Table 1) using antibodies directed against α-actinin (Fig. 2A,B) and the NH2-terminal (epitope T12) and I-band (epitope T11) regions of the titin protein (Fig. 2C–F) and desmin (Fig. 2G,H). Normal myocardial samples showed a typical regular cross-striated staining pattern throughout the cell body with the different antibodies, corresponding to the normal sarcomeric apparatus of an adult myocyte (Fig. 2A,C,E). Likewise, intercalated discs were heavily stained with anti-desmin antibody (Fig. 2G). In AF samples, there was a significant decrease in the surface area stained by anti-α-actinin (79±3 vs. 91±3%, in AF vs. SR; P<0.05), anti-titin T12 (70±1 vs. 81±3%, in AF vs. SR; P<0.001), and more pronounced with anti-titin T11 (15±2 vs. 63±4%, in AF vs. SR; P<0.0001) (Fig. 3). In fibrillating atria, sarcomeric organization was observed essentially at the periphery of abnormal myocytes, which stained positively with anti-α-actinin and anti-titin T12 (Fig. 2B,D); only weak striated staining was observed with anti-T11 (Fig. 2F). In some areas, titin T12 staining was also organized in a punctuate pattern, indicating disorganization of this protein (Fig. 2D). Desmin was mainly observed at the level of the intercalated discs (dense staining), while weak, poorly organized labeling was observed at the level of Z lines (Fig. 2H). Control tissue sections were negative with anti-smooth muscle α-actin and anti-β-myosin heavy chain antibodies (Fig. 4A,C; Table 3), while these two contractile proteins of the fetal phenotype were detected in 76% of myocytes in diseased sections (Fig. 4B,D; Table 3). These results indicated that myolytic myocytes of diseased atria had marked structural and phenotypic alterations of their contractile apparatus.

Fig. 2

Frozen atrial sections stained with anti-α-actinin (A,B), anti-titin T12 (C,D) and T11 epitope (E,F) and anti-desmin (G,H) antibodies; control (A, C, E, G) and diseased atria (B, D, F, H) (confocal microscopy). Note the sarcomeric organization at the periphery of the myolytic myocytes (arrowhead), the depletion of contractile material (stars), and the intercalated discs stained with anti-desmin antibody (arrow). Bar=25 μm.

Fig. 3

Changes in the expression of α-actinin sarcomeric ■, titin T12 Embedded Image and titin T11 □. * P<0.05, ** P<0.001 and *** P<0.0001 for the comparison between the labeling area of control and dilated and fibrillating atria.

Fig. 4

Frozen atrial sections stained with anti-β-MHC (A,B) and anti-smooth-muscle-α-actin (C,D) antibodies, control (A, C), and diseased atria (B, D) (confocal microscopy). Note the presence of vessel strongly stained with anti-smooth-muscle-α-actin antibodies in C and the detection in a large number of myocytes in diseased sections of the two contactile proteins which are representative of fetal phenotype. Bar=25 μm.

View this table:
Table 3

Presence and intensity of glycogen and sarcomeric α-actinin, α-smooth actin, β-MHC, titin T12 and T11 labeling evaluated in situ and in four different cultures

In situDay 8Day 21
α-Actinin sarcomeric+++++++++++++
(n = 5 cultures)      
Titin T12++++++++++
(n = 4 cultures)      
Titin T11+++++0±
(n = 4 cultures)      
Glycogen granules0++++++++++++0
(n = 3 cultures)      
α-Smooth actin0++++0
(n = 3 cultures)      
(n = 3 cultures)      
  • Score 0 corresponds to the absence of labeling; positive staining is graded ±, +, ++ or +++.

    CA, cytosine β-d-arabino-furanoside.

