- Email: [email protected]
Myofibroblasts in diseased hearts: New players in cardiac arrhythmias?
CELL TO BEDSIDE
Myofibroblasts in diseased hearts: New players in cardiac arrhythmias? Stephan Rohr, MD From the Department of Physiology, University of Bern, Bern, Switzerland. Cardiac pathologies leading to the development of organ fibrosis typically are associated with the appearance of interstitial myofibroblasts. This cell type plays a central role in excessive extracellular matrix deposition, thereby contributing to arrhythmogenic slow and discontinuous conduction by causing disorganization of the three-dimensional network of electrically coupled cardiomyocytes. Besides this involvement in structural remodeling, myofibroblasts recently have been discovered in-vitro to promote arrhythmogenesis by direct modification of cardiomyocyte electrophysiology following establishment of heterocellular electrical coupling. In particular, myofibroblasts were found to rescue impulse conduction between disjoined cardiac tissues by acting as passive electrical conduits for excitatory current flow. Although, in principle, such recovery of blocked conduction might be beneficial, propagation across myofibroblast conduits is substantially delayed, thereby promoting arrhythmogenic slow and discontinuous conduction. Second, moderately polarized myofibroblasts were found to induce cell density– dependent depolarization of cardiomyocytes, which causes arrhythmogenic slow conduction due to
the reduction of fast inward currents. Finally, critical depolarization of cardiomyocytes by myofibroblasts was discovered to lead to the appearance of ectopic activity in a model of the infarct border zone. These findings obtained in vitro suggest that electrotonic interactions following gap junctional coupling between myofibroblasts and cardiomyocytes in structurally remodeled fibrotic hearts might directly initiate the main mechanisms underlying arrhythmogenesis, that is, abnormal automaticity and abnormal impulse conduction. If, in the future, similar arrhythmogenic mechanisms can be shown to be operational in intact hearts, myofibroblasts might emerge as a novel noncardiomyocyte target for antiarrhythmic therapy.
Introduction
sembly of cardiomyocytes, fibroblasts contribute importantly to the uniformity of the excitable substrate and, thus, to continuous and fast electrical activation of the working myocardium under physiologic conditions. Given that the fibrillar components of the scaffold are subject to constant turnover, which amounts up to approximately 5% per day,2 it is evident that formation and degradation of these fibers needs to be closely controlled in order to maintain the structural and functional integrity of the myocardium over time. This control is exerted by a large number of biophysical and molecular signaling events that control fibroblast growth and function. If the delicate balance between extracellular matrix (ECM) production and degradation is lost, the working myocardium undergoes structural remodeling that has far-reaching adverse consequences for both the electrical and the pump function of the heart.
Although a wealth of information exists regarding the role of abnormal cardiomyocyte electrophysiology in arrhythmogenesis, much less is known about whether stromal cells of the heart in general and cardiac fibroblasts in particular might be actively involved in arrhythmogenesis. This is surprising insofar as that fibroblasts represent the most numerous cell population in normal human hearts, with individual cells being in intimate contact with cardiomyocytes. In fact, cardiac fibroblasts outnumber cardiomyocytes by a factor of 2–3 while occupying approximately 20% of the volume of the working myocardium.1 Under physiologic conditions, fibroblasts produce and maintain a three-dimensional network of collagen and elastin fibers, which acts as a scaffold for cardiomyocytes and integrates the mechanical forces of individual cells, thus resulting in an efficient pump function of the entire organ. Moreover, by providing the structural backbone for the regular three-dimensional asThis work was supported by the Swiss National Science Foundation (Grant 320000-118247/1). Address reprint requests and correspondence: Dr. Stephan Rohr, Department of Physiology, University of Bern, Buehlplatz 5, CH-3012 Bern, Switzerland. E-mail address: [email protected]. ch. (Received February 12, 2009; accepted February 22, 2009.)
KEYWORDS Myofibroblast; Fibroblast; Arrhythmia; Ectopic activity; Slow conduction; Structural remodeling; Fibrosis; Infarction; Connexin; Cell therapy (Heart Rhythm 2009;6:848 – 856) © 2009 Heart Rhythm Society. All rights reserved.
Structural remodeling and arrhythmogenesis Structural remodeling of the myocardium is the result of a variety of complex cellular reactions to injury and involves both cardiomyocytes and noncardiomyocytes. Histologically, it is characterized by cardiomyocyte hypertrophy, activation and proliferation of fibroblasts, increased ECM
1547-5271/$ -see front matter © 2009 Heart Rhythm Society. All rights reserved.
doi:10.1016/j.hrthm.2009.02.038
Rohr
Myofibroblasts and Cardiac Arrhythmogenesis
849
Figure 1 Etiology and functional consequences of cardiac fibrosis. Microscopic images illustrate connective tissue distribution (collagen: red) in normal myocardium (left) and in a patient with extensive left ventricular fibrosis (right) that results in isolation of bundles of cardiomyocytes. Bar ⫽ 100 m. (Images courtesy Dr. T. Schaffner, Department of Pathology, University of Bern, Bern, Switzerland.)
deposition, and cell death.3 The leading causes of structural remodeling are pressure overload, volume overload, ischemic heart disease, and genetics and old age, which result in changes of the size and shape of the heart and the appearance of interstitial and/or replacement fibrosis (Figure 1). Functionally, structural remodeling is associated with compromised diastolic and/or systolic pump function and an increased likelihood of occurrence of life-threatening arrhythmias.4 Arrhythmias in structurally remodeled myocardia are based on adverse changes in cardiomyocyte electrophysiology (‘electrical remodeling’) and/or on changes in the extent and organization of the ECM (“fibrotic remodeling”). Electrical remodeling relates to arrhythmogenic alterations in the function and expression of ion channels, exchangers, calcium handling proteins, and gap junctions occurring over extended periods of time in response to pathologic stimuli. Fibrotic remodeling, on the other hand, predisposes the heart to arrhythmias due to the presence of excessive amounts of ECM, which is organized into collagenous septa that separate bundles of cardiomyocytes over substantial distances. The resulting reduction of lateral gap junctional coupling results in functional discontinuities within the excitable substrate that disrupts the orderly spread of electrical activation and results in nonuniform conduction5 or zigzag activation of the myocardium.6 Common to both of these abnormal activation patterns is a slowing of macroscopic
conduction velocity and an increased likelihood of occurrence of unidirectional conduction blocks that are central to the initiation of reentrant activations. Arrhythmogenic mechanisms associated with both electrical remodeling and fibrotic remodeling are cardiomyocyte centered in the sense that they concern either pathologic changes in cardiomyocyte electrophysiology or that they relate to the disruption of the orderly three-dimensional network of electrotonically coupled cardiomyocyte by fibrosis. But what about a direct involvement of connective tissue cells that goes beyond the mere hypersecretion of ECM?
Cardiac fibrosis and the appearance of myofibroblasts The absence of arrhythmias in healthy hearts suggests that fibroblasts, even though they greatly outnumber cardiomyocytes, exert no arrhythmogenic effects per se under physiologic conditions. However, under pathologic conditions such as hypertensive heart disease and infarction, an additional type of cell makes its appearance in the working myocardium. These so-called myofibroblasts (sometimes termed activated fibroblasts) are considered to be importantly involved in the establishment of reactive and reparative fibrosis.7,8 Myofibroblasts were first described in the 1970s as fibroblastic cells located within granulation tissue of skin wounds where they initiate contraction of the granulation tissue and disappear thereafter by programmed cell
850
Heart Rhythm, Vol 6, No 6, June 2009 the initial scar formation and the continuous rebuilding of the scar thereafter.16 Although the exact origin of myofibroblasts is still under debate (circulating or resident adventitial or interstitial fibroblasts undergoing a phenotype switch toward myofibroblasts, circulating bone marrow– derived progenitor cells),16 –18 ample evidence indicates that their appearance is triggered by local inflammatory reactions, mechanical stress, and humoral factors.17,19 –21 Induction of myofibroblasts is triggered in particular by transforming growth factor-1 (TGF-1), but other factors such as basic fibroblast growth, angiotensin II, catecholamines, and insulin-like growth factor also are involved in the determination of phenotype and function of cardiac myo/fibroblasts.21
Gap junctional coupling of myofibroblasts to cardiomyocytes
Figure 2 Section of a 4-day-old infarct in the left ventricular free wall of an adult rat. A: Phase contrast image with the infarct outlined by the red dashed line. B: Same region stained for nuclei (DAPI, blue) and ␣-SMA (green). The filamentous green staining corresponds to stress fibers of numerous myofibroblasts within the infarct area. In addition, myofibroblasts can be seen in the immediate vicinity of the infarct (yellow arrows), where they are intercalated between surviving border zone cardiomyocytes.
