Electrotonic Modulation of Cardiac Impulse Conduction by Myofibroblasts
http://www.100md.com
《循环研究杂志》
the Department of Physiology, University of Bern, Switzerland.
Abstract
Structural remodeling of the myocardium associated with mechanical overload or cardiac infarction is accompanied by the appearance of myofibroblasts. These fibroblast-like cells express -smooth muscle actin (SMA) and have been shown to express connexins in tissues other than heart. The present study examined whether myofibroblasts of cardiac origin establish heterocellular gap junctional coupling with cardiomyocytes and whether ensuing electrotonic interactions affect impulse propagation. For this purpose, impulse conduction characteristics (conduction velocity [] and maximal upstroke velocity [dV/dtmax]) were assessed optically in cultured strands of cardiomyocytes, which were coated with fibroblasts of cardiac origin. Immunocytochemistry showed that cultured fibroblasts underwent a phenotype switch to SMA-positive myofibroblasts that expressed connexin 43 and 45 both among themselves and at contact sites with cardiomyocytes. Myofibroblasts affected and dV/dtmax in a cell density-dependent manner; a gradual increase of myofibroblast-to-cardiomyocyte ratios up to 7:100 caused an increase of both and dV/dtmax, which was followed by a progressive decline at higher ratios. On full coverage of the strands with myofibroblasts (ratio >20:100), fell <200 mm/s. This biphasic dependence of and dV/dtmax on myofibroblast density is reminiscent of "supernormal conduction" and is explained by a myofibroblast density-dependent gradual depolarization of the cardiomyocyte strands from eC78 mV to eC50 mV as measured using microelectrode recordings. These findings suggest that myofibroblasts, apart from their role in structural remodeling, might contribute to arrhythmogenesis by direct electrotonic modulation of conduction and that prevention of their appearance might represent an antiarrhythmic therapeutic target.
Key Words: electrophysiology slow conduction cardiac myofibroblasts fibrosis gap junctions
Introduction
Two thirds of the cells of normal hearts are noncardiomyocytes, with fibroblasts constituting the largest fraction. At 2 months of age, fibroblasts outnumber cardiomyocytes by a factor of 2 in human hearts.1eC3 Under physiological conditions, fibroblasts are responsible for providing cardiomyocytes with a mechanical scaffold, which integrates the contractile activity of individual cells so as to result in the coordinated pump function of the organ. Accordingly, fibroblasts are found throughout the myocardium, where they form a 3D cellular network surrounding groups of cardiomyocytes.4 The integrity of this structure is adversely affected by a large number of cardiac diseases ranging from volume to pressure overload and to myocardial infarction. Under these pathological conditions, complex reactions involving changes in extracellular matrix production, cell proliferation, and cell death cause structural remodeling of the ventricular wall, which compromises pump function and predisposes the heart to arrhythmias.5 Moreover, it has been shown that these disease states are associated with the appearance of myofibroblasts.6,7 This cell type, which plays a central role in wound healing in general, is characterized by de novo expression of -smooth muscle actin (SMA).8 In the heart, myofibroblasts are found, for example, in hypertensive heart disease and in infarcted myocardia, where they are involved in the establishment of fibrosis and the formation of the infarct scar.6,7
Based on previous studies showing that myofibroblasts express connexins in tissues other than heart,9,10 the question arises whether this cell type might similarly be capable of forming functional gap junctions in diseased myocardia. If this were to be the case, the intriguing possibility arises that myofibroblasts might be involved in arrhythmogenesis not only by contributing to the formation of electrically insulating collagenous septa causing discontinuous and zig-zag conduction11,12 but also by direct electrotonic modulation of impulse conduction. Accordingly, in the present study, we investigated whether impulse propagation along strands of cultured cardiomyocytes might be affected by the presence of cardiac fibroblasts. Immunocytochemistry identified these fibroblasts as connexin-expressing myofibroblasts, which depressed conduction along the strands in a density-dependent manner and caused uniform slow conduction on completely covering the cardiomyocytes. The findings suggest that myofibroblasts appearing in diseased myocardia might contribute to arrhythmogenesis by direct electrotonic modulation of impulse propagation.
Materials and Methods
Patterned Growth Cell Cultures
Patterned growth cell cultures from neonatal rat hearts were prepared according to previously published procedures.13 Experiments were approved by the state veterinary department. Briefly, hearts from 6 to 10 neonatal rats (Wistar; 1 to 2 days old; Zentrale Tierstlle, Inselspital Bern) were excised, the ventricles were minced, and the resulting small tissue pieces were dissociated in Hank’s balanced salt solution (HBSS; without Ca2+ and Mg2+; Bioconcept) containing trypsin (0.1%; Roche Diagnostics) and pancreatin (60 e/mL; Sigma). The dispersed cells were, after centrifugation, resuspended in medium 199 (M-199) with HBSS (Sigma) containing 10% neonatal calf serum (Bioconcept), penicillin (20 000 U/L; Fakola), streptomycin (34 eol/L; Fakola), and vitamin B12 (15 eol/L; Sigma). Dissociated ventricular myocytes were preplated for 2 hours in culture flasks at a density <1x103 cells/mm2 to reduce the percentage of nonmyocardial cells.14 Cardiomyocytes were seeded at a density of 1.5x103 cells/mm2 on coverslips pretreated to result in patterned growth cell strands measuring 80 ex10 mm.13 The preparations were grown in supplemented medium M-199 (see above) containing, in addition, vitamin C (18 eol/L; Sigma) and epinephrine (10 eol/L; Sigma). After 1 day, the serum content was reduced to 5%. Bromodeoxyuridine (BrdU; 100 eol/L; Sigma) was routinely added to the growth medium of control cardiomyocyte cultures.
Fibroblast Coating
Flasks containing preplated cardiac fibroblasts were kept in culture with supplemented M-199 for 8 days. Before being harvested with trypsin containing dissociation buffer, fibroblasts were live stained for 20 minutes with 5 eol/L DiI (Invitrogen) dissolved in supplemented M-199. After dissociation, fibroblasts were washed and resuspended in the same medium before being seeded at densities up to 400 cells/mm2 onto 1-day-old cardiomyocyte preparations. Experiments were performed with 3- to 4-day-old preparations.
Optical Recording of Electrical Activation
The characteristics of impulse propagation were assessed optically after staining the preparations for 5 minutes with the voltage sensitive dye di-8-ANEPPS (135 eol/L; Biotium). Changes in fluorescence corresponding to transmembrane voltage changes were assessed using a custom-made fiber optic recording setup described in detail previously.15 Recordings were made with a x20, 0.75 numerical aperture objective, which permitted the measurement of impulse propagation characteristics over a distance of 750 e with a spatial resolution of 50 e.
Experimental Protocol
After staining, preparations were superfused with HBSS (Sigma) containing (mmol/L) 137 NaCl, 5.4 KCl, 1.3 CaCl2, 0.8 MgSO4, 4.2 NaHCO3, 0.5 KH2PO4, 0.3 NaH2PO4, and 10 HEPES, which was titrated to pH 7.40 with 1 mol/L NaOH and contained 1% serum. Tetrodotoxin (TTX) used in one series of experiments was obtained from Latoxan. Preparations were stimulated at 2 Hz with rectangular pulses (duration 1 ms; suprathreshold intensity) for 10 s before a given optical recording. All experiments were performed at 36°C.
Data Analysis
Optical raw data were digitally low-pass filtered at corner frequencies (fo) ranging from 0.1 to 0.5 kHz, and action potential amplitudes (APA) were set to 100%. Assuming an average APA of 100 mV, the scaled values given as %APA translate directly into millivolts. Local activation times for each recording site were determined as described before and conduction velocities (; mm/s) were calculated from the slope of a linear least square fit of activation times along the preparation.15 Values for maximal upstroke velocities (dV/dtmax) were calculated in relation to %APA and are given as %APA/ms. Under the assumption of an average APA of 100 mV, %APA/ms corresponds to V/s.
Microelectrode Recordings
Maximal diastolic potentials (MDPs) in strands of cardiomyocytes and in monolayer cultures of myofibroblasts were assessed using conventional microelectrode recording techniques as described previously (for details, see the online supplement, available at http://circres.ahajournals.org).16
Immunocytochemistry
Cultured preparations were stained for vimentin, SMA, connexin 43 (Cx43), and Cx45 using protocols outlined in detail in the online supplement. In those experiments in which fibroblast densities were correlated with optical or electrical measurements, measurement locations were stored during the experiments (PC-controlled custom-made x-y table for the microscope) for later recall after preparations had undergone immunocytochemistry.
Fibroblast Density
The density of fibroblasts was determined by counting live-stained or immunocytochemically identified fibroblasts within the strand sections imaged during optical recordings (750x80 e; total area 60 000 e2). Because strand sections had identical dimensions in all experiments, fibroblast densities are given as cell count per measurement area (MA).
Statistics
Values are given as mean±SD. Data were compared using the Student t test (two-tailed; homoscedastic or heteroscedastic where appropriate), and differences were considered significant at P<0.05.
Results
Phenotype of Cultured Fibroblasts
Cardiac fibroblasts in cell culture tend to undergo a phenotype switch to myofibroblasts.17,18 The hallmark of this transition, which is sensitive to culture conditions, is the de novo expression of SMA.19,20 As depicted by Figure 1A, fibroblasts grown under our culture conditions showed abundant expression of SMA and vimentin 24 hours to 72 hours after plating on glass coverslips, which identifies these cells as myofibroblasts. Because rigid substrates have been shown to favor the transition of fibroblasts to myofibroblasts in culture,20 SMA expression was additionally assessed in fibroblasts growing on top of cardiomyocytes (Figure 1B). Despite the fact that cardiomyocytes can be considered to represent a "soft" substrate for fibroblasts, SMA expression was not compromised, which indicates that fibroblasts also maintained the myofibroblast phenotype when grown in this more in vivo-like configuration. Double immunolabeling for SMA and Cx43 and Cx45, respectively, revealed that myofibroblasts of cardiac origin express both types of connexin and that these connexins are located at contact sites among myofibroblasts and between myofibroblasts and cardiomyocytes (Figure 2A). As illustrated by the high magnification images of a double-immunolabeling experiment for Cx43 and Cx45 in Figure 2B, both connexins showed a punctate expression pattern along contact sites between cardiomyocytes and myofibroblasts (Figure 2B).
Modification of Impulse Conduction by a Coat of Myofibroblasts
The effects of myofibroblasts on impulse conduction along linear strands of cardiomyocytes were investigated in preparations that received a coat of myofibroblasts of cardiac origin after 24 hours in culture. Before seeding, myofibroblasts were live stained with DiI to permit the assessment of their density and spatial distribution along the strands. Myofibroblast densities are given as cell count per measurement area (MA). MAs were identical for all experiments and corresponded to the sections of the preparations imaged during optical recordings (750x80 e). Two days after seeding, the density of myofibroblasts on top of cardiomyocyte strands ranged from 0 to 30 cells per MA, where 30 cells per MA typically indicated complete coverage of the cardiomyocyte strands by myofibroblasts. Given that MAs contained on average 132 cardiomyocytes (determined in phase contrast images of 80 e wide control strands; n=20), complete coverage of cardiomyocytes by myofibroblasts occurred at a myofibroblast-to-cardiomyocyte ratio of 23:100. An example of an optical measurement of impulse propagation in a preparation with only a few adherent myofibroblasts is shown in Figure 3A. Maximal upstroke velocities (dV/dtmax; 26.4±1.4 %APA/ms; fo=0.1 kHz; n=15) and conduction velocity (; 377 mm/s) were fast and propagation was uniform. In contrast and as shown in Figure 3B, preparations with an increased myofibroblast density showed a marked decline in both dV/dtmax (19.4±2.1 %APA/ms; fo=0.1 kHz; n=15) and (243 mm/s), whereas propagation remained uniform. The summary of all individual measurements (n=137) shown in Figure 4A illustrates that increasing myofibroblast densities were accompanied by a gradual decline of from 400 mm/s (no attached myofibroblasts) to 170 mm/s (complete coverage of cardiomyocyte strands by myofibroblasts). Binning of these data (Figure 4B) revealed a myofibroblast density-dependent decay of , which was best fitted by an exponential (r2=0.998). Similar to the decay of , dV/dtmax progressively declined with increasing myofibroblast density (Figure 4C and 4D).
Slowing of Conduction by Endogenous Myofibroblasts
Whereas the experiments above illustrate the extent by which exogenously added myofibroblasts affect conduction, they did not take into account the effects of endogenous myofibroblasts that are invariably present in primary cultures of cardiomyocytes.16,21 To investigate the possibility that this background of endogenous myofibroblasts by itself modulates conduction, we determined impulse propagation characteristics in preparations obtained under the following culture conditions known to modify the myofibroblast content: (1) omission of preplating of the cell suspension obtained after trypsinization; (2) standard conditions (ie, use of preplating); (3) standard conditions with BrdU added to the growth medium. As shown by the immunostainings of 3-day-old preparations in Figure 5A, omission of preplating resulted in preparations being highly contaminated by endogenous myofibroblasts (29.3±8.1 cells per MA; n=42). This number was substantially reduced (9.6±4.4 cells per MA; n=39) after preplating, whereas the additional inclusion of BrdU had no further significant effect on the myofibroblast content (9.5±3.0 cells per MA; n=44) in these 3-day-old preparations. As shown in Figure 5B, and dV/dtmax were highest in the preplating/BrdU group (450±34 mm/s; 74.8±3.2 %APA/ms; fo=0.5 kHz; n=44) and were slightly reduced in the absence of BrdU (430±54 mm/s; 74.6±6.8 %APA/ms; n=39). Omission of preplating resulted in a significant reduction of both and dV/dtmax to 315±68 mm/s and 49.2±11.0 %APA/ms, respectively (n=42), indicating that endogenous myofibroblasts acted similarly on conduction as exogenously added myofibroblasts.
Interestingly, when combining all data and plotting both and dV/dtmax as a function of the density of endogenous myofibroblasts (Figure 6), both parameters showed a biphasic dependence; with increasing myofibroblast density, and dV/dtmax first increased to reach a peak at 9 myofibroblasts per MA before declining to values exhibiting the characteristics of slow conduction at higher myofibroblast densities. This result is highly reminiscent of the biphasic dependence of and dV/dtmax on the concentration of extracellular potassium, which, by modulating the resting membrane potential, induces the well-known phenomenon of "supernormal conduction."22,23 Accordingly, this biphasic relationship suggests that myofibroblasts might affect conduction by a cell density-dependent gradual depolarization of the cardiomyocytes.
Mechanism of Myofibroblast Induced Slowing of Conduction
Based on previous reports showing that cultured cardiac fibroblasts are less polarized (eC20 to eC40 mV) than cardiomyocytes (eC60 to eC80 mV),16,24,25 it can be hypothesized that on gap junctional coupling, fibroblasts cause partial depolarization of cardiomyocytes into the range of sodium channel inactivation, thus causing slowing of conduction. Alternatively, if cardiomyocytes should be able to fully compensate for the depolarizing influence of fibroblasts25 or if fibroblasts were to be well polarized themselves,26 slowing of conduction might be explained by the capacitive load exerted by the fibroblast cell membranes on the cardiomyocytes. To differentiate between these two mechanisms, we assessed the effect of 20 eol/L TTX on measured along cardiomyocyte strands coated with myofibroblasts. As depicted by Figure 7, TTX failed to affect in preparations exhibiting myofibroblast-induced slow conduction (150 mm/s) whereas with decreasing myofibroblast density, gradually increased and TTX became progressively more effective in slowing conduction. In preparations with a low myofibroblast count and, hence, fast conduction, TTX reduced by up to 60%, which is in accordance with values reported previously for cardiomyocyte strands not coated with myofibroblasts.27 These findings suggest that increasing numbers of myofibroblasts, which themselves exhibited a membrane potential of eC20.4±4.0 mV (n=14), caused a progressive depolarization of the cardiomyocytes, which rendered the preparations ultimately insensitive to TTX. We assessed this hypothesis by correlating MDPs measured in cardiomyocyte strands to the local overall density of myofibroblasts (endogenous and exogenous). As shown by the scatter plot of all measurements in Figure 8A, increasing myofibroblast densities caused a gradual decline in MDP from around eC80 mV to values around eC50 mV. Binning of the data (Figure 8B) resulted in a relationship between MDP and myofibroblast density that was best fitted with a power function (r2=0.970).
Discussion
The results of this study indicate that heterocellular gap junctional coupling between cardiomyocytes and myofibroblasts of cardiac origin substantially influences the characteristics of cardiac impulse propagation. In particular, the study shows that: (1) cardiac fibroblasts in culture undergo a phenotype switch to myofibroblasts that express Cx43 and Cx45 both among themselves and with cardiomyocytes; (2) myofibroblasts slow conduction primarily by partial depolarization of cardiomyocytes; and (3) the presence of endogenous myofibroblasts importantly determines the electrophysiological characteristics of primary cardiac cell cultures.