3.2 Cultured atrial myocytes resemble the myolytic myocytes of diseased atria

After a week in culture, whatever the clinical characteristics of the donor (Table 2), myocytes lost their rod shape, spread, and increased in size. Staining with anti-α-actinin, anti-titin T12 and anti-titin T11 antibodies revealed marked depletion of contractile material (Fig. 5A–C; Table 3). A cross-striated sarcomeric organization was mainly observed at the periphery of the cells (Fig. 5A–C), whereas the center of cells lacking contractile material was filled with dense glycogen granules (Fig. 5G,H). The T12 and, to a lesser extent, the T11 epitope of titin protein were detected on the remaining sarcomeric apparatus. Desmin was absent from the cell body and was only detected in the extension of the plasma membrane (Fig. 5D). Cultured myocytes re-expressed smooth muscle α-actin and β-myosin heavy chain (Fig. 5E,F). The former protein was organized as a network of stress fibers throughout the cell; these fibers were anti-β-MHC-negative (Fig. 5E,F, circle and arrowhead). Taken together, these results showed that myocytes underwent marked growth and dedifferentiation after a week in culture and shared similarities with myolytic myocytes of diseased atria.

Fig. 5

Immunostaining (visualized by confocal microscopy) of myocytes after 1 week in culture, with antibodies directed against α-actinin (A), titin T12 (B) and T11 (C), desmin (D), βMHC (E) and smooth-muscle-α-actin (F). Note the depletion of contractile material (stars) and its replacement by glycogen granules stained with PAS+diastase (G, arrow) but not with PAS alone (H). The dedifferentiation is independent of the clinical characteristics of the donor: A, D, E, F, G and H are from patient in sinus rhythm, and B and C from patients with AF. Bar=25 μm.

3.3 Fibroblast proliferation inhibits redifferentiation of myocytes in vitro

After 3–4 weeks of culture in the presence of CA, when myocytes had reached confluence, few cells were positive with anti-fibronectin immunoglobulin, which stains fibroblasts (Fig. 6A). At this stage the myocytes were entirely occupied by a cross-striated sarcomeric apparatus that stained positive with the anti-α-actinin antibody (Table 3). As illustrated in Fig. 6C, confluent myocytes re-established tight cell–cell contacts, and punctate Cx43 labeling delineated the entire cell outline (arrowhead). Titin T12 staining (Fig. 7A) was superimposable on α-actinin staining, whereas glycogen granules were no longer visible (Fig. 6C,E) (Table 3). Titin T11 was mainly present in an amorphous state (Fig. 7C; Table 3), but in some areas it formed striations (arrow Fig. 7C). Desmin also showed a staining pattern consistent with a predominantly sarcomeric localization (Fig. 7E). Smooth muscle α-actin was absent from confluent and sarcomerized myocytes (Table 3), whereas β-myosin heavy chain staining yielded a typical H-band pattern (Fig. 7G).

Fig. 7

Myocytes cultured for more than 3 weeks, with (A, C, E, G) or without (B, D, F, H) CA, and stained with anti-T12 (A,B) and T11 (C,D) titin, anti-desmin (E,F) and anti-β-MHC (G,H) antibodies (confocal microscopy). Note the striated organization of titin T12 and T11 and desmin (arrow head). This redifferentiation process was also independent of the clinical characteristics of the donor: F, G and H are from patients in sinus rhythm, and A, B, C, D and E from patients with AF. Bar=25 μm.

Fig. 6

Immunostaining of human atrial myocytes cultured for 3 weeks in the presence (A,C,E) and absence (B,D,F) of CA; double immunostaining with anti-sarcomeric α-actinin (red) and anti-fibronectin (green) antibodies (A,B, donor in sinus rhythm) (only fibroblasts were stained with anti-fibronectin (arrow head)) and with anti-sarcomeric α-actinin (red) and anti-connexin 43 (green) antibodies (C,D, donor in AF). Glycogen granules (stained with PAS+diastase, E,F) were only observed in the absence of CA (arrow) (donor in sinus rhythm). Bar=25 μm.

In the absence of CA, fibroblasts proliferated and formed contacts with myocytes, resulting in a confluent culture system (Fig. 6B). Cx43 staining yielded a dispersed disorganized punctuate pattern at the periphery of myocytes (Fig. 6D arrow) and also at the level of the cell body, probably corresponding to dense and tight contacts between myocytes and fibroblasts (Fig. 6D circle). In these culture conditions the sarcomeric apparatus remained poorly organized, being mainly visible at the periphery of cells and along myocyte extensions (Fig. 6D; Table 3). In the central part of myocytes, sarcomeric α-actinin staining gave a filamentous and punctuate pattern (Fig. 6D) and an accumulation of glycogen granules was still present (Fig. 6F). Only partial reorganization of titin T12 was observed (Fig. 7B) and, in the majority of myocytes, titin T12 and titin T11 staining was poorly organized throughout the cell body (Fig. 7B,D). Desmin was only detected at the periphery of myocytes, at the level of cell–cell contacts (Fig. 7F).