death.9 In contrast to fibroblasts, myofibroblasts express ␣-smooth muscle actin (␣-SMA), which serves as a routine marker for this cell type. In recent years, myofibroblasts have been found to play a central role in hepatic, pulmonary, pancreatic, and renal fibrosis in which they seem to be the primary source of excessive ECM production and participate in the progression of disease toward organ failure.10 In the heart, myofibroblasts are not normally present with the exception of valve leaflets, where ␣–SMA–positive interstitial cells are found under physiologic conditions.11 Myofibroblasts are, however, constitutive cellular elements of many of the cardiac pathologies mentioned. For example, animal models of pressure overload have shown that early inflammatory changes are followed by myofibroblast proliferation within days and subsequent development of perivascular and interstitial fibrosis.12,13 Similarly and as shown in Figure 2, myofibroblasts appear in infarcts in very large numbers a few days after the onset of myocardial infarction and, due to incomplete apoptosis during the healing process, persist locally for many years.14,15 In infarcts, myofibroblasts are recognized for their central role in both
Given the close apposition between cardiomyocytes and myofibroblasts in diseased myocardia, the question arises as to whether gap junctions might electrically interconnect the two cell types. Whereas connexin (Cx) expression is a well-established feature of myofibroblasts in tissues different from heart, such as suburothelial tissue,22 intestine,23 breast cancer stroma,24 and healing skin wounds,25 the question of gap junctional coupling between myofibroblasts and cardiomyocytes in diseased myocardia is still open. Interestingly, however, fibroblasts residing in infarct scars have shown abundant expression of Cx43 and Cx45.26 Given that other investigators have shown that “fibroblasts” in infarct scars are, in fact, myofibroblasts (cf. also Figure 2),15,27 it is tempting to speculate that the connexin-expressing “fibroblasts” in the former study actually were connexin-expressing myofibroblasts residing in the infarct region. In contrast to the lack of data characterizing formation of gap junctions between myofibroblasts and cardiomyocytes in intact myocardia, heterocellular electrotonic coupling is a long-known feature of cultured cardiomyocytes and fibroblasts. As early as 1969, Hyde et al28 observed that fibroblasts inserted between cardiomyocytes synchronized spontaneous contractions and, using sharp electrode impalements, showed that cardiomyocytes in contact with what they considered to be fibroblasts exhibit reduced membrane potentials. In the meantime, fibroblasts in this model system were found to display a myofibroblast phenotype and to express Cx43 and Cx45 (but not Cx40) among themselves and at contact sites with cardiomyocytes (Figure 3).29 Thus, at least in vitro, myofibroblasts are capable of establishing gap junctional coupling both among themselves and with cardiomyocytes, which has, as outlined in more detail later, profound arrhythmogenic consequences.
Myofibroblast mend broken pathways of impulse conduction It is generally assumed that cardiac impulse conduction is blocked at sites where the cardiomyocyte network is disrupted by collagenous septa or at sites of sutures following heart transplantation. Interestingly, for the latter case, it has
Rohr
Myofibroblasts and Cardiac Arrhythmogenesis
851
Figure 3 Homocellular and heterocellular gap junctional coupling of cultured cardiac myofibroblasts. Top row: Immunofluorescent labeling for Cx43 (green) and ␣-SMA (red) shows expression of Cx43 both at myofibroblast–myofibroblast borders (yellow arrows) and at myofibroblast– cardiomyocyte contact sites (white arrows). The corresponding phase contrast image shows the spatial arrangement of myofibroblasts (MFB) and cardiomyocytes (CM). Bottom row: Same as top row but for Cx45, which is similarly expressed at both homocellular and heterocellular cell junctions. (Modified with permission from Miragoli M, Gaudesius G, Rohr S. Electrotonic modulation of cardiac impulse conduction by myofibroblasts. Circ Res 2006;98:801– 810.)
been reported that recipient and donor atria are sporadically capable of establishing electrical synchronization despite scar formation in the region of the suture.30 Because myofibroblasts are a typical cellular component of scar tissue, we tested the hypothesis that these cells are capable of relaying electrical activity between disjoined regions of cardiac tissue by acting as a passive electrical conduit for excitatory current flow.31 For this purpose, we developed a photolithographic method that permits the construction of a geometrically defined in vitro cell model of a transplantation scar.32 The model consisted of two strands of neonatal rat ventricular cardiomyocytes interconnected by a region containing myofibroblasts of cardiac origin only (Figure 4A). The preparations were stimulated with an extracellular electrode placed several length constants away from the measurement site, and impulse propagation across the myofibroblast insert was followed optically using voltage-sensitive dyes.33 Similar to intracellular electrode recordings, these dyes indicate changes in transmembrane potential in real time and therefore permit characterization of electrical activation patterns with high spatiotemporal resolution. As illustrated by the panel depicting action potential upstrokes along the preparation, local stimulation was followed by activation of the entire preparation, indicating successful propagation across the myofibroblast insert. However, as shown by the plot of local activation times versus distance, successful bridging of electrical activation by myofibroblasts came at the price of a substantial local activation delay of approximately 30 ms. As shown in Figure 4B,
conduction delays across myofibroblast inserts were a function of insert width and rose monotonically up to approximately 60 to 70 ms for the largest inserts still supporting synchronization of electrical activity. At widths greater than 300 m, propagation invariably failed. These experiments demonstrate that gap junctional myofibroblast-to-cardiomyocyte coupling can restore conduction across discontinuities in networks of cardiomyocytes. By acting as passive electrical conduits for excitatory current flow, myofibroblasts might contribute to the explanation of electrical synchronization of recipient and donor tissue following heart transplantation. With regard to arrhythmogenesis, “mending of broken pathways” might have both proarrhythmic and antiarrhythmic consequences depending on the setting. Whereas elimination of conduction blocks per se can be expected to counteract arrhythmogenesis, the substantial conduction delays incurred at sites where myofibroblasts act as conduits for excitatory current flow predict the emergence of discontinuous slow conduction. In fact, in the example shown in Figure 4, apparent local conduction velocities within the central 200 m encompassing the myofibroblast insert amounted to approximately 5 mm/s, which corresponds to an approximately 100-fold slowing of conduction compared with cardiomyocyte only strands (⬃400 mm/s). If such instances of local propagation delays were to be concatenated in settings such as interstitial fibrosis, macroscopic impulse conduction would become very slow and the risks for arrhythmogenesis would increase accordingly.
852
Heart Rhythm, Vol 6, No 6, June 2009 blasts (Figure 5A). Impulse conduction characteristics along these preparations were assessed optically and, following the experiments, the preparations were stained for ␣-SMA in order to determine local myofibroblast densities. Compared with uncoated cardiomyocyte preparations, conduction was considerably slowed in the presence of myofibroblasts. At the same time, maximal upstroke velocities of the propagated action potentials were reduced. This suggested that myofibroblasts, which exhibit a low intrinsic membrane potential of approximately – 40 to –50 mV, induced partial depolarization of cardiomyocytes following heterocellular gap junctional coupling and, hence, caused a reduction of the sodium inward current. Intracellular measurements using sharp electrodes confirmed this hypothesis (Figure 5B). Maximal diastolic potentials measured in cardiomyocytes decreased from approximately – 80 mV at myofibroblast densities less than 5% to approximately –55 mV at myofi-
Figure 4 Myofibroblasts serve as passive electrical conduits for impulse conduction. A: Experimental preparations consisted of two strands of cardiomyocytes that were joined by a myofibroblast insert. Electrical activity was recorded optically from the sites marked by circles in the phase contrast image of the preparation (orange: cardiomyocyte region; green: myofibroblast region). Following stimulation, action potential upstrokes could be recorded along the entire preparations, indicating that myofibroblasts successfully synchronized electrical activity between the two cardiomyocyte strands. The plot of local activation times versus distance indicates that conduction across the 134-m-wide myofibroblast insert was delayed by approximately 30 ms. B: Propagation delays across myofibroblast inserts rose monotonically with insert width. At widths greater than 300 m, synchronization of electrical activity invariably failed. (Modified with permission from Gaudesius G, Miragoli M, Thomas SP, et al. Coupling of cardiac electrical activity over extended distances by fibroblasts of cardiac origin. Circ Res 2003;93:421– 428.)