Fibroblasts, Myofibroblasts, and Connexin Expression
Proliferation and activity of extracellular matrix producing cells is central to tissue fibrosis. In the normal myocardium, interstitial fibroblasts are responsible for collagen synthesis, whereas after infarction or in the context of hypertensive heart disease, phenotypically transformed fibroblasts termed myofibroblasts additionally participate in fibrogenesis.6,28,29 This cell type, which is not normally found in healthy hearts with the exception of valve leaflets, is characterized by the expression of SMA.8 As in the case of diseased hearts, cultured cardiac fibroblasts undergo a phenotype switch to myofibroblasts, and this process has been shown to be favored, for example, by hyperoxic culture conditions or by growing fibroblasts on rigid substrates.18eC20 Whereas these results were obtained in monolayer cultures, we show in the present study that fibroblasts of cardiac origin develop and maintain the myofibroblast phenotype also when grown in a spatial configuration resembling the in vivo situation (ie, when layered on top of cardiomyocytes). Because cardiomyocytes can be considered to represent a "soft substrate" for the fibroblasts, the observed phenotype switch is likely to be dependent on additional mechanisms such as, for example, mechanical stretch exerted by the contracting cardiomyocytes.30
The question of whether fibroblasts in intact hearts express gap junctions among themselves and with cardiomyocytes is still debated. Whereas early studies failed to show robust evidence for homocellular or heterocellular gap junctional coupling,31 such coupling has recently been described in vivo in the sinoatrial node,32 whereas for fibroblasts residing in the ventricular wall of healthy myocardium, the presence of connexins is still disputed.4,33 Regarding myofibroblasts, expression of gap junctions has been shown to occur in tissues other than heart.9,10,34 The results of the present study extend these findings to cardiac tissue by showing that myofibroblasts of cardiac origin express Cx43 and Cx45 both among themselves and with cardiomyocytes, thus forming the basis of potential heterocellular electrotonic interactions. Whereas direct experimental proof of such homocellular/heterocellular gap junctional coupling in intact cardiac tissue is still missing, the findings that healing infarcts are mainly populated by myofibroblasts,6,35 and noncardiomyocytes populating healing infarcts in sheep hearts display abundant expression of Cx43 and Cx4533 provide indirect evidence that myofibroblasts appearing in diseased hearts might indeed have the potential to form homocellular/heterocellular gap junctions by which they possibly interfere with normal cardiac electrophysiology.
Myofibroblast-Induced Slow Conduction
Although the role of myofibroblasts as a central element in structural cardiac remodeling after various injuries to the heart is well established, the question of whether this cell type might contribute to arrhythmogenesis by direct electrotonic interaction with cardiomyocytes is still open. Cell culture experiments have shown previously that heterocellular electrical coupling between cardiomyocytes and fibroblasts (which, based on the present and other studies,18,20 were presumably myofibroblasts) substantially influences impulse propagation. In monolayer cultures of neonatal rat ventricular cardiomyocytes, it was demonstrated that clusters of fibroblasts transfected with the voltage-sensitive potassium channel Kv1.3 induce local conduction blocks.36 Moreover, it was shown that fibroblasts of cardiac origin are capable of relaying electrical activation between strands of cardiomyocytes for distances up to 300 e.37 In contrast to these investigations addressing the effects of nonuniformly distributed fibroblasts on conduction, the present study shows that uniformly distributed myofibroblasts as might be found in diffuse cardiac fibrosis substantially reduce and dV/dtmax as a function of myofibroblast density. Uniform slow conduction (<200 mm/s) was observed at myofibroblast-to-cardiomyocyte ratios >19:100 (> 27:100 when corrected for endogenous myofibroblasts; see below). At these densities, myofibroblasts, which exhibited membrane potentials (eC14 to eC25 mV) falling within the range of potentials reported before by others for cultured cardiac fibroblasts (eC10 to eC20 mV; eC20 to eC40 mV),24,25 depressed the MDP of cardiomyocytes to values
Endogenous Fibroblasts and Conduction
Primary cultures of cardiomyocytes are invariably "contaminated" by noncardiomyocytes among which fibroblasts represent the largest fraction. Because these cells tend to overgrow the cardiomyocytes with time in culture, measures that reduce their number in the initial cell suspension (preplating, density gradient centrifugation) and that inhibit their proliferation (BrdU, cytosine arabinofuranoside, irradiation) are routinely used in the establishment of these cultures.14,16,21 As determined in 3-day-old control strands obtained from preplated cell suspensions, the average density of endogenous myofibroblasts amounted to 10 cells per MA, which corresponds to a myofibroblast-to-cardiomyocyte ratio of 8:100. In contrast, omission of preplating resulted in cell strands containing a substantially higher percentage of endogenous myofibroblasts (22:100), which caused a significant reduction of both and dV/dtmax.
Interestingly, when combining the data of all the three types of preparations and plotting and dV/dtmax as a function of the number of myofibroblasts per MA, the resulting relationship shows a biphasic shape with peak values for either parameter occurring at a myofibroblast-to-cardiomyocyte ratio of 7:100. Based on the finding that myofibroblasts gradually depolarize cardiomyocyte strands in a cell density-dependent manner up to levels of eC50 mV, where they become TTX insensitive, this biphasic relationship can be explained by the well-known phenomenon of supernormal conduction, which accompanies gradual depolarizations of cardiac tissue by potassium.22,23
The finding that supernormal conduction occurred at myofibroblast-to-cardiomyocyte ratios (7:100) similar to the average myofibroblast-to-cardiomyocyte ratio found in uncoated control strands (8:100) has two implications. It explains, why and dV/dtmax declined monotonically in the case of exogenously added myofibroblasts (Figure 4). In these experiments, changes in and dV/dtmax were related to the number of exogenously added myofibroblasts only. If these numbers are corrected for "background myofibroblasts" by adding the average number of endogenous myofibroblasts present in uncoated strands (10), the relationship is shifted to the right and coincides, as expected, with the monotonically descending right limb of the biphasic relationship described above. It also predicts that the dependence of on extracellular potassium concentration ([K+]o) in control strands will equally show a peak at potassium concentrations close to that of the superfusion solution ([K+]o=5.4 mmol/L). This is indeed what has been reported previously for cultured cardiomyocyte strands,27 and it explains in retrospect why levels of [K+]o inducing supernormal conduction in these preparations (5.8 mmol/L) tend to be lower than those observed in intact tissue (up to 10.8 mmol/L)38; ie, endogenous myofibroblasts present in cultured preparations "pre-depolarize" the cardiomyocytes at normal [K+]o sufficiently to elicit supernormal conduction, and any further increase of [K+]o is bound to slow conduction by progressive inactivation of sodium channels.
These findings indicate that heterocellular electrical interactions between myofibroblasts and cardiomyocytes can modify the set point of occurrence of supernormal conduction in respect to [K+]o. Moreover, the results illustrate that myofibroblasts play a key role in the determination of conduction characteristics of cultured myocardial cells that needs to be taken into account when aiming at establishing culture models exhibiting close to in vivo properties in regard to and dV/dtmax.
Study Limitations
Although the findings of this study indicate that impulse propagation characteristics in networks of cardiomyocytes in vitro are strongly affected by the presence of electrically coupled myofibroblasts, the extrapolation of these data to the situation in vivo has to await future characterizations of myofibroblasts in diseased hearts in respect to their size and relative density, their cellular electrophysiology, and their ability to form homocellular and heterocellular gap junctions. Under the assumption that these characteristics are similar to those shown in the present study, the extent to which myofibroblasts will affect conduction in intact diseased hearts will depend on such additional factors as: (1) differences in current density and current composition of adult cardiomyocytes versus cultured neonatal cardiomyocytes, (2) differences in size of cells in situ versus cultured cells, (3) the degree of homocellular and heterocellular gap junctional coupling, (4) the ratio of myofibroblasts to cardiomyocytes within electrotonic reach of each other, and (5) the specific cellular tissue architecture of regions where myofibroblasts intermingle with cardiomyocytes. The combination of all of these factors will ultimately determine to which extent source-to-sink relationships are changed and, thus, conduction is affected in vivo.
Perspectives
Although there exists abundant information regarding the role of myofibroblasts in structural remodeling after a vast array of injuries to the heart, the possibility that this cell type might directly interact with cardiomyocyte electrophysiology via gap junctions has received virtually no attention in the past. In this context, the results of the present study open the perspective that sites of intimate contact between myofibroblasts and cardiomyocytes in diseased hearts (eg, infarct border zones) might contribute to local arrhythmogenic slowing of conduction. Accordingly, controlling the phenotype and proliferation of myofibroblasts may represent a new and specific therapeutic target in the management of cardiac arrhythmias. Indeed, antifibrotic therapies aimed at reversing structural remodeling by influencing metabolic and proliferative activities of myofibroblasts have been shown to be antiarrhythmic, and it remains to be investigated whether, apart from influences on the cellular tissue architecture, a loss of electrotonic interactions between myofibroblasts and cardiomyocytes might also be involved in this antiarrhythmic effect.
Acknowledgments
This study was supported by the Swiss University Conference and by the Swiss National Science Foundation (grant 3100-105916 to S.R.). We thank Regula Fle筩kiger-Labrada for her expert technical assistance and Professor H.P. Clamann for helpful discussions on this manuscript.
References
Adler CP, Ringlage WP, Bhm N. DNS-Gehalt und Zellzahl in Herz und Leber von Kindern. Pathol Res Pract. 1981; 172.
Maisch B. Extracellular matrix and cardiac interstitium: restriction is not a restricted phenomenon. Herz. 1995; 20: 75eC80.
Nag AC. Study of non-muscle cells of the adult mammalian heart: a fine structural analysis and distribution. Cytobios. 1980; 28: 41eC61.
Goldsmith EC, Hoffman A, Morales MO, Potts JD, Price RL, McFadden A, Rice M, Borg TK. Organization of fibroblasts in the heart. Dev Dyn. 2004; 230.
Manabe I, Shindo T, Nagai R. Gene expression in fibroblasts and fibrosis: involvement in cardiac hypertrophy. Circ Res. 2002; 91: 1103eC1113.
Sun Y, Kiami MF, Postlethwaite AE, Weber KT. Infarct scar as living tissue. Basic Res Cardiol. 2002; 97: 343eC347.
Weber KT. Fibrosis in hypertensive heart disease: focus on cardiac fibroblasts. J Hypertens. 2004; 22: 47eC50.
Darby I, Skalli O, Gabbiani G. -smooth muscle actin is transiently expressed by myofibroblasts during experimental wound healing. Lab Invest. 1990; 63: 21eC29.
Gabbiani G, Chaponnier C, Huttner I. Cytoplasmic filaments and gap junctions in epithelial cells and myofibroblasts during wound healing. J Cell Biol. 1978; 76: 561eC648.
Spanakis SG, Petridou S, Masur SK. Functional gap junctions in corneal fibroblasts and myofibroblasts. Invest Ophthalmol Vis Sci. 1998; 39: 1320eC1328.
de Bakker JM, van Capelle FJ, Janse MJ, Tasseron S, Vermeulen JT, de Jonge N, Lahpor JR. Slow conduction in the infarcted human heart. ‘zigzag’ course of activation. Circulation. 1993; 88: 915eC926.
Spach MS, Boineau JP. Microfibrosis produces electrical load variations due to loss of side-to-side cell connections: a major mechanism of structural heart disease arrhythmias. Pacing Clin Electrophysiol. 1997; 20: 397eC413.
Rohr S, Fle筩kiger-Labrada R, Kucera J. Photolithographically defined deposition of attachment factors as a versatile method for patterning the growth of different cell types in culture. Eur J Physiol. 2003; 446: 125eC132.
Blondel B, Roijen I, Cheneval JP. Heart cells in culture: a simple method for increasing the proportion of myoblasts. Experientia. 1971; 27: 356eC358.
Rohr S, Kucera JP. Optical recording system based on a fiber optic image conduit: assessment of microscopic activation patterns in cardiac tissue. Biophys J. 1998; 75: 1062eC1075.
Rohr S, Schlly DM, Kleeber AG. Patterned growth of neonatal rat heart cells in culture. Circ Res. 1991; 68: 114eC130.
Brouty-Boye D, Kolonias D, Savaraj N, Lampidis TJ. -smooth muscle actin expression in cultured cardiac fibroblasts of newborn rat. In Vitro Cell Dev Biol. 1992; 28A: 293eC296.
Wang J, Seth A, McCulloch CAG. Force regulates smooth muscle actin in cardiac fibroblasts. Am J Physiol Heart Circ Physiol. 2000; 279: H2776eCH2785.
Roy S, Khanna S, Bickerstaff AA, Subramanian SV, Atalay M, Bierl M, Pendyala S, Levy D, Sharma N, Venojrvi M, Strauch A, Orosz CG, Sen CK. Oxygen sensing by primary cardiac fibroblasts. Circ Res. 2003; 92: 264eC271.
Wang J, Chen H, Seth A, McCulloch CA. Mechanical force regulation of myofibroblast differentiation in cardiac fibroblasts. Am J Physiol Heart Circ Physiol. 2003; 285: H1871eCH1881.
Lokuta A, Kirby MS, Gaa ST, Lederer J, Rogers TB. On establishing primary cultures of neonatal rat ventricular myocytes for analysis over long periods. J Cardiovasc Electrophysiol. 1994; 5: 50eC61.
Kagiyama Y, Hill JL, Gettes LS. Interaction of acidosis and increased extracellular potassium on action potential characteristics and conduction in guinea pig ventricular muscle. Circ Res. 1982; 51: 614eC623.
Shaw RM, Rudy Y. Electrophysiologic effects of acute myocardial ischemia: a mechanistic investigation of action potential conduction and conduction failure. Circ Res. 1997; 80: 124eC138.
Hyde AB, BBlondel, A Matter JP. Homo- and heterocellular junctions in cell cultures: an electrophysiological and morphological study. Prog Brain Res. 1969; 31: 283eC311.
Rook MB, van Ginneken AC, de Jonge B, el Aoumari A, Gros D, Jongsma HJ. Differences in gap junction channels between cardiac myocytes, fibroblasts, and heterologous pairs. Am J Physiol Cell Physiol. 1992; 263: C959eCC977.
Chilton L, Ohya S, Freed D, George E, Drobic V, Shibukawa Y, MacCannell KA, Imaizumi Y, Clark RB, Dixon IMC, Giles WR. K+ currents regulate the resting membrane potential, proliferation, and contractile responses in ventricular fibroblasts and myofibroblasts. Am J Physiol Heart Circ Physiol. 2005; 288: H2931eCH2939.
Rohr S, Kucera JP, Kleber AG. Slow conduction in cardiac tissue. I. Effects of a reduction of excitability vs. a reduction of electrical coupling on microconduction. Circ Res. 1998; 83: 781eC794.
Leslie KO, Taatjes DJ, Schwarz J, von Turkovich M, Low RB. Cardiac myofibroblasts express alpha smooth muscle actin during right ventricular pressure overload in the rabbit. Am J Pathol. 1991; 139: 207eC216.
Weber K. Cardiac interstitium. In: Poole-Wilson P, Colucci WBM, et al, eds. Heart Failure. New York, NY: Churchill Livingstone; 1997: 13eC31.
Hinz B, Mastrangelo D, Iselin CE, Chaponnier C, Gabbiani G. Mechanical tension controls granulation tissue contractile activity and myofibroblast differentiation. Am J Pathol. 2001; 159: 1009eC1020.
De Maziere AMGL, Van Ginneken ACG, Wilders R, Jongsma HJ, Bouman LN. Spatial and functional relationship between myocytes and fibroblasts in the rabbit sinoatrial node. J Mol Cell Cardiol. 1992; 24: 567eC578.
Camelliti P, Green CR, LeGrice I, Kohl P. Fibroblast network in rabbit sinoatrial node: structural and functional identification of homogeneous and heterogeneous cell coupling. Circ Res. 2004; 94: 828eC835.
Camelliti P, Devlin GP, Matthews KG, Kohl P, Green CR. Spatially and temporally distinct expression of fibroblast connexins after sheep ventricular infarction. Cardiovasc Res. 2004; 62: 415eC425.
Powell DW, Mifflin RC, Valentich JD, Crowe SE, Saada JI, West AB. Myofibroblasts. I. Paracrine cells important in health and disease. Am J Physiol Cell Physiol. 1999; 277: C1eCC19.
Cameron Va, Rademaker MT, Ellmers LJ, Expiner EA, Nicholls MG, Richards M. Atrial (ANP) and brain natriuretic peptide (BNP) expression after myocardial infarction in sheep: ANP is synthesized by fibroblasts infiltrating the infarct. Endocrinology. 2000; 141: 4690eC4697.
Feld Y, Melamed-Frank M, Kehat I, Tal D, Marom S, Gepstein L. Electrophysiological modulation of cardiomyocytic tissue by transfected fibroblasts expressing potassium channels: a novel strategy to manipulate excitability. Circ. 2002; 105: 522eC529.
Gaudesius G, Miragoli M, Thomas SP, Rohr S. Coupling of cardiac electrical activity over extended distances by fibroblasts of cardiac origin. Circ Res. 2003; 93: 421eC428.
Hisatome I, Arita M. Effects of catecholamines on the residual sodium channel dependent slow conduction in guinea pig ventricular muscles under normoxia and hypoxia. Cardiovasc Res. 1995; 29: 65eC73.(Michele Miragoli, Giedriu)
Abstract
Structural remodeling of the myocardium associated with mechanical overload or cardiac infarction is accompanied by the appearance of myofibroblasts. These fibroblast-like cells express -smooth muscle actin (SMA) and have been shown to express connexins in tissues other than heart. The present study examined whether myofibroblasts of cardiac origin establish heterocellular gap junctional coupling with cardiomyocytes and whether ensuing electrotonic interactions affect impulse propagation. For this purpose, impulse conduction characteristics (conduction velocity [] and maximal upstroke velocity [dV/dtmax]) were assessed optically in cultured strands of cardiomyocytes, which were coated with fibroblasts of cardiac origin. Immunocytochemistry showed that cultured fibroblasts underwent a phenotype switch to SMA-positive myofibroblasts that expressed connexin 43 and 45 both among themselves and at contact sites with cardiomyocytes. Myofibroblasts affected and dV/dtmax in a cell density-dependent manner; a gradual increase of myofibroblast-to-cardiomyocyte ratios up to 7:100 caused an increase of both and dV/dtmax, which was followed by a progressive decline at higher ratios. On full coverage of the strands with myofibroblasts (ratio >20:100), fell <200 mm/s. This biphasic dependence of and dV/dtmax on myofibroblast density is reminiscent of "supernormal conduction" and is explained by a myofibroblast density-dependent gradual depolarization of the cardiomyocyte strands from eC78 mV to eC50 mV as measured using microelectrode recordings. These findings suggest that myofibroblasts, apart from their role in structural remodeling, might contribute to arrhythmogenesis by direct electrotonic modulation of conduction and that prevention of their appearance might represent an antiarrhythmic therapeutic target.
Key Words: electrophysiology slow conduction cardiac myofibroblasts fibrosis gap junctions
Introduction
Two thirds of the cells of normal hearts are noncardiomyocytes, with fibroblasts constituting the largest fraction. At 2 months of age, fibroblasts outnumber cardiomyocytes by a factor of 2 in human hearts.1eC3 Under physiological conditions, fibroblasts are responsible for providing cardiomyocytes with a mechanical scaffold, which integrates the contractile activity of individual cells so as to result in the coordinated pump function of the organ. Accordingly, fibroblasts are found throughout the myocardium, where they form a 3D cellular network surrounding groups of cardiomyocytes.4 The integrity of this structure is adversely affected by a large number of cardiac diseases ranging from volume to pressure overload and to myocardial infarction. Under these pathological conditions, complex reactions involving changes in extracellular matrix production, cell proliferation, and cell death cause structural remodeling of the ventricular wall, which compromises pump function and predisposes the heart to arrhythmias.5 Moreover, it has been shown that these disease states are associated with the appearance of myofibroblasts.6,7 This cell type, which plays a central role in wound healing in general, is characterized by de novo expression of -smooth muscle actin (SMA).8 In the heart, myofibroblasts are found, for example, in hypertensive heart disease and in infarcted myocardia, where they are involved in the establishment of fibrosis and the formation of the infarct scar.6,7
Based on previous studies showing that myofibroblasts express connexins in tissues other than heart,9,10 the question arises whether this cell type might similarly be capable of forming functional gap junctions in diseased myocardia. If this were to be the case, the intriguing possibility arises that myofibroblasts might be involved in arrhythmogenesis not only by contributing to the formation of electrically insulating collagenous septa causing discontinuous and zig-zag conduction11,12 but also by direct electrotonic modulation of impulse conduction. Accordingly, in the present study, we investigated whether impulse propagation along strands of cultured cardiomyocytes might be affected by the presence of cardiac fibroblasts. Immunocytochemistry identified these fibroblasts as connexin-expressing myofibroblasts, which depressed conduction along the strands in a density-dependent manner and caused uniform slow conduction on completely covering the cardiomyocytes. The findings suggest that myofibroblasts appearing in diseased myocardia might contribute to arrhythmogenesis by direct electrotonic modulation of impulse propagation.
Materials and Methods
Patterned Growth Cell Cultures
Patterned growth cell cultures from neonatal rat hearts were prepared according to previously published procedures.13 Experiments were approved by the state veterinary department. Briefly, hearts from 6 to 10 neonatal rats (Wistar; 1 to 2 days old; Zentrale Tierstlle, Inselspital Bern) were excised, the ventricles were minced, and the resulting small tissue pieces were dissociated in Hank’s balanced salt solution (HBSS; without Ca2+ and Mg2+; Bioconcept) containing trypsin (0.1%; Roche Diagnostics) and pancreatin (60 e/mL; Sigma). The dispersed cells were, after centrifugation, resuspended in medium 199 (M-199) with HBSS (Sigma) containing 10% neonatal calf serum (Bioconcept), penicillin (20 000 U/L; Fakola), streptomycin (34 eol/L; Fakola), and vitamin B12 (15 eol/L; Sigma). Dissociated ventricular myocytes were preplated for 2 hours in culture flasks at a density <1x103 cells/mm2 to reduce the percentage of nonmyocardial cells.14 Cardiomyocytes were seeded at a density of 1.5x103 cells/mm2 on coverslips pretreated to result in patterned growth cell strands measuring 80 ex10 mm.13 The preparations were grown in supplemented medium M-199 (see above) containing, in addition, vitamin C (18 eol/L; Sigma) and epinephrine (10 eol/L; Sigma). After 1 day, the serum content was reduced to 5%. Bromodeoxyuridine (BrdU; 100 eol/L; Sigma) was routinely added to the growth medium of control cardiomyocyte cultures.
Fibroblast Coating
Flasks containing preplated cardiac fibroblasts were kept in culture with supplemented M-199 for 8 days. Before being harvested with trypsin containing dissociation buffer, fibroblasts were live stained for 20 minutes with 5 eol/L DiI (Invitrogen) dissolved in supplemented M-199. After dissociation, fibroblasts were washed and resuspended in the same medium before being seeded at densities up to 400 cells/mm2 onto 1-day-old cardiomyocyte preparations. Experiments were performed with 3- to 4-day-old preparations.
Optical Recording of Electrical Activation
The characteristics of impulse propagation were assessed optically after staining the preparations for 5 minutes with the voltage sensitive dye di-8-ANEPPS (135 eol/L; Biotium). Changes in fluorescence corresponding to transmembrane voltage changes were assessed using a custom-made fiber optic recording setup described in detail previously.15 Recordings were made with a x20, 0.75 numerical aperture objective, which permitted the measurement of impulse propagation characteristics over a distance of 750 e with a spatial resolution of 50 e.
Experimental Protocol
After staining, preparations were superfused with HBSS (Sigma) containing (mmol/L) 137 NaCl, 5.4 KCl, 1.3 CaCl2, 0.8 MgSO4, 4.2 NaHCO3, 0.5 KH2PO4, 0.3 NaH2PO4, and 10 HEPES, which was titrated to pH 7.40 with 1 mol/L NaOH and contained 1% serum. Tetrodotoxin (TTX) used in one series of experiments was obtained from Latoxan. Preparations were stimulated at 2 Hz with rectangular pulses (duration 1 ms; suprathreshold intensity) for 10 s before a given optical recording. All experiments were performed at 36°C.
Data Analysis
Optical raw data were digitally low-pass filtered at corner frequencies (fo) ranging from 0.1 to 0.5 kHz, and action potential amplitudes (APA) were set to 100%. Assuming an average APA of 100 mV, the scaled values given as %APA translate directly into millivolts. Local activation times for each recording site were determined as described before and conduction velocities (; mm/s) were calculated from the slope of a linear least square fit of activation times along the preparation.15 Values for maximal upstroke velocities (dV/dtmax) were calculated in relation to %APA and are given as %APA/ms. Under the assumption of an average APA of 100 mV, %APA/ms corresponds to V/s.
Microelectrode Recordings
Maximal diastolic potentials (MDPs) in strands of cardiomyocytes and in monolayer cultures of myofibroblasts were assessed using conventional microelectrode recording techniques as described previously (for details, see the online supplement, available at http://circres.ahajournals.org).16
Immunocytochemistry
Cultured preparations were stained for vimentin, SMA, connexin 43 (Cx43), and Cx45 using protocols outlined in detail in the online supplement. In those experiments in which fibroblast densities were correlated with optical or electrical measurements, measurement locations were stored during the experiments (PC-controlled custom-made x-y table for the microscope) for later recall after preparations had undergone immunocytochemistry.
Fibroblast Density
The density of fibroblasts was determined by counting live-stained or immunocytochemically identified fibroblasts within the strand sections imaged during optical recordings (750x80 e; total area 60 000 e2). Because strand sections had identical dimensions in all experiments, fibroblast densities are given as cell count per measurement area (MA).
Statistics
Values are given as mean±SD. Data were compared using the Student t test (two-tailed; homoscedastic or heteroscedastic where appropriate), and differences were considered significant at P<0.05.
Results
Phenotype of Cultured Fibroblasts
Cardiac fibroblasts in cell culture tend to undergo a phenotype switch to myofibroblasts.17,18 The hallmark of this transition, which is sensitive to culture conditions, is the de novo expression of SMA.19,20 As depicted by Figure 1A, fibroblasts grown under our culture conditions showed abundant expression of SMA and vimentin 24 hours to 72 hours after plating on glass coverslips, which identifies these cells as myofibroblasts. Because rigid substrates have been shown to favor the transition of fibroblasts to myofibroblasts in culture,20 SMA expression was additionally assessed in fibroblasts growing on top of cardiomyocytes (Figure 1B). Despite the fact that cardiomyocytes can be considered to represent a "soft" substrate for fibroblasts, SMA expression was not compromised, which indicates that fibroblasts also maintained the myofibroblast phenotype when grown in this more in vivo-like configuration. Double immunolabeling for SMA and Cx43 and Cx45, respectively, revealed that myofibroblasts of cardiac origin express both types of connexin and that these connexins are located at contact sites among myofibroblasts and between myofibroblasts and cardiomyocytes (Figure 2A). As illustrated by the high magnification images of a double-immunolabeling experiment for Cx43 and Cx45 in Figure 2B, both connexins showed a punctate expression pattern along contact sites between cardiomyocytes and myofibroblasts (Figure 2B).
Modification of Impulse Conduction by a Coat of Myofibroblasts
The effects of myofibroblasts on impulse conduction along linear strands of cardiomyocytes were investigated in preparations that received a coat of myofibroblasts of cardiac origin after 24 hours in culture. Before seeding, myofibroblasts were live stained with DiI to permit the assessment of their density and spatial distribution along the strands. Myofibroblast densities are given as cell count per measurement area (MA). MAs were identical for all experiments and corresponded to the sections of the preparations imaged during optical recordings (750x80 e). Two days after seeding, the density of myofibroblasts on top of cardiomyocyte strands ranged from 0 to 30 cells per MA, where 30 cells per MA typically indicated complete coverage of the cardiomyocyte strands by myofibroblasts. Given that MAs contained on average 132 cardiomyocytes (determined in phase contrast images of 80 e wide control strands; n=20), complete coverage of cardiomyocytes by myofibroblasts occurred at a myofibroblast-to-cardiomyocyte ratio of 23:100. An example of an optical measurement of impulse propagation in a preparation with only a few adherent myofibroblasts is shown in Figure 3A. Maximal upstroke velocities (dV/dtmax; 26.4±1.4 %APA/ms; fo=0.1 kHz; n=15) and conduction velocity (; 377 mm/s) were fast and propagation was uniform. In contrast and as shown in Figure 3B, preparations with an increased myofibroblast density showed a marked decline in both dV/dtmax (19.4±2.1 %APA/ms; fo=0.1 kHz; n=15) and (243 mm/s), whereas propagation remained uniform. The summary of all individual measurements (n=137) shown in Figure 4A illustrates that increasing myofibroblast densities were accompanied by a gradual decline of from 400 mm/s (no attached myofibroblasts) to 170 mm/s (complete coverage of cardiomyocyte strands by myofibroblasts). Binning of these data (Figure 4B) revealed a myofibroblast density-dependent decay of , which was best fitted by an exponential (r2=0.998). Similar to the decay of , dV/dtmax progressively declined with increasing myofibroblast density (Figure 4C and 4D).
Slowing of Conduction by Endogenous Myofibroblasts
Whereas the experiments above illustrate the extent by which exogenously added myofibroblasts affect conduction, they did not take into account the effects of endogenous myofibroblasts that are invariably present in primary cultures of cardiomyocytes.16,21 To investigate the possibility that this background of endogenous myofibroblasts by itself modulates conduction, we determined impulse propagation characteristics in preparations obtained under the following culture conditions known to modify the myofibroblast content: (1) omission of preplating of the cell suspension obtained after trypsinization; (2) standard conditions (ie, use of preplating); (3) standard conditions with BrdU added to the growth medium. As shown by the immunostainings of 3-day-old preparations in Figure 5A, omission of preplating resulted in preparations being highly contaminated by endogenous myofibroblasts (29.3±8.1 cells per MA; n=42). This number was substantially reduced (9.6±4.4 cells per MA; n=39) after preplating, whereas the additional inclusion of BrdU had no further significant effect on the myofibroblast content (9.5±3.0 cells per MA; n=44) in these 3-day-old preparations. As shown in Figure 5B, and dV/dtmax were highest in the preplating/BrdU group (450±34 mm/s; 74.8±3.2 %APA/ms; fo=0.5 kHz; n=44) and were slightly reduced in the absence of BrdU (430±54 mm/s; 74.6±6.8 %APA/ms; n=39). Omission of preplating resulted in a significant reduction of both and dV/dtmax to 315±68 mm/s and 49.2±11.0 %APA/ms, respectively (n=42), indicating that endogenous myofibroblasts acted similarly on conduction as exogenously added myofibroblasts.
Interestingly, when combining all data and plotting both and dV/dtmax as a function of the density of endogenous myofibroblasts (Figure 6), both parameters showed a biphasic dependence; with increasing myofibroblast density, and dV/dtmax first increased to reach a peak at 9 myofibroblasts per MA before declining to values exhibiting the characteristics of slow conduction at higher myofibroblast densities. This result is highly reminiscent of the biphasic dependence of and dV/dtmax on the concentration of extracellular potassium, which, by modulating the resting membrane potential, induces the well-known phenomenon of "supernormal conduction."22,23 Accordingly, this biphasic relationship suggests that myofibroblasts might affect conduction by a cell density-dependent gradual depolarization of the cardiomyocytes.
Mechanism of Myofibroblast Induced Slowing of Conduction
Based on previous reports showing that cultured cardiac fibroblasts are less polarized (eC20 to eC40 mV) than cardiomyocytes (eC60 to eC80 mV),16,24,25 it can be hypothesized that on gap junctional coupling, fibroblasts cause partial depolarization of cardiomyocytes into the range of sodium channel inactivation, thus causing slowing of conduction. Alternatively, if cardiomyocytes should be able to fully compensate for the depolarizing influence of fibroblasts25 or if fibroblasts were to be well polarized themselves,26 slowing of conduction might be explained by the capacitive load exerted by the fibroblast cell membranes on the cardiomyocytes. To differentiate between these two mechanisms, we assessed the effect of 20 eol/L TTX on measured along cardiomyocyte strands coated with myofibroblasts. As depicted by Figure 7, TTX failed to affect in preparations exhibiting myofibroblast-induced slow conduction (150 mm/s) whereas with decreasing myofibroblast density, gradually increased and TTX became progressively more effective in slowing conduction. In preparations with a low myofibroblast count and, hence, fast conduction, TTX reduced by up to 60%, which is in accordance with values reported previously for cardiomyocyte strands not coated with myofibroblasts.27 These findings suggest that increasing numbers of myofibroblasts, which themselves exhibited a membrane potential of eC20.4±4.0 mV (n=14), caused a progressive depolarization of the cardiomyocytes, which rendered the preparations ultimately insensitive to TTX. We assessed this hypothesis by correlating MDPs measured in cardiomyocyte strands to the local overall density of myofibroblasts (endogenous and exogenous). As shown by the scatter plot of all measurements in Figure 8A, increasing myofibroblast densities caused a gradual decline in MDP from around eC80 mV to values around eC50 mV. Binning of the data (Figure 8B) resulted in a relationship between MDP and myofibroblast density that was best fitted with a power function (r2=0.970).
Discussion
The results of this study indicate that heterocellular gap junctional coupling between cardiomyocytes and myofibroblasts of cardiac origin substantially influences the characteristics of cardiac impulse propagation. In particular, the study shows that: (1) cardiac fibroblasts in culture undergo a phenotype switch to myofibroblasts that express Cx43 and Cx45 both among themselves and with cardiomyocytes; (2) myofibroblasts slow conduction primarily by partial depolarization of cardiomyocytes; and (3) the presence of endogenous myofibroblasts importantly determines the electrophysiological characteristics of primary cardiac cell cultures.
Fibroblasts, Myofibroblasts, and Connexin Expression
Proliferation and activity of extracellular matrix producing cells is central to tissue fibrosis. In the normal myocardium, interstitial fibroblasts are responsible for collagen synthesis, whereas after infarction or in the context of hypertensive heart disease, phenotypically transformed fibroblasts termed myofibroblasts additionally participate in fibrogenesis.6,28,29 This cell type, which is not normally found in healthy hearts with the exception of valve leaflets, is characterized by the expression of SMA.8 As in the case of diseased hearts, cultured cardiac fibroblasts undergo a phenotype switch to myofibroblasts, and this process has been shown to be favored, for example, by hyperoxic culture conditions or by growing fibroblasts on rigid substrates.18eC20 Whereas these results were obtained in monolayer cultures, we show in the present study that fibroblasts of cardiac origin develop and maintain the myofibroblast phenotype also when grown in a spatial configuration resembling the in vivo situation (ie, when layered on top of cardiomyocytes). Because cardiomyocytes can be considered to represent a "soft substrate" for the fibroblasts, the observed phenotype switch is likely to be dependent on additional mechanisms such as, for example, mechanical stretch exerted by the contracting cardiomyocytes.30
The question of whether fibroblasts in intact hearts express gap junctions among themselves and with cardiomyocytes is still debated. Whereas early studies failed to show robust evidence for homocellular or heterocellular gap junctional coupling,31 such coupling has recently been described in vivo in the sinoatrial node,32 whereas for fibroblasts residing in the ventricular wall of healthy myocardium, the presence of connexins is still disputed.4,33 Regarding myofibroblasts, expression of gap junctions has been shown to occur in tissues other than heart.9,10,34 The results of the present study extend these findings to cardiac tissue by showing that myofibroblasts of cardiac origin express Cx43 and Cx45 both among themselves and with cardiomyocytes, thus forming the basis of potential heterocellular electrotonic interactions. Whereas direct experimental proof of such homocellular/heterocellular gap junctional coupling in intact cardiac tissue is still missing, the findings that healing infarcts are mainly populated by myofibroblasts,6,35 and noncardiomyocytes populating healing infarcts in sheep hearts display abundant expression of Cx43 and Cx4533 provide indirect evidence that myofibroblasts appearing in diseased hearts might indeed have the potential to form homocellular/heterocellular gap junctions by which they possibly interfere with normal cardiac electrophysiology.
Myofibroblast-Induced Slow Conduction
Although the role of myofibroblasts as a central element in structural cardiac remodeling after various injuries to the heart is well established, the question of whether this cell type might contribute to arrhythmogenesis by direct electrotonic interaction with cardiomyocytes is still open. Cell culture experiments have shown previously that heterocellular electrical coupling between cardiomyocytes and fibroblasts (which, based on the present and other studies,18,20 were presumably myofibroblasts) substantially influences impulse propagation. In monolayer cultures of neonatal rat ventricular cardiomyocytes, it was demonstrated that clusters of fibroblasts transfected with the voltage-sensitive potassium channel Kv1.3 induce local conduction blocks.36 Moreover, it was shown that fibroblasts of cardiac origin are capable of relaying electrical activation between strands of cardiomyocytes for distances up to 300 e.37 In contrast to these investigations addressing the effects of nonuniformly distributed fibroblasts on conduction, the present study shows that uniformly distributed myofibroblasts as might be found in diffuse cardiac fibrosis substantially reduce and dV/dtmax as a function of myofibroblast density. Uniform slow conduction (<200 mm/s) was observed at myofibroblast-to-cardiomyocyte ratios >19:100 (> 27:100 when corrected for endogenous myofibroblasts; see below). At these densities, myofibroblasts, which exhibited membrane potentials (eC14 to eC25 mV) falling within the range of potentials reported before by others for cultured cardiac fibroblasts (eC10 to eC20 mV; eC20 to eC40 mV),24,25 depressed the MDP of cardiomyocytes to values
Endogenous Fibroblasts and Conduction
Primary cultures of cardiomyocytes are invariably "contaminated" by noncardiomyocytes among which fibroblasts represent the largest fraction. Because these cells tend to overgrow the cardiomyocytes with time in culture, measures that reduce their number in the initial cell suspension (preplating, density gradient centrifugation) and that inhibit their proliferation (BrdU, cytosine arabinofuranoside, irradiation) are routinely used in the establishment of these cultures.14,16,21 As determined in 3-day-old control strands obtained from preplated cell suspensions, the average density of endogenous myofibroblasts amounted to 10 cells per MA, which corresponds to a myofibroblast-to-cardiomyocyte ratio of 8:100. In contrast, omission of preplating resulted in cell strands containing a substantially higher percentage of endogenous myofibroblasts (22:100), which caused a significant reduction of both and dV/dtmax.
Interestingly, when combining the data of all the three types of preparations and plotting and dV/dtmax as a function of the number of myofibroblasts per MA, the resulting relationship shows a biphasic shape with peak values for either parameter occurring at a myofibroblast-to-cardiomyocyte ratio of 7:100. Based on the finding that myofibroblasts gradually depolarize cardiomyocyte strands in a cell density-dependent manner up to levels of eC50 mV, where they become TTX insensitive, this biphasic relationship can be explained by the well-known phenomenon of supernormal conduction, which accompanies gradual depolarizations of cardiac tissue by potassium.22,23
The finding that supernormal conduction occurred at myofibroblast-to-cardiomyocyte ratios (7:100) similar to the average myofibroblast-to-cardiomyocyte ratio found in uncoated control strands (8:100) has two implications. It explains, why and dV/dtmax declined monotonically in the case of exogenously added myofibroblasts (Figure 4). In these experiments, changes in and dV/dtmax were related to the number of exogenously added myofibroblasts only. If these numbers are corrected for "background myofibroblasts" by adding the average number of endogenous myofibroblasts present in uncoated strands (10), the relationship is shifted to the right and coincides, as expected, with the monotonically descending right limb of the biphasic relationship described above. It also predicts that the dependence of on extracellular potassium concentration ([K+]o) in control strands will equally show a peak at potassium concentrations close to that of the superfusion solution ([K+]o=5.4 mmol/L). This is indeed what has been reported previously for cultured cardiomyocyte strands,27 and it explains in retrospect why levels of [K+]o inducing supernormal conduction in these preparations (5.8 mmol/L) tend to be lower than those observed in intact tissue (up to 10.8 mmol/L)38; ie, endogenous myofibroblasts present in cultured preparations "pre-depolarize" the cardiomyocytes at normal [K+]o sufficiently to elicit supernormal conduction, and any further increase of [K+]o is bound to slow conduction by progressive inactivation of sodium channels.
These findings indicate that heterocellular electrical interactions between myofibroblasts and cardiomyocytes can modify the set point of occurrence of supernormal conduction in respect to [K+]o. Moreover, the results illustrate that myofibroblasts play a key role in the determination of conduction characteristics of cultured myocardial cells that needs to be taken into account when aiming at establishing culture models exhibiting close to in vivo properties in regard to and dV/dtmax.
Study Limitations
Although the findings of this study indicate that impulse propagation characteristics in networks of cardiomyocytes in vitro are strongly affected by the presence of electrically coupled myofibroblasts, the extrapolation of these data to the situation in vivo has to await future characterizations of myofibroblasts in diseased hearts in respect to their size and relative density, their cellular electrophysiology, and their ability to form homocellular and heterocellular gap junctions. Under the assumption that these characteristics are similar to those shown in the present study, the extent to which myofibroblasts will affect conduction in intact diseased hearts will depend on such additional factors as: (1) differences in current density and current composition of adult cardiomyocytes versus cultured neonatal cardiomyocytes, (2) differences in size of cells in situ versus cultured cells, (3) the degree of homocellular and heterocellular gap junctional coupling, (4) the ratio of myofibroblasts to cardiomyocytes within electrotonic reach of each other, and (5) the specific cellular tissue architecture of regions where myofibroblasts intermingle with cardiomyocytes. The combination of all of these factors will ultimately determine to which extent source-to-sink relationships are changed and, thus, conduction is affected in vivo.
Perspectives
Although there exists abundant information regarding the role of myofibroblasts in structural remodeling after a vast array of injuries to the heart, the possibility that this cell type might directly interact with cardiomyocyte electrophysiology via gap junctions has received virtually no attention in the past. In this context, the results of the present study open the perspective that sites of intimate contact between myofibroblasts and cardiomyocytes in diseased hearts (eg, infarct border zones) might contribute to local arrhythmogenic slowing of conduction. Accordingly, controlling the phenotype and proliferation of myofibroblasts may represent a new and specific therapeutic target in the management of cardiac arrhythmias. Indeed, antifibrotic therapies aimed at reversing structural remodeling by influencing metabolic and proliferative activities of myofibroblasts have been shown to be antiarrhythmic, and it remains to be investigated whether, apart from influences on the cellular tissue architecture, a loss of electrotonic interactions between myofibroblasts and cardiomyocytes might also be involved in this antiarrhythmic effect.
Acknowledgments
This study was supported by the Swiss University Conference and by the Swiss National Science Foundation (grant 3100-105916 to S.R.). We thank Regula Fle筩kiger-Labrada for her expert technical assistance and Professor H.P. Clamann for helpful discussions on this manuscript.
References
Adler CP, Ringlage WP, Bhm N. DNS-Gehalt und Zellzahl in Herz und Leber von Kindern. Pathol Res Pract. 1981; 172.
Maisch B. Extracellular matrix and cardiac interstitium: restriction is not a restricted phenomenon. Herz. 1995; 20: 75eC80.
Nag AC. Study of non-muscle cells of the adult mammalian heart: a fine structural analysis and distribution. Cytobios. 1980; 28: 41eC61.
Goldsmith EC, Hoffman A, Morales MO, Potts JD, Price RL, McFadden A, Rice M, Borg TK. Organization of fibroblasts in the heart. Dev Dyn. 2004; 230.
Manabe I, Shindo T, Nagai R. Gene expression in fibroblasts and fibrosis: involvement in cardiac hypertrophy. Circ Res. 2002; 91: 1103eC1113.
Sun Y, Kiami MF, Postlethwaite AE, Weber KT. Infarct scar as living tissue. Basic Res Cardiol. 2002; 97: 343eC347.
Weber KT. Fibrosis in hypertensive heart disease: focus on cardiac fibroblasts. J Hypertens. 2004; 22: 47eC50.
Darby I, Skalli O, Gabbiani G. -smooth muscle actin is transiently expressed by myofibroblasts during experimental wound healing. Lab Invest. 1990; 63: 21eC29.
Gabbiani G, Chaponnier C, Huttner I. Cytoplasmic filaments and gap junctions in epithelial cells and myofibroblasts during wound healing. J Cell Biol. 1978; 76: 561eC648.
Spanakis SG, Petridou S, Masur SK. Functional gap junctions in corneal fibroblasts and myofibroblasts. Invest Ophthalmol Vis Sci. 1998; 39: 1320eC1328.
de Bakker JM, van Capelle FJ, Janse MJ, Tasseron S, Vermeulen JT, de Jonge N, Lahpor JR. Slow conduction in the infarcted human heart. ‘zigzag’ course of activation. Circulation. 1993; 88: 915eC926.
Spach MS, Boineau JP. Microfibrosis produces electrical load variations due to loss of side-to-side cell connections: a major mechanism of structural heart disease arrhythmias. Pacing Clin Electrophysiol. 1997; 20: 397eC413.
Rohr S, Fle筩kiger-Labrada R, Kucera J. Photolithographically defined deposition of attachment factors as a versatile method for patterning the growth of different cell types in culture. Eur J Physiol. 2003; 446: 125eC132.
Blondel B, Roijen I, Cheneval JP. Heart cells in culture: a simple method for increasing the proportion of myoblasts. Experientia. 1971; 27: 356eC358.
Rohr S, Kucera JP. Optical recording system based on a fiber optic image conduit: assessment of microscopic activation patterns in cardiac tissue. Biophys J. 1998; 75: 1062eC1075.
Rohr S, Schlly DM, Kleeber AG. Patterned growth of neonatal rat heart cells in culture. Circ Res. 1991; 68: 114eC130.
Brouty-Boye D, Kolonias D, Savaraj N, Lampidis TJ. -smooth muscle actin expression in cultured cardiac fibroblasts of newborn rat. In Vitro Cell Dev Biol. 1992; 28A: 293eC296.
Wang J, Seth A, McCulloch CAG. Force regulates smooth muscle actin in cardiac fibroblasts. Am J Physiol Heart Circ Physiol. 2000; 279: H2776eCH2785.
Roy S, Khanna S, Bickerstaff AA, Subramanian SV, Atalay M, Bierl M, Pendyala S, Levy D, Sharma N, Venojrvi M, Strauch A, Orosz CG, Sen CK. Oxygen sensing by primary cardiac fibroblasts. Circ Res. 2003; 92: 264eC271.
Wang J, Chen H, Seth A, McCulloch CA. Mechanical force regulation of myofibroblast differentiation in cardiac fibroblasts. Am J Physiol Heart Circ Physiol. 2003; 285: H1871eCH1881.
Lokuta A, Kirby MS, Gaa ST, Lederer J, Rogers TB. On establishing primary cultures of neonatal rat ventricular myocytes for analysis over long periods. J Cardiovasc Electrophysiol. 1994; 5: 50eC61.
Kagiyama Y, Hill JL, Gettes LS. Interaction of acidosis and increased extracellular potassium on action potential characteristics and conduction in guinea pig ventricular muscle. Circ Res. 1982; 51: 614eC623.
Shaw RM, Rudy Y. Electrophysiologic effects of acute myocardial ischemia: a mechanistic investigation of action potential conduction and conduction failure. Circ Res. 1997; 80: 124eC138.
Hyde AB, BBlondel, A Matter JP. Homo- and heterocellular junctions in cell cultures: an electrophysiological and morphological study. Prog Brain Res. 1969; 31: 283eC311.
Rook MB, van Ginneken AC, de Jonge B, el Aoumari A, Gros D, Jongsma HJ. Differences in gap junction channels between cardiac myocytes, fibroblasts, and heterologous pairs. Am J Physiol Cell Physiol. 1992; 263: C959eCC977.
Chilton L, Ohya S, Freed D, George E, Drobic V, Shibukawa Y, MacCannell KA, Imaizumi Y, Clark RB, Dixon IMC, Giles WR. K+ currents regulate the resting membrane potential, proliferation, and contractile responses in ventricular fibroblasts and myofibroblasts. Am J Physiol Heart Circ Physiol. 2005; 288: H2931eCH2939.
Rohr S, Kucera JP, Kleber AG. Slow conduction in cardiac tissue. I. Effects of a reduction of excitability vs. a reduction of electrical coupling on microconduction. Circ Res. 1998; 83: 781eC794.
Leslie KO, Taatjes DJ, Schwarz J, von Turkovich M, Low RB. Cardiac myofibroblasts express alpha smooth muscle actin during right ventricular pressure overload in the rabbit. Am J Pathol. 1991; 139: 207eC216.
Weber K. Cardiac interstitium. In: Poole-Wilson P, Colucci WBM, et al, eds. Heart Failure. New York, NY: Churchill Livingstone; 1997: 13eC31.
Hinz B, Mastrangelo D, Iselin CE, Chaponnier C, Gabbiani G. Mechanical tension controls granulation tissue contractile activity and myofibroblast differentiation. Am J Pathol. 2001; 159: 1009eC1020.
De Maziere AMGL, Van Ginneken ACG, Wilders R, Jongsma HJ, Bouman LN. Spatial and functional relationship between myocytes and fibroblasts in the rabbit sinoatrial node. J Mol Cell Cardiol. 1992; 24: 567eC578.
Camelliti P, Green CR, LeGrice I, Kohl P. Fibroblast network in rabbit sinoatrial node: structural and functional identification of homogeneous and heterogeneous cell coupling. Circ Res. 2004; 94: 828eC835.
Camelliti P, Devlin GP, Matthews KG, Kohl P, Green CR. Spatially and temporally distinct expression of fibroblast connexins after sheep ventricular infarction. Cardiovasc Res. 2004; 62: 415eC425.
Powell DW, Mifflin RC, Valentich JD, Crowe SE, Saada JI, West AB. Myofibroblasts. I. Paracrine cells important in health and disease. Am J Physiol Cell Physiol. 1999; 277: C1eCC19.
Cameron Va, Rademaker MT, Ellmers LJ, Expiner EA, Nicholls MG, Richards M. Atrial (ANP) and brain natriuretic peptide (BNP) expression after myocardial infarction in sheep: ANP is synthesized by fibroblasts infiltrating the infarct. Endocrinology. 2000; 141: 4690eC4697.
Feld Y, Melamed-Frank M, Kehat I, Tal D, Marom S, Gepstein L. Electrophysiological modulation of cardiomyocytic tissue by transfected fibroblasts expressing potassium channels: a novel strategy to manipulate excitability. Circ. 2002; 105: 522eC529.
Gaudesius G, Miragoli M, Thomas SP, Rohr S. Coupling of cardiac electrical activity over extended distances by fibroblasts of cardiac origin. Circ Res. 2003; 93: 421eC428.
Hisatome I, Arita M. Effects of catecholamines on the residual sodium channel dependent slow conduction in guinea pig ventricular muscles under normoxia and hypoxia. Cardiovasc Res. 1995; 29: 65eC73.(Michele Miragoli, Giedriu)