3.4 Recovery of cell excitability during culture is inhibited by fibroblast proliferation

To overcome the technical difficulty of recording action potentials in confluent myocyte monolayers, excitability was studied during culture by recording the intracellular Ca2+ transient (Cai-tr) as a marker of a functional excitation–contraction coupling process. On the day of isolation (day 0), after 2 h of attachment to laminin, cardiomyocytes were responsive to electrical pulses, and Cai-tr followed the frequency of the electrical field stimulation, from 0.1 to 1 Hz (Fig. 8). In contrast, on day 6 of culture, the cells were quiescent or showed only rare spontaneous Cai-tr, and did not respond to electrical stimulation (data not shown). On day 20 of culture in the presence of CA, myocytes were quiescent or sometimes showed spontaneous activity and in addition they responded to field stimulation delivered at frequencies between 0.5 and 1 Hz (Fig. 8). In contrast, when fibroblasts proliferated (in the absence of CA), the cultured myocytes remained unexcitable (data not shown). Taken together, these results were in keeping with a sequential loss and recovery of conduction protein expression, observed after 6 and 20 days of culture, respectively.

Fig. 8

Effect of electrical stimulation on the intracellular free calcium concentration recorded in Fura-2 AM-loaded human atrial myocytes maintained in culture for 0 and 20 days in the presence of CA.

4 Discussion

This study shows that myolytic atrial myocytes, which occupy large areas of fibrillating and diseased atrial myocardium, are dedifferentiated, with characteristics of embryonic muscle cells. By reproducing this dedifferentiation process in vitro, we found evidence that fibroblast proliferation might be an important factor in the worsening and irreversibility of this cellular remodeling phenomenon.

As previously reported [5], in the right appendage of diseased human atria, a majority of myocytes are enlarged, showing a marked depletion of their contractile apparatus and its replacement by an accumulation of glycogen granules. Several lines of evidence indicate that these myolytic myocytes have structural and phenotypic characteristics observed during ontogenic development of cardiac muscle [21]. For instance, desmin is totally absent from the sarcomere of myolytic myocytes, and is only present at the level of the intercalated disc. This intermediate-filament protein is involved in maintaining the structural and functional integrity of the muscle, and is detected in the late stage of contractile apparatus formation [11,22]. Only the T12 epitope of titin is detected in myolytic atrial myocytes, and is spatially associated with sarcomeric α-actinin in a Z-band-like sarcomeric structure. This is reminiscent of the α-actinin/titin aggregates (I-Z-I-like complexes) observed early during myofibrillogenesis which, together with clusters of myosin and actin, contribute to sarcomere assembly [9,23–25]. If, in myolytic atrial myocytes, sarcomeric α-actinin and titin T12 are the remnants of an I-Z-I-like complex, this might indicate that myolysis during AF could result from disassembly of the myofibrillar structure.

In addition, the proteins that composed the contractile apparatus were those of an immature myocardium; for example, smooth muscle α-actin and β-MHC were detected in atrial myocardium from patients but not healthy controls. Both proteins belong to the fetal program and are re-expressed during hypertophic remodeling of the rat left ventricle, reflecting a general adaptation of this myocardium to changes in its working conditions [26]. In atria from patients with mitral valve disease, a switch from α- to β-myosin heavy chain has been attributed to the atrial pressure and volume overload, suggesting that this myocardium can also adapt to changes in its loading conditions [27]. In the goat, following a period of AF induced by electrostimulation, a number of myocytes are also myolytic and in a relatively dedifferentiated state [3,4]. Taken together, these findings indicate that myolysis and dedifferentiation of myocytes could result from a common and non specific adaptive response to changes in the working conditions of the atrial myocardium including rapid beating or hemodynamic overload.

The resemblance between myolytic atrial myocytes and long-term-cultured atrial myocytes is another indication that the former are in a dedifferentiated state [12–15,23–25,28]. Like myolytic myocytes of fibrillating atria, long-term-cultured atrial myocytes lost their contractile apparatus, which remained mainly at the cell periphery, their empty cytosol being filled with glycogen granules. In addition, the T12 but not the T11 epitope of titin was detected, sarcomeric desmin was lacking, and smooth muscle α-actin and β-myosin heavy chain were re-expressed. This growth and dedifferentiation of myocytes in long-term culture is well known, and has been extensively used as a model to identify factors that regulate cardiac phenotype plasticity [29]. For instance, using the capacity of cultured myocytes to redifferentiate when they establish cell–cell contacts in vitro, it was recently shown that the atrial and ventricular tissue-specific pattern of gap junctions is determined and regulated by factors intrinsic to myocytes [30]. The study by Kostin et al. [30] points to the major role of cell–cell contacts in maintaining a normal architecture and structure of the myocardium [29,31]. In our study, confluent human atrial myocytes in culture also regained characteristics of differentiated myocytes, with a well-developed sarcomeric apparatus and normal excitability, indicating the restoration of structures that control the excitation–contraction coupling process. However, the reappearance of only discrete striated titin T-11 staining indicates that sarcomere formation was not complete after 3 weeks in culture; a longer culture period may be required, or certain factors may be lacking in culture, such as electrical and mechanical activity [31].

The other important finding of this in vitro study is that when fibroblasts proliferate they establish tight junctions with myocytes (as indicated by diffuse and poorly organized Cx43 staining) and prevent myocytes from re-differentiating and recovering their excitability. This indicates that the nature of the cell–cell contact plays a key role in maintaining the normal structure and function of atrial myocytes. This is reminiscent of the recent observation that rabbit ventricular myocytes undergo hibernation-like dedifferentiation when co-culture with cardiac fibroblasts [16].

4.1 Possible clinical implications and limitations of this study

The lack of contractile activity and passive stretching contribute to the myolysis and dedifferentiation of cardiac myocytes in vitro [32]. During AF, contractile activity is down-regulated and the atrial myocardium becomes subject to abnormal passive stretching that might induce this cellular remodeling and ultrastructural changes. The massive loss of sarcomeres in a number of atrial myocytes could contribute to atrial contractile dysfunction [33] and to the subsequent increase in passive stretching [10,34], and thereby participate in a vicious circle that perpetuates AF [35]. This would also explain why the restoration of sinus rhythm and normal atrial contractile activity is important for maintaining the contractile apparatus and, possibly, for preserving the ability of myolytic myocytes to regain a more differentiated structure and phenotype. Indeed, in vitro, this dedifferentiation process is reversible when myocytes re-establish cell–cell contacts. In contrast, fibroblast-myocyte contacts considerably delayed and inhibit myocyte re-differentiation suggesting that factors intrinsic to myocytes may be necessary to maintain a differentiated cardiac phenotype as reported for the organization of gap junction [30]. Another explanation could be that fibroblast proliferation alter the electrical and mechanical coupling between myocytes that are both important factors to maintain a normal cardiac phenotype [20,32]. Given the marked accumulation of extracellular matrix in diseased atria and also with age, it is likely that proliferation of fibroblasts occurs also in situ and thus could represent a factor of irreversibility and aggravation of cellular remodeling [2,5] and contribute to the increase incidence of atrial fibrillation with aging. Further studies are required to elucidate the role of fibroblast proliferation in extracellular matrix accumulation and myocytes remodelling during chronic atrial fibrillation.


This work was supported by grants from the Société Française de Cardiologie (SFC), Servier Laboratory, the Association Française contre les Myopathies (AFM) and the Fondation pour la Recherche Médicale (FRM). We thank Drs Loı̈c Macé and Patrice Dervanian for providing atrial samples.


  1. [1]
  2. [2]
  3. [3]
  4. [4]
  5. [5]
  6. [6]
  7. [7]
  8. [8]
  9. [9]
  10. [10]
  11. [11]
  12. [12]
  13. [13]
  14. [14]
  15. [15]
  16. [16]
  17. [17]
  18. [18]
  19. [19]
  20. [20]
  21. [21]
  22. [22]
  23. [23]
  24. [24]
  25. [25]
  26. [26]
  27. [27]
  28. [28]
  29. [29]
  30. [30]
  31. [31]
  32. [32]
  33. [33]
  34. [34]
  35. [35]
View Abstract