Myofibroblasts induce slow conduction Structurally, the border zone of healing infarcts is characterized by surviving cardiomyocytes in close contact with a large number of myofibroblasts (cf. Figure 2). Functionally, this very same region is known to contribute significantly to postinfarction arrhythmogenesis based on remodeling of the cellular microarchitecture, modifications of gap junctional coupling among cardiomyocytes, and changes in the electrophysiology of border zone cardiomyocytes (for review see Peters and Wit34). Given that myofibroblasts communicate via gap junctions with cardiomyocytes, we investigated whether this type of interaction might have additional arrhythmogenic consequences. For this purpose, we generated a simplified model of the infarct border zone, which consisted of strands of cardiomyocytes coated with myofibro-
Figure 5 Myofibroblasts induce slow conduction. A: Schematic diagram of the experimental preparation consisting of cardiomyocyte strands coated with myofibroblasts. Impulse propagation along the preparation following stimulation was measured optically at the sites indicated by circles in the phase contrast picture. Postexperimental staining for ␣-SMA (green overlay) served to determine the percentage area of the strands covered with myofibroblasts. As indicated by the plot of local activation times versus distance, conduction along the hybrid cell preparations was continuous but slow. B: Myofibroblasts reduce resting membrane potential of cardiomyocytes in a cell density– dependent manner. C: Maximal action potential upstroke velocities and conduction velocities along the preparations show a biphasic dependence on myofibroblast density. (Modified with permission from Miragoli M, Gaudesius G, Rohr S. Electrotonic modulation of cardiac impulse conduction by myofibroblasts. Circ Res 2006;98:801– 810.)
Rohr
Myofibroblasts and Cardiac Arrhythmogenesis
broblast densities greater than 40%. This graded depolarization was paralleled by a biphasic change in conduction velocities that initially slightly increased from approximately 390 mm/s to 440 mm/s before, at myofibroblast densities greater than 10%, monotonically falling to 260 mm/s (Figure 5C). This biphasic dependence is, in fact, the expected result for a gradual depolarization of cardiomyocytes by myofibroblasts, as it replays exactly what is known for many decades for intact cardiac tissue subjected to gradual depolarization by potassium: also there, conduction velocity transiently increases (“supernormal conduction”)
853 before increasing sodium channel inactivation causes progressive conduction slowing. These data suggest that heterocellular electrotonic interactions in the border zone of healing infarcts or along laminae of interstitial fibrotic tissue might cause arrhythmogenic slow conduction based on partial depolarization of cardiomyocytes by myofibroblasts. Under the assumption that myofibroblasts in diseased hearts exert similar effects on cardiomyocytes as those found in vitro, the degree of conduction slowing in the infarct border zone likely depends on the relative densities and the three-dimensional
Figure 6 Myofibroblasts induce abnormal automaticity. A, top: Schematic diagram of the layout of the experimental preparations consisting of cardiomyocyte strands coated with myofibroblasts. Middle left: Overview of a preparation consisting of 24 identical strands (0.6 ⫻ 4.5 mm each) that were scanned simultaneously for the presence of spontaneous electrical activity with a high-speed camera following staining with voltage sensitive dyes. Snapshot shows two strands exhibiting spontaneous activity (green: area depolarized by propagating action potential; pink star: origin of abnormal automaticity). Middle right: Details of the cellular microarchitecture (phase contrast image) and the density and distribution of myofibroblasts (vimentin staining) of a selected strand. Bottom: Continuous optical recording of action potentials at the sites denoted with a blue and red circle in the overview shows regular spontaneous electrical activity. B: Incidence of abnormal automaticity as a function of myofibroblast density. (Modified with permission from Miragoli M, Salvarani N, Rohr S. Myofibroblasts induce ectopic activity in cardiac tissue. Circ Res 2007;101:755–758.)
854
Heart Rhythm, Vol 6, No 6, June 2009
coupling pattern of the two cell populations. Within the epicardial/endocardial border zone consisting of a thin and essentially two-dimensional sheet of surviving cardiomyocytes, the large number of myofibroblasts beneath might slow conduction considerably. In contrast, within the myocardial wall, the depolarizing effect of myofibroblasts likely is offset by the dense three-dimensional network of wellcoupled cardiomyocytes bordering onto the infarct area. The combination of the two then would result in heterogeneous conduction properties of the excitable substrate around the infarct, which might additionally contribute to the established arrhythmogenic nature of the infarct border zone.
Myofibroblasts induce abnormal automaticity In diseased hearts, cardiac fibrosis not only induces arrhythmogenic slow and discontinuous conduction; it also is suspected to promote ectopic activity, which is central to the generation of focal and reentrant tachyarrhythmias.35 The presence of ectopic activity under these conditions is thought to be favored by anisotropies in tissue structure and in electrical coupling that ultimately permits a few cardiomyocytes exhibiting abnormal automaticity to drive the surrounding myocardium.36 Given that myofibroblasts reduce the resting membrane potentials of cardiomyocytes
Figure 7
following heterocellular electrical coupling, we hypothesized that ectopic activity in fibrotic myocardium alternatively might be based on the concept of depolarizationinduced automaticity.37 The model used to investigate this hypothesis was designed similar to the infarct border zone model and consisted of strands of cardiomyocytes coated with myofibroblasts (Figure 6A). Following staining with voltage-sensitive dyes, spontaneous activity was recorded for repeated periods of 4 seconds with a macroscope that permitted simultaneous monitoring of 24 individual strands. The overview of the preparations in Figure 6A corresponds to a single frame taken from such a recording. It illustrates the general layout of the patterned growth preparations and shows two strands undergoing spontaneous activation. As demonstrated by the continuous recordings in one strand, ectopic activity was regular with action potentials occurring at a frequency of approximately 1 Hz. As shown in Figure 6B, occurrence of spontaneous activity was steeply dependent on myofibroblast density. Whereas preparations were completely quiescent at myofibroblast densities less than 16%, more than 80% exhibited spontaneous activity at densities greater than 40%. Spontaneous activity was the result of depolarization-induced automaticity as demonstrated by the finding that hyperpolarization of the preparations with
Summary of mechanisms by which myofibroblasts might contribute to arrhythmogenesis. ECM ⫽ extracellular matrix.
Rohr
Myofibroblasts and Cardiac Arrhythmogenesis
KATP channel openers led to immediate cessation of abnormal automaticity.38 Moreover, graded reductions of resting membrane potentials in single cardiomyocytes using patch clamp techniques started to induce spontaneous activity at levels of membrane depolarization (approximately – 67 mV) highly similar to those present at minimal myofibroblast densities necessary to induce ectopic activity (approximately – 66 mV). These findings suggest that myofibroblasts in intimate contact with cardiomyocytes at sites such as the infarct border zone or along muscle bundles ensheathed by connective tissue might elicit focal ectopic activity. Mechanistically, this situation closely resembles the concept of arrhythmogenic “injury current” flow between damaged and healthy cardiomyocytes in acute ischemia except that, in the present case, myofibroblasts rather than injured cardiomyocytes are the source of depolarizing current. The chances of this novel arrhythmogenic mechanism becoming operational in the fibrotically remodeled hearts in situ will likely depend on a number of different parameters, such as the membrane potentials of myofibroblasts and cardiomyocytes in situ, the respective cell sizes, the degree of homocellular and heterocellular coupling among the two cell types, and the specifics of local cellular tissue architecture affecting source-to-load matching.
Conclusions and perspectives The findings presented demonstrate that myofibroblasts, as schematically summarized in Figure 7, are capable of exerting highly arrhythmogenic effects following heterocellular electrotonic coupling to cardiomyocytes. By inducing (dis)continuous slow conduction as well as ectopic activity, myofibroblasts are, in fact, capable of launching the main mechanisms driving focal and reentrant tachyarrhythmias, that is, abnormal impulse conduction and abnormal impulse formation.39 Although this finding opens the perspective that this particular cell type might constitute a new active player in arrhythmogenesis, extrapolation of the in vitro findings presented to intact diseased myocardia has to await demonstrations of analogous structural and functional myofibroblast-to-cardiomyocyte interactions in situ. If such interactions were to be found, novel approaches to antiarrhythmic treatments might emerge that could be targeted specifically at myofibroblasts and not, as in the past, at cardiomyocytes. The advantages of such myofibroblast-centered antiarrhythmic treatments would be manifold, with two major aspects: (1) relief of arrhythmias irrespective of the underlying cause (mechanical overload, ischemic heart disease, etc.) and (2) treating myofibroblasts instead of modifying cardiomyocyte electrophysiology could circumvent the potentially dangerous complexities associated with the latter approach.40 Among a substantial number of options, therapies affecting myofibroblast appearance and function could include treatments as diverse as the use of statins, inhibition of the renin-angiotensin-aldosterone system, and interference with cytokines known to modulate the phenotype of this cell. All of these approaches have been shown to affect myofibro-
855 blast biology in vitro and to counteract arrhythmogenesis in vivo. Whether these antiarrhythmic effects were solely due to a reduction of fibrosis, were based on modifications of the electrotonic cross-talk between cardiomyocytes and myofibroblasts, or occurred secondary to a change in paracrine activity of the myofibroblasts remains to be determined. Apart from their possible relevance for arrhythmogenesis in fibrotically remodeled hearts, the results presented also might have relevance for cell therapies using communication competent cell types that, compared with cardiomyocytes, have a reduced intrinsic membrane potential.41 In analogy to myofibroblasts, such cells might critically depolarize neighboring cardiomyocytes after establishment of heterocellular gap junctional coupling, thus contributing to arrhythmogenesis by induction of slow conduction and precipitation of abnormal automaticity. Further studies are needed to determine whether this mechanism contributes to arrhythmias following cell therapy and whether special attention must be given to the degree of membrane polarization of the cells transplanted.
Acknowledgment I wish to thank Dr. Christian Mühlfeld for providing sections of infarcted rat hearts.
References 1. Maisch B. Extracellular matrix and cardiac interstitium: restriction is not a restricted phenomenon. Herz 1995;20:75– 80. 2. McAnulty RJ, Laurent GJ. Collagen synthesis and degradation in vivo. Evidence for rapid rates of collagen turnover with extensive degradation of newly synthesized collagen in tissues of the adult rat. Coll Relat Res 1987;7:93–104. 3. Swynghedauw B. Molecular mechanisms of myocardial remodeling. Physiol Rev 1999;79:215–262. 4. Adamson PB, Barr RC, Callans DJ, et al. The perplexing complexity of cardiac arrhythmias: beyond electrical remodeling. Heart Rhythm 2005;2:650 – 659. 5. 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;20:397– 413. 6. De Bakker JMT, Van Capelle FJL, Janse MJ, et al. Slow conduction in the infarcted human heart: zigzag course of activation. Circulation 1993;88:915– 926. 7. Weber KT. Extracellular matrix remodeling in heart failure: a role for de novo angiotensin II generation. Circulation 1997;96:4065– 4082. 8. Weber KT. Fibrosis in hypertensive heart disease: focus on cardiac fibroblasts. J Hypertension 2004;22:47–50. 9. Gabbiani G. The myofibroblast in wound healing and fibrocontractive diseases. J Pathol 2003;200:500 –503. 10. Kisseleva T, Brenner DA. Fibrogenesis of parenchymal organs. Proc Am Thorac Soc 2008;5:338 –342. 11. Aikawa E, Whittaker P, Farber M, et al. Human semilunar cardiac valve remodeling by activated cells from fetus to adult: implications for postnatal adaptation, pathology, and tissue engineering. Circulation 2006;113:1344 – 1352. 12. Leslie KO, Taatjes DJ, Schwarz J, et al. Cardiac myofibroblasts express alpha smooth muscle actin during right ventricular pressure overload in the rabbit. Am J Pathol 1991;139:207–216. 13. Kuwahara F, Kai H, Tokuda K, et al. Transforming growth factor function blocking prevents myocardial fibrosis and diastolic dysfunction in pressureoverloaded rats. Circ 2002;106:130 –135. 14. Peterson DJ, Ju H, Panagia M, et al. Expression of Gi-2␣ and Gs␣ in myofibroblasts localized to the infarct scar in heart failure due to myocardial infarction. Cardiovasc Res 1999;41:575–585. 15. Willems IE, Havenith MG, De Mey JG, et al. The alpha-smooth muscle actinpositive cells in healing human myocardial scars. Am J Pathol 1994;145:868 – 875. 16. Sun Y, Kiani MF, Postlethwaite AE, et al. Infarct scar as living tissue. Basic Res Cardiol 2002;97:343–347.
856 17. Desmouliere A, Chapponier C, Gabbiani G. Tissue repair, contraction, and the myofibroblast. Wound Rep Reg 2005;13:7–12. 18. van Amerongen M, Bou-Gharios G, Popa E, et al. Bone marrow-derived myofibroblasts contribute functionally to scar formation after myocardial infarction. J Pathol 2008;214:377–386. 19. Weber KT. From inflammation to fibrosis: a stiff stretch of highway. Hypertension 2004;43:716 –719. 20. Kuwahara F, Kai H, Tokuda K, et al. Hypertensive myocardial fibrosis and diastolic dysfunction. Hypertension 2004;43:739 –745. 21. Manabe I, Shindo T, Nagai R. Gene expression in fibroblasts and fibrosis. Circ Res 2002;91:1103–1113. 22. Sui GP, Rothery S, Dupont E, et al. Gap junctions and connexin expression in human suburothelial interstitial cells. BJU Int 2002;90:118 –129. 23. Powell DW, Mifflin RC, Valentich JD, et al. Myofibroblasts. II. Intestinal subepithelial myofibroblasts. Am J Physiol 1999;277:C183–C201. 24. Jamieson S, Going JJ, D’Arcy R, et al. Expression of gap junction proteins connexin 26 and connexin 43 in normal human breast and in breast tumours. J Pathol 1998;184:37– 43. 25. Gabbiani G, Chaponnier C, Huttner I. Cytoplasmic filaments and gap junctions in epithelial cells and myofibroblasts during wound healing. J Cell Biol 1978; 76:561–568. 26. Camelliti P, Devlin GP, Matthews KG, et al. Spatially and temporally distinct expression of fibroblast connexins after sheep ventricular infarction. Cardiovasc Res 2004;62:415. 27. Sun Y, Weber KT. Angiotensin converting enzyme and myofibroblasts during tissue repair in the rat heart. J Mol Cell Cardiol 1996;28:851– 858. 28. Hyde AB, Blondel B, Matter A, et al. Homo- and heterocellular junctions in cell cultures: an electrophysiological and morphological study. Prog Brain Res 1969;31:283–311. 29. Miragoli M, Gaudesius G, Rohr S. Electrotonic modulation of cardiac impulse conduction by myofibroblasts. Circ Res 2006;98:801– 810. 30. Lefroy DC, Fang JC, Stevenson LW, et al. Recipient-to-donor atrioatrial con-
Heart Rhythm, Vol 6, No 6, June 2009
31.
32.
33.
34. 35.
36.
37. 38. 39.
40.
41.
duction after orthotopic heart transplantation: surface electrocardiographic features and estimated prevalence. Am J Cardiol 1998;82:444 – 450. Gaudesius G, Miragoli M, Thomas SP, et al. Coupling of cardiac electrical activity over extended distances by fibroblasts of cardiac origin. Circ Res 2003; 93:421– 428. Rohr S, Flückiger-Labrada R, Kucera JP. Photolithographically defined deposition of attachment factors as a versatile method for patterning the growth of different cell types in culture. Eur J Physiol 2003;446:125–132. 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;75:1062–1075. Peters NS, Wit AL. Myocardial architecture and ventricular arrhythmogenesis. Circulation 1998;97:1746 –1754. De Bakker JMT, Stein M, van Rijen HVM. Three-dimensional anatomic structure as substrate for ventricular tachycardia/ventricular fibrillation. Heart Rhythm 2005;2:777–779. Wilders R, Wagner MB, Golod GA, et al. Effects of anisotropy on the development of cardiac arrhythmias associated with focal activity. Eur J Physiol 2000;441:301–312. Mohabir R, Ferrier GR. Effects of ischemic conditions and reperfusion on depolarization-induced automaticity. Am J Physiol 1988;255:H992–H999. Miragoli M, Salvarani N, Rohr S. Myofibroblasts induce ectopic activity in cardiac tissue. Circ Res 2007;101:755–758. Zlochiver S, MuÒoz V, Vikstrom KL, et al. Electrotonic myofibroblast-tomyocyte coupling increases propensity to reentrant arrhythmias in two-dimensional cardiac monolayers. Biophys J 2008;95:4469 – 4480. Echt DS, Liebson PR, Mitchell LB, et al. Mortality and morbidity in patients receiving encainide, flecainide, or placebo. The cardiac arrhythmia suppression trial. N Engl J Med 1991;324:781–788. Heubach JF, Graf EM, Leutheuser J, et al. Electrophysiological properties of human mesenchymal stem cells. J Physiol 2004;554:659 – 672.
Myofibroblasts in diseased hearts: New players in cardiac arrhythmias? Stephan Rohr, MD From the Department of Physiology, University of Bern, Bern, Switzerland. Cardiac pathologies leading to the development of organ fibrosis typically are associated with the appearance of interstitial myofibroblasts. This cell type plays a central role in excessive extracellular matrix deposition, thereby contributing to arrhythmogenic slow and discontinuous conduction by causing disorganization of the three-dimensional network of electrically coupled cardiomyocytes. Besides this involvement in structural remodeling, myofibroblasts recently have been discovered in-vitro to promote arrhythmogenesis by direct modification of cardiomyocyte electrophysiology following establishment of heterocellular electrical coupling. In particular, myofibroblasts were found to rescue impulse conduction between disjoined cardiac tissues by acting as passive electrical conduits for excitatory current flow. Although, in principle, such recovery of blocked conduction might be beneficial, propagation across myofibroblast conduits is substantially delayed, thereby promoting arrhythmogenic slow and discontinuous conduction. Second, moderately polarized myofibroblasts were found to induce cell density– dependent depolarization of cardiomyocytes, which causes arrhythmogenic slow conduction due to
the reduction of fast inward currents. Finally, critical depolarization of cardiomyocytes by myofibroblasts was discovered to lead to the appearance of ectopic activity in a model of the infarct border zone. These findings obtained in vitro suggest that electrotonic interactions following gap junctional coupling between myofibroblasts and cardiomyocytes in structurally remodeled fibrotic hearts might directly initiate the main mechanisms underlying arrhythmogenesis, that is, abnormal automaticity and abnormal impulse conduction. If, in the future, similar arrhythmogenic mechanisms can be shown to be operational in intact hearts, myofibroblasts might emerge as a novel noncardiomyocyte target for antiarrhythmic therapy.
Introduction
sembly of cardiomyocytes, fibroblasts contribute importantly to the uniformity of the excitable substrate and, thus, to continuous and fast electrical activation of the working myocardium under physiologic conditions. Given that the fibrillar components of the scaffold are subject to constant turnover, which amounts up to approximately 5% per day,2 it is evident that formation and degradation of these fibers needs to be closely controlled in order to maintain the structural and functional integrity of the myocardium over time. This control is exerted by a large number of biophysical and molecular signaling events that control fibroblast growth and function. If the delicate balance between extracellular matrix (ECM) production and degradation is lost, the working myocardium undergoes structural remodeling that has far-reaching adverse consequences for both the electrical and the pump function of the heart.
Although a wealth of information exists regarding the role of abnormal cardiomyocyte electrophysiology in arrhythmogenesis, much less is known about whether stromal cells of the heart in general and cardiac fibroblasts in particular might be actively involved in arrhythmogenesis. This is surprising insofar as that fibroblasts represent the most numerous cell population in normal human hearts, with individual cells being in intimate contact with cardiomyocytes. In fact, cardiac fibroblasts outnumber cardiomyocytes by a factor of 2–3 while occupying approximately 20% of the volume of the working myocardium.1 Under physiologic conditions, fibroblasts produce and maintain a three-dimensional network of collagen and elastin fibers, which acts as a scaffold for cardiomyocytes and integrates the mechanical forces of individual cells, thus resulting in an efficient pump function of the entire organ. Moreover, by providing the structural backbone for the regular three-dimensional asThis work was supported by the Swiss National Science Foundation (Grant 320000-118247/1). Address reprint requests and correspondence: Dr. Stephan Rohr, Department of Physiology, University of Bern, Buehlplatz 5, CH-3012 Bern, Switzerland. E-mail address: [email protected]. ch. (Received February 12, 2009; accepted February 22, 2009.)
KEYWORDS Myofibroblast; Fibroblast; Arrhythmia; Ectopic activity; Slow conduction; Structural remodeling; Fibrosis; Infarction; Connexin; Cell therapy (Heart Rhythm 2009;6:848 – 856) © 2009 Heart Rhythm Society. All rights reserved.
Structural remodeling and arrhythmogenesis Structural remodeling of the myocardium is the result of a variety of complex cellular reactions to injury and involves both cardiomyocytes and noncardiomyocytes. Histologically, it is characterized by cardiomyocyte hypertrophy, activation and proliferation of fibroblasts, increased ECM
1547-5271/$ -see front matter © 2009 Heart Rhythm Society. All rights reserved.
doi:10.1016/j.hrthm.2009.02.038
Rohr
Myofibroblasts and Cardiac Arrhythmogenesis
849
Figure 1 Etiology and functional consequences of cardiac fibrosis. Microscopic images illustrate connective tissue distribution (collagen: red) in normal myocardium (left) and in a patient with extensive left ventricular fibrosis (right) that results in isolation of bundles of cardiomyocytes. Bar ⫽ 100 m. (Images courtesy Dr. T. Schaffner, Department of Pathology, University of Bern, Bern, Switzerland.)
deposition, and cell death.3 The leading causes of structural remodeling are pressure overload, volume overload, ischemic heart disease, and genetics and old age, which result in changes of the size and shape of the heart and the appearance of interstitial and/or replacement fibrosis (Figure 1). Functionally, structural remodeling is associated with compromised diastolic and/or systolic pump function and an increased likelihood of occurrence of life-threatening arrhythmias.4 Arrhythmias in structurally remodeled myocardia are based on adverse changes in cardiomyocyte electrophysiology (‘electrical remodeling’) and/or on changes in the extent and organization of the ECM (“fibrotic remodeling”). Electrical remodeling relates to arrhythmogenic alterations in the function and expression of ion channels, exchangers, calcium handling proteins, and gap junctions occurring over extended periods of time in response to pathologic stimuli. Fibrotic remodeling, on the other hand, predisposes the heart to arrhythmias due to the presence of excessive amounts of ECM, which is organized into collagenous septa that separate bundles of cardiomyocytes over substantial distances. The resulting reduction of lateral gap junctional coupling results in functional discontinuities within the excitable substrate that disrupts the orderly spread of electrical activation and results in nonuniform conduction5 or zigzag activation of the myocardium.6 Common to both of these abnormal activation patterns is a slowing of macroscopic
conduction velocity and an increased likelihood of occurrence of unidirectional conduction blocks that are central to the initiation of reentrant activations. Arrhythmogenic mechanisms associated with both electrical remodeling and fibrotic remodeling are cardiomyocyte centered in the sense that they concern either pathologic changes in cardiomyocyte electrophysiology or that they relate to the disruption of the orderly three-dimensional network of electrotonically coupled cardiomyocyte by fibrosis. But what about a direct involvement of connective tissue cells that goes beyond the mere hypersecretion of ECM?
Cardiac fibrosis and the appearance of myofibroblasts The absence of arrhythmias in healthy hearts suggests that fibroblasts, even though they greatly outnumber cardiomyocytes, exert no arrhythmogenic effects per se under physiologic conditions. However, under pathologic conditions such as hypertensive heart disease and infarction, an additional type of cell makes its appearance in the working myocardium. These so-called myofibroblasts (sometimes termed activated fibroblasts) are considered to be importantly involved in the establishment of reactive and reparative fibrosis.7,8 Myofibroblasts were first described in the 1970s as fibroblastic cells located within granulation tissue of skin wounds where they initiate contraction of the granulation tissue and disappear thereafter by programmed cell
850
Heart Rhythm, Vol 6, No 6, June 2009 the initial scar formation and the continuous rebuilding of the scar thereafter.16 Although the exact origin of myofibroblasts is still under debate (circulating or resident adventitial or interstitial fibroblasts undergoing a phenotype switch toward myofibroblasts, circulating bone marrow– derived progenitor cells),16 –18 ample evidence indicates that their appearance is triggered by local inflammatory reactions, mechanical stress, and humoral factors.17,19 –21 Induction of myofibroblasts is triggered in particular by transforming growth factor-1 (TGF-1), but other factors such as basic fibroblast growth, angiotensin II, catecholamines, and insulin-like growth factor also are involved in the determination of phenotype and function of cardiac myo/fibroblasts.21
Gap junctional coupling of myofibroblasts to cardiomyocytes
Figure 2 Section of a 4-day-old infarct in the left ventricular free wall of an adult rat. A: Phase contrast image with the infarct outlined by the red dashed line. B: Same region stained for nuclei (DAPI, blue) and ␣-SMA (green). The filamentous green staining corresponds to stress fibers of numerous myofibroblasts within the infarct area. In addition, myofibroblasts can be seen in the immediate vicinity of the infarct (yellow arrows), where they are intercalated between surviving border zone cardiomyocytes.
death.9 In contrast to fibroblasts, myofibroblasts express ␣-smooth muscle actin (␣-SMA), which serves as a routine marker for this cell type. In recent years, myofibroblasts have been found to play a central role in hepatic, pulmonary, pancreatic, and renal fibrosis in which they seem to be the primary source of excessive ECM production and participate in the progression of disease toward organ failure.10 In the heart, myofibroblasts are not normally present with the exception of valve leaflets, where ␣–SMA–positive interstitial cells are found under physiologic conditions.11 Myofibroblasts are, however, constitutive cellular elements of many of the cardiac pathologies mentioned. For example, animal models of pressure overload have shown that early inflammatory changes are followed by myofibroblast proliferation within days and subsequent development of perivascular and interstitial fibrosis.12,13 Similarly and as shown in Figure 2, myofibroblasts appear in infarcts in very large numbers a few days after the onset of myocardial infarction and, due to incomplete apoptosis during the healing process, persist locally for many years.14,15 In infarcts, myofibroblasts are recognized for their central role in both
Given the close apposition between cardiomyocytes and myofibroblasts in diseased myocardia, the question arises as to whether gap junctions might electrically interconnect the two cell types. Whereas connexin (Cx) expression is a well-established feature of myofibroblasts in tissues different from heart, such as suburothelial tissue,22 intestine,23 breast cancer stroma,24 and healing skin wounds,25 the question of gap junctional coupling between myofibroblasts and cardiomyocytes in diseased myocardia is still open. Interestingly, however, fibroblasts residing in infarct scars have shown abundant expression of Cx43 and Cx45.26 Given that other investigators have shown that “fibroblasts” in infarct scars are, in fact, myofibroblasts (cf. also Figure 2),15,27 it is tempting to speculate that the connexin-expressing “fibroblasts” in the former study actually were connexin-expressing myofibroblasts residing in the infarct region. In contrast to the lack of data characterizing formation of gap junctions between myofibroblasts and cardiomyocytes in intact myocardia, heterocellular electrotonic coupling is a long-known feature of cultured cardiomyocytes and fibroblasts. As early as 1969, Hyde et al28 observed that fibroblasts inserted between cardiomyocytes synchronized spontaneous contractions and, using sharp electrode impalements, showed that cardiomyocytes in contact with what they considered to be fibroblasts exhibit reduced membrane potentials. In the meantime, fibroblasts in this model system were found to display a myofibroblast phenotype and to express Cx43 and Cx45 (but not Cx40) among themselves and at contact sites with cardiomyocytes (Figure 3).29 Thus, at least in vitro, myofibroblasts are capable of establishing gap junctional coupling both among themselves and with cardiomyocytes, which has, as outlined in more detail later, profound arrhythmogenic consequences.
Myofibroblast mend broken pathways of impulse conduction It is generally assumed that cardiac impulse conduction is blocked at sites where the cardiomyocyte network is disrupted by collagenous septa or at sites of sutures following heart transplantation. Interestingly, for the latter case, it has
Rohr
Myofibroblasts and Cardiac Arrhythmogenesis
851
Figure 3 Homocellular and heterocellular gap junctional coupling of cultured cardiac myofibroblasts. Top row: Immunofluorescent labeling for Cx43 (green) and ␣-SMA (red) shows expression of Cx43 both at myofibroblast–myofibroblast borders (yellow arrows) and at myofibroblast– cardiomyocyte contact sites (white arrows). The corresponding phase contrast image shows the spatial arrangement of myofibroblasts (MFB) and cardiomyocytes (CM). Bottom row: Same as top row but for Cx45, which is similarly expressed at both homocellular and heterocellular cell junctions. (Modified with permission from Miragoli M, Gaudesius G, Rohr S. Electrotonic modulation of cardiac impulse conduction by myofibroblasts. Circ Res 2006;98:801– 810.)
been reported that recipient and donor atria are sporadically capable of establishing electrical synchronization despite scar formation in the region of the suture.30 Because myofibroblasts are a typical cellular component of scar tissue, we tested the hypothesis that these cells are capable of relaying electrical activity between disjoined regions of cardiac tissue by acting as a passive electrical conduit for excitatory current flow.31 For this purpose, we developed a photolithographic method that permits the construction of a geometrically defined in vitro cell model of a transplantation scar.32 The model consisted of two strands of neonatal rat ventricular cardiomyocytes interconnected by a region containing myofibroblasts of cardiac origin only (Figure 4A). The preparations were stimulated with an extracellular electrode placed several length constants away from the measurement site, and impulse propagation across the myofibroblast insert was followed optically using voltage-sensitive dyes.33 Similar to intracellular electrode recordings, these dyes indicate changes in transmembrane potential in real time and therefore permit characterization of electrical activation patterns with high spatiotemporal resolution. As illustrated by the panel depicting action potential upstrokes along the preparation, local stimulation was followed by activation of the entire preparation, indicating successful propagation across the myofibroblast insert. However, as shown by the plot of local activation times versus distance, successful bridging of electrical activation by myofibroblasts came at the price of a substantial local activation delay of approximately 30 ms. As shown in Figure 4B,
conduction delays across myofibroblast inserts were a function of insert width and rose monotonically up to approximately 60 to 70 ms for the largest inserts still supporting synchronization of electrical activity. At widths greater than 300 m, propagation invariably failed. These experiments demonstrate that gap junctional myofibroblast-to-cardiomyocyte coupling can restore conduction across discontinuities in networks of cardiomyocytes. By acting as passive electrical conduits for excitatory current flow, myofibroblasts might contribute to the explanation of electrical synchronization of recipient and donor tissue following heart transplantation. With regard to arrhythmogenesis, “mending of broken pathways” might have both proarrhythmic and antiarrhythmic consequences depending on the setting. Whereas elimination of conduction blocks per se can be expected to counteract arrhythmogenesis, the substantial conduction delays incurred at sites where myofibroblasts act as conduits for excitatory current flow predict the emergence of discontinuous slow conduction. In fact, in the example shown in Figure 4, apparent local conduction velocities within the central 200 m encompassing the myofibroblast insert amounted to approximately 5 mm/s, which corresponds to an approximately 100-fold slowing of conduction compared with cardiomyocyte only strands (⬃400 mm/s). If such instances of local propagation delays were to be concatenated in settings such as interstitial fibrosis, macroscopic impulse conduction would become very slow and the risks for arrhythmogenesis would increase accordingly.
852
Heart Rhythm, Vol 6, No 6, June 2009 blasts (Figure 5A). Impulse conduction characteristics along these preparations were assessed optically and, following the experiments, the preparations were stained for ␣-SMA in order to determine local myofibroblast densities. Compared with uncoated cardiomyocyte preparations, conduction was considerably slowed in the presence of myofibroblasts. At the same time, maximal upstroke velocities of the propagated action potentials were reduced. This suggested that myofibroblasts, which exhibit a low intrinsic membrane potential of approximately – 40 to –50 mV, induced partial depolarization of cardiomyocytes following heterocellular gap junctional coupling and, hence, caused a reduction of the sodium inward current. Intracellular measurements using sharp electrodes confirmed this hypothesis (Figure 5B). Maximal diastolic potentials measured in cardiomyocytes decreased from approximately – 80 mV at myofibroblast densities less than 5% to approximately –55 mV at myofi-
Figure 4 Myofibroblasts serve as passive electrical conduits for impulse conduction. A: Experimental preparations consisted of two strands of cardiomyocytes that were joined by a myofibroblast insert. Electrical activity was recorded optically from the sites marked by circles in the phase contrast image of the preparation (orange: cardiomyocyte region; green: myofibroblast region). Following stimulation, action potential upstrokes could be recorded along the entire preparations, indicating that myofibroblasts successfully synchronized electrical activity between the two cardiomyocyte strands. The plot of local activation times versus distance indicates that conduction across the 134-m-wide myofibroblast insert was delayed by approximately 30 ms. B: Propagation delays across myofibroblast inserts rose monotonically with insert width. At widths greater than 300 m, synchronization of electrical activity invariably failed. (Modified with permission from Gaudesius G, Miragoli M, Thomas SP, et al. Coupling of cardiac electrical activity over extended distances by fibroblasts of cardiac origin. Circ Res 2003;93:421– 428.)
Myofibroblasts induce slow conduction Structurally, the border zone of healing infarcts is characterized by surviving cardiomyocytes in close contact with a large number of myofibroblasts (cf. Figure 2). Functionally, this very same region is known to contribute significantly to postinfarction arrhythmogenesis based on remodeling of the cellular microarchitecture, modifications of gap junctional coupling among cardiomyocytes, and changes in the electrophysiology of border zone cardiomyocytes (for review see Peters and Wit34). Given that myofibroblasts communicate via gap junctions with cardiomyocytes, we investigated whether this type of interaction might have additional arrhythmogenic consequences. For this purpose, we generated a simplified model of the infarct border zone, which consisted of strands of cardiomyocytes coated with myofibro-
Figure 5 Myofibroblasts induce slow conduction. A: Schematic diagram of the experimental preparation consisting of cardiomyocyte strands coated with myofibroblasts. Impulse propagation along the preparation following stimulation was measured optically at the sites indicated by circles in the phase contrast picture. Postexperimental staining for ␣-SMA (green overlay) served to determine the percentage area of the strands covered with myofibroblasts. As indicated by the plot of local activation times versus distance, conduction along the hybrid cell preparations was continuous but slow. B: Myofibroblasts reduce resting membrane potential of cardiomyocytes in a cell density– dependent manner. C: Maximal action potential upstroke velocities and conduction velocities along the preparations show a biphasic dependence on myofibroblast density. (Modified with permission from Miragoli M, Gaudesius G, Rohr S. Electrotonic modulation of cardiac impulse conduction by myofibroblasts. Circ Res 2006;98:801– 810.)
Rohr
Myofibroblasts and Cardiac Arrhythmogenesis
broblast densities greater than 40%. This graded depolarization was paralleled by a biphasic change in conduction velocities that initially slightly increased from approximately 390 mm/s to 440 mm/s before, at myofibroblast densities greater than 10%, monotonically falling to 260 mm/s (Figure 5C). This biphasic dependence is, in fact, the expected result for a gradual depolarization of cardiomyocytes by myofibroblasts, as it replays exactly what is known for many decades for intact cardiac tissue subjected to gradual depolarization by potassium: also there, conduction velocity transiently increases (“supernormal conduction”)
853 before increasing sodium channel inactivation causes progressive conduction slowing. These data suggest that heterocellular electrotonic interactions in the border zone of healing infarcts or along laminae of interstitial fibrotic tissue might cause arrhythmogenic slow conduction based on partial depolarization of cardiomyocytes by myofibroblasts. Under the assumption that myofibroblasts in diseased hearts exert similar effects on cardiomyocytes as those found in vitro, the degree of conduction slowing in the infarct border zone likely depends on the relative densities and the three-dimensional
Figure 6 Myofibroblasts induce abnormal automaticity. A, top: Schematic diagram of the layout of the experimental preparations consisting of cardiomyocyte strands coated with myofibroblasts. Middle left: Overview of a preparation consisting of 24 identical strands (0.6 ⫻ 4.5 mm each) that were scanned simultaneously for the presence of spontaneous electrical activity with a high-speed camera following staining with voltage sensitive dyes. Snapshot shows two strands exhibiting spontaneous activity (green: area depolarized by propagating action potential; pink star: origin of abnormal automaticity). Middle right: Details of the cellular microarchitecture (phase contrast image) and the density and distribution of myofibroblasts (vimentin staining) of a selected strand. Bottom: Continuous optical recording of action potentials at the sites denoted with a blue and red circle in the overview shows regular spontaneous electrical activity. B: Incidence of abnormal automaticity as a function of myofibroblast density. (Modified with permission from Miragoli M, Salvarani N, Rohr S. Myofibroblasts induce ectopic activity in cardiac tissue. Circ Res 2007;101:755–758.)
854
Heart Rhythm, Vol 6, No 6, June 2009
coupling pattern of the two cell populations. Within the epicardial/endocardial border zone consisting of a thin and essentially two-dimensional sheet of surviving cardiomyocytes, the large number of myofibroblasts beneath might slow conduction considerably. In contrast, within the myocardial wall, the depolarizing effect of myofibroblasts likely is offset by the dense three-dimensional network of wellcoupled cardiomyocytes bordering onto the infarct area. The combination of the two then would result in heterogeneous conduction properties of the excitable substrate around the infarct, which might additionally contribute to the established arrhythmogenic nature of the infarct border zone.
Myofibroblasts induce abnormal automaticity In diseased hearts, cardiac fibrosis not only induces arrhythmogenic slow and discontinuous conduction; it also is suspected to promote ectopic activity, which is central to the generation of focal and reentrant tachyarrhythmias.35 The presence of ectopic activity under these conditions is thought to be favored by anisotropies in tissue structure and in electrical coupling that ultimately permits a few cardiomyocytes exhibiting abnormal automaticity to drive the surrounding myocardium.36 Given that myofibroblasts reduce the resting membrane potentials of cardiomyocytes
Figure 7
following heterocellular electrical coupling, we hypothesized that ectopic activity in fibrotic myocardium alternatively might be based on the concept of depolarizationinduced automaticity.37 The model used to investigate this hypothesis was designed similar to the infarct border zone model and consisted of strands of cardiomyocytes coated with myofibroblasts (Figure 6A). Following staining with voltage-sensitive dyes, spontaneous activity was recorded for repeated periods of 4 seconds with a macroscope that permitted simultaneous monitoring of 24 individual strands. The overview of the preparations in Figure 6A corresponds to a single frame taken from such a recording. It illustrates the general layout of the patterned growth preparations and shows two strands undergoing spontaneous activation. As demonstrated by the continuous recordings in one strand, ectopic activity was regular with action potentials occurring at a frequency of approximately 1 Hz. As shown in Figure 6B, occurrence of spontaneous activity was steeply dependent on myofibroblast density. Whereas preparations were completely quiescent at myofibroblast densities less than 16%, more than 80% exhibited spontaneous activity at densities greater than 40%. Spontaneous activity was the result of depolarization-induced automaticity as demonstrated by the finding that hyperpolarization of the preparations with
Summary of mechanisms by which myofibroblasts might contribute to arrhythmogenesis. ECM ⫽ extracellular matrix.
Rohr
Myofibroblasts and Cardiac Arrhythmogenesis
KATP channel openers led to immediate cessation of abnormal automaticity.38 Moreover, graded reductions of resting membrane potentials in single cardiomyocytes using patch clamp techniques started to induce spontaneous activity at levels of membrane depolarization (approximately – 67 mV) highly similar to those present at minimal myofibroblast densities necessary to induce ectopic activity (approximately – 66 mV). These findings suggest that myofibroblasts in intimate contact with cardiomyocytes at sites such as the infarct border zone or along muscle bundles ensheathed by connective tissue might elicit focal ectopic activity. Mechanistically, this situation closely resembles the concept of arrhythmogenic “injury current” flow between damaged and healthy cardiomyocytes in acute ischemia except that, in the present case, myofibroblasts rather than injured cardiomyocytes are the source of depolarizing current. The chances of this novel arrhythmogenic mechanism becoming operational in the fibrotically remodeled hearts in situ will likely depend on a number of different parameters, such as the membrane potentials of myofibroblasts and cardiomyocytes in situ, the respective cell sizes, the degree of homocellular and heterocellular coupling among the two cell types, and the specifics of local cellular tissue architecture affecting source-to-load matching.
Conclusions and perspectives The findings presented demonstrate that myofibroblasts, as schematically summarized in Figure 7, are capable of exerting highly arrhythmogenic effects following heterocellular electrotonic coupling to cardiomyocytes. By inducing (dis)continuous slow conduction as well as ectopic activity, myofibroblasts are, in fact, capable of launching the main mechanisms driving focal and reentrant tachyarrhythmias, that is, abnormal impulse conduction and abnormal impulse formation.39 Although this finding opens the perspective that this particular cell type might constitute a new active player in arrhythmogenesis, extrapolation of the in vitro findings presented to intact diseased myocardia has to await demonstrations of analogous structural and functional myofibroblast-to-cardiomyocyte interactions in situ. If such interactions were to be found, novel approaches to antiarrhythmic treatments might emerge that could be targeted specifically at myofibroblasts and not, as in the past, at cardiomyocytes. The advantages of such myofibroblast-centered antiarrhythmic treatments would be manifold, with two major aspects: (1) relief of arrhythmias irrespective of the underlying cause (mechanical overload, ischemic heart disease, etc.) and (2) treating myofibroblasts instead of modifying cardiomyocyte electrophysiology could circumvent the potentially dangerous complexities associated with the latter approach.40 Among a substantial number of options, therapies affecting myofibroblast appearance and function could include treatments as diverse as the use of statins, inhibition of the renin-angiotensin-aldosterone system, and interference with cytokines known to modulate the phenotype of this cell. All of these approaches have been shown to affect myofibro-
855 blast biology in vitro and to counteract arrhythmogenesis in vivo. Whether these antiarrhythmic effects were solely due to a reduction of fibrosis, were based on modifications of the electrotonic cross-talk between cardiomyocytes and myofibroblasts, or occurred secondary to a change in paracrine activity of the myofibroblasts remains to be determined. Apart from their possible relevance for arrhythmogenesis in fibrotically remodeled hearts, the results presented also might have relevance for cell therapies using communication competent cell types that, compared with cardiomyocytes, have a reduced intrinsic membrane potential.41 In analogy to myofibroblasts, such cells might critically depolarize neighboring cardiomyocytes after establishment of heterocellular gap junctional coupling, thus contributing to arrhythmogenesis by induction of slow conduction and precipitation of abnormal automaticity. Further studies are needed to determine whether this mechanism contributes to arrhythmias following cell therapy and whether special attention must be given to the degree of membrane polarization of the cells transplanted.
Acknowledgment I wish to thank Dr. Christian Mühlfeld for providing sections of infarcted rat hearts.
References 1. Maisch B. Extracellular matrix and cardiac interstitium: restriction is not a restricted phenomenon. Herz 1995;20:75– 80. 2. McAnulty RJ, Laurent GJ. Collagen synthesis and degradation in vivo. Evidence for rapid rates of collagen turnover with extensive degradation of newly synthesized collagen in tissues of the adult rat. Coll Relat Res 1987;7:93–104. 3. Swynghedauw B. Molecular mechanisms of myocardial remodeling. Physiol Rev 1999;79:215–262. 4. Adamson PB, Barr RC, Callans DJ, et al. The perplexing complexity of cardiac arrhythmias: beyond electrical remodeling. Heart Rhythm 2005;2:650 – 659. 5. 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;20:397– 413. 6. De Bakker JMT, Van Capelle FJL, Janse MJ, et al. Slow conduction in the infarcted human heart: zigzag course of activation. Circulation 1993;88:915– 926. 7. Weber KT. Extracellular matrix remodeling in heart failure: a role for de novo angiotensin II generation. Circulation 1997;96:4065– 4082. 8. Weber KT. Fibrosis in hypertensive heart disease: focus on cardiac fibroblasts. J Hypertension 2004;22:47–50. 9. Gabbiani G. The myofibroblast in wound healing and fibrocontractive diseases. J Pathol 2003;200:500 –503. 10. Kisseleva T, Brenner DA. Fibrogenesis of parenchymal organs. Proc Am Thorac Soc 2008;5:338 –342. 11. Aikawa E, Whittaker P, Farber M, et al. Human semilunar cardiac valve remodeling by activated cells from fetus to adult: implications for postnatal adaptation, pathology, and tissue engineering. Circulation 2006;113:1344 – 1352. 12. Leslie KO, Taatjes DJ, Schwarz J, et al. Cardiac myofibroblasts express alpha smooth muscle actin during right ventricular pressure overload in the rabbit. Am J Pathol 1991;139:207–216. 13. Kuwahara F, Kai H, Tokuda K, et al. Transforming growth factor function blocking prevents myocardial fibrosis and diastolic dysfunction in pressureoverloaded rats. Circ 2002;106:130 –135. 14. Peterson DJ, Ju H, Panagia M, et al. Expression of Gi-2␣ and Gs␣ in myofibroblasts localized to the infarct scar in heart failure due to myocardial infarction. Cardiovasc Res 1999;41:575–585. 15. Willems IE, Havenith MG, De Mey JG, et al. The alpha-smooth muscle actinpositive cells in healing human myocardial scars. Am J Pathol 1994;145:868 – 875. 16. Sun Y, Kiani MF, Postlethwaite AE, et al. Infarct scar as living tissue. Basic Res Cardiol 2002;97:343–347.
856 17. Desmouliere A, Chapponier C, Gabbiani G. Tissue repair, contraction, and the myofibroblast. Wound Rep Reg 2005;13:7–12. 18. van Amerongen M, Bou-Gharios G, Popa E, et al. Bone marrow-derived myofibroblasts contribute functionally to scar formation after myocardial infarction. J Pathol 2008;214:377–386. 19. Weber KT. From inflammation to fibrosis: a stiff stretch of highway. Hypertension 2004;43:716 –719. 20. Kuwahara F, Kai H, Tokuda K, et al. Hypertensive myocardial fibrosis and diastolic dysfunction. Hypertension 2004;43:739 –745. 21. Manabe I, Shindo T, Nagai R. Gene expression in fibroblasts and fibrosis. Circ Res 2002;91:1103–1113. 22. Sui GP, Rothery S, Dupont E, et al. Gap junctions and connexin expression in human suburothelial interstitial cells. BJU Int 2002;90:118 –129. 23. Powell DW, Mifflin RC, Valentich JD, et al. Myofibroblasts. II. Intestinal subepithelial myofibroblasts. Am J Physiol 1999;277:C183–C201. 24. Jamieson S, Going JJ, D’Arcy R, et al. Expression of gap junction proteins connexin 26 and connexin 43 in normal human breast and in breast tumours. J Pathol 1998;184:37– 43. 25. Gabbiani G, Chaponnier C, Huttner I. Cytoplasmic filaments and gap junctions in epithelial cells and myofibroblasts during wound healing. J Cell Biol 1978; 76:561–568. 26. Camelliti P, Devlin GP, Matthews KG, et al. Spatially and temporally distinct expression of fibroblast connexins after sheep ventricular infarction. Cardiovasc Res 2004;62:415. 27. Sun Y, Weber KT. Angiotensin converting enzyme and myofibroblasts during tissue repair in the rat heart. J Mol Cell Cardiol 1996;28:851– 858. 28. Hyde AB, Blondel B, Matter A, et al. Homo- and heterocellular junctions in cell cultures: an electrophysiological and morphological study. Prog Brain Res 1969;31:283–311. 29. Miragoli M, Gaudesius G, Rohr S. Electrotonic modulation of cardiac impulse conduction by myofibroblasts. Circ Res 2006;98:801– 810. 30. Lefroy DC, Fang JC, Stevenson LW, et al. Recipient-to-donor atrioatrial con-
Heart Rhythm, Vol 6, No 6, June 2009
31.
32.
33.
34. 35.
36.
37. 38. 39.
40.
41.
duction after orthotopic heart transplantation: surface electrocardiographic features and estimated prevalence. Am J Cardiol 1998;82:444 – 450. Gaudesius G, Miragoli M, Thomas SP, et al. Coupling of cardiac electrical activity over extended distances by fibroblasts of cardiac origin. Circ Res 2003; 93:421– 428. Rohr S, Flückiger-Labrada R, Kucera JP. Photolithographically defined deposition of attachment factors as a versatile method for patterning the growth of different cell types in culture. Eur J Physiol 2003;446:125–132. 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;75:1062–1075. Peters NS, Wit AL. Myocardial architecture and ventricular arrhythmogenesis. Circulation 1998;97:1746 –1754. De Bakker JMT, Stein M, van Rijen HVM. Three-dimensional anatomic structure as substrate for ventricular tachycardia/ventricular fibrillation. Heart Rhythm 2005;2:777–779. Wilders R, Wagner MB, Golod GA, et al. Effects of anisotropy on the development of cardiac arrhythmias associated with focal activity. Eur J Physiol 2000;441:301–312. Mohabir R, Ferrier GR. Effects of ischemic conditions and reperfusion on depolarization-induced automaticity. Am J Physiol 1988;255:H992–H999. Miragoli M, Salvarani N, Rohr S. Myofibroblasts induce ectopic activity in cardiac tissue. Circ Res 2007;101:755–758. Zlochiver S, MuÒoz V, Vikstrom KL, et al. Electrotonic myofibroblast-tomyocyte coupling increases propensity to reentrant arrhythmias in two-dimensional cardiac monolayers. Biophys J 2008;95:4469 – 4480. Echt DS, Liebson PR, Mitchell LB, et al. Mortality and morbidity in patients receiving encainide, flecainide, or placebo. The cardiac arrhythmia suppression trial. N Engl J Med 1991;324:781–788. Heubach JF, Graf EM, Leutheuser J, et al. Electrophysiological properties of human mesenchymal stem cells. J Physiol 2004;554:659 – 672.