Bachelor thesis
The role of fibroblasts and the extracellular matrix in cardiac remodeling and regeneration
Hans van Luit -‐ s2196549 Supervisor: prof. dr. M.C. Harmsen June 19th 2015
Abstract
The low turnover number of cardiomyocytes (CM) contributes to the progressive adverse effects after heart injury. Much research focusses on the proliferation of CM. However, cardiac fibroblasts may play a major role in the process of cardiac regeneration since they make up for half of the total cell number in the myocardium. Therefore, in this review, the regeneration of myocardium is discussed in a broader perspective by investigating the interactions between fibroblasts, macrophages and CM. Under pathological conditions, the extracellular matrix (ECM) stiffens, and type I-‐receptor transforming growth factor (TGFβ1) bioavailability and signalling increases, which stimulate fibroblasts to differentiate into myofibroblasts. These α-‐smooth muscle actin (SMA) expressing fibroblasts, show an increased production of ECM proteins in response to TGFβ1 and increased stiffness. Macrophages play a role in these processes via angiotensin II (ATII) signalling and TGFβ1 production. CM-‐fibroblast interactions are complex and involve many types of signal transduction including TGFβ1, FGF2 and FGF16, Notch signalling, indirect signalling via the ECM and signalling via gap junctions between both cell types. Recent developments regarding cardiac research showed increased CM proliferation and decreased formation of chronic scars in neonatal rodents after cardiac injury. Additionally, evidence exists for increased CM proliferation in young humans compared to adult, indicating a different role for fibroblasts during early stages of life compared to adulthood. Some of these differences can be explained by altered CM-‐fibroblast interactions. This review provides recent insights in the relations between different cell types involved in cardiac remodeling and discusses some recent developments in cardiac regeneration in a broader perspective.
Abbreviations
MI: myocardial infarct, CM: cardiomyocytes, ECM: extracellular matrix, SMA: α-‐smooth muscle actin, TGFβ1: type I-‐receptor fibrogenic growth factors, ATII: angiotensin II, ACE: angiotensin-‐
converting enzyme, CTGF: connective tissue growth factor, MMPs: matrix metalloproteinases, TRβI: TGFβ type I receptor, TRβII: TGFβ type II receptor, TAK1: TGFβ-‐activated kinase 1, LAP:
latency-‐associated protein, LTBP1: TGF-‐β1 binding protein-‐1, LLC: large latency complex, LOX:
lysyl oxidase, FGF: fibroblast growth factor, HBEGF: heparin-‐binding EGF-‐like growth factor, Nrg: Neuregulin.
ABSTRACT ... 2
1 – INTRODUCTION ... 4
2 – CARDIAC FIBROBLASTS ... 5
2.1 The role of fibroblasts and myofibroblasts ... 5
2.2 The differentiation of cardiac fibroblasts to myofibroblasts ... 5
3 – MACROPHAGE-‐FIBROBLAST INTERACTIONS ... 6
3.1 Angiotensin signalling ... 6
3.2 TGFβ signalling ... 7
4 – FIBROBLASTS AND MECHANICAL STRESS ... 7
4.1 Mechanical stress and TGF-‐β1 ... 7
4.2 Mechanical stress induces SMA protein expression and control of SMA fibers ... 7
4.3 Mechanical stress and the AT1-‐receptor ... 8
5 – FIBROBLAST – CARDIOMYOCYTE INTERACTIONS ... 9
5.1 Interactions between fibroblasts and cardiomyocytes ... 9
5.2 Interactions via soluble mediators ... 9
5.3 Indirect cardiomyocyte-‐fibroblast interactions via the ECM ... 10
5.4 Interactions via cell-‐cell contact ... 11
6 -‐ RECENT DEVELOPMENTS IN RESEARCH ON CARDIAC REGENERATION ... 12
6.1 The proliferative window of CM ... 12
6.2 Neuregulin signalling ... 12
7 – DISCUSSION ... 13
7.1 Summary of interactions between cell types ... 13
7.2 Future perspectives ... 14
8 -‐ LIST OF REFERENCES ... 14
1 – Introduction
In the Netherlands, 130,000 people suffer from heart failure (Hartstichting, 2012). Heart failure is a common long-‐term side effect after a myocardiac infarct (MI). Cardiac infarcts are caused by the blockage of blood supply to the heart. This generally results in necrosis and apoptosis of cardiac tissue, including the cells responsible for the contractile property of the heart: CM (CM).
For a major part, these cells are replaced by scar tissue, formed by differentiated fibroblasts:
myofibroblasts. Due to the loss of CM and the increase of scar tissue, a patient suffering from heart failure is unable to pump sufficient blood to meet the body's requirements. In summary, the main cause of heart failure is the lack of an appropriate regenerative response of the heart to cell death. Finding a way to initiate such a response could open a way to cure cardiac diseases like heart failure.
However, Goldemberg showed in 1886 that the adult mammalian heart was terminally differentiated, and, therefore, lacks any regenerative capacity (Goldemberg, 1886). This dogma held for over a hundred years. It was not until the early seventies, when first evidence of cardiac regeneration in lower vertebrates was provided (e.g. Rumyantsev, 1961).
Evidence of human cardiomyocyte proliferation was only provided recently. One of the most influential researches in this area was that of Bergmann and colleagues. In this research, advantage was taken of the integration of carbon-‐14, generated by nuclear bomb tests during the Cold War, into DNA. The presence of carbon-‐14 in CM was monitored during several years and this data was used to establish the age of CM in humans. This experiment revealed that CM slowly renew, with a rate of approximately 1% per year at the age of 25 (Bergmann et al., 2009).
The slow regeneration rate of CM forms one of the problems regarding cardiac regeneration.
Another major problem in cardiac regeneration is the formation of scar tissue, which prevents the regeneration and impairs the function of CM at the site of injury. However, Porrello and colleagues showed that cardiac regeneration without chronic scarring in mammals is not impossible. However, this research was performed with mice and the regenerative capacities of CM disappeared seven days after birth (Porrello, 2011).
Although plenty of research on proliferation of CM has been performed, less attention is paid to cardiac fibroblasts and macrophages and their relation to myocardial regeneration, in particular their influence on CM. In this review, processes related to the relation between cardiac fibroblasts and CM under normal and pathological conditions are reviewed an illustrated. First, in section 2, the role of cardiac fibroblast and their ability to differentiate will be discussed, followed by the interactions between macrophages and fibroblast in section 3.
Thereafter, the role of mechanical stress and the interactions between fibroblasts and CM will be discussed in section 4 and 5, respectively. Subsequently, in section 6, recent developments regarding cardiac regeneration will be reviewed. Finally, in the discussion section, a summary of these interactions will be given and some connections to recent developments in cardiac regeneration will be made.
2 – Cardiac fibroblasts
2.1 The role of fibroblasts and myofibroblasts
Fibroblasts in the myocardium (cardiac muscle) make up for approximately 10% of the heart's volume. However, due to their small size, they account for half of the heart's total cell number (Shiraishi, 1992; Srivastava, 2006). Fibroblasts are responsible for the deposition and organisation of the extracellular matrix (ECM) and thus, maintaining the cardiac ECM homeostasis. Besides this structural role, fibroblasts play an important role at influencing the electrical activity of CM, by the ability of fibroblasts to couple with CM (Goldsmith, 2014).
Crosstalk between cardiac fibroblasts and CM is various and involves the secretion of growth factors, cytokines and paracrines by both cells. Additionally, communication between cells is facilitated by components of the matrix. Each of these interactions could be modulated by mechanical stress (Abramochkin, 2014).
In pathological conditions, like heart failure, cardiac fibroblasts are able to differentiate into myofibroblasts, which express α-‐smooth muscle actin (SMA) contractile units (Gerling, 2003). Myofibroblasts play an important role in a process called scar contraction. This takes place in the later phase of wound healing (days to weeks after the injury) and is executed by cross-‐linking collagen fibers, which increases the tensile forces and cause contraction of the myofibroblasts. This process of 'scar thinning' leads to changes in tissue density and stiffness (van den Borne, 2010). Myofibroblasts are, in response to type I-‐receptor fibrogenic growth factors (TGF-‐β1) also able to excrete matrix molecules to a greater extinct than normal fibroblasts, especially the stiff, cross-‐linked type I collagen protein (Swynghedauw, 1999). Both of these adaptions are necessary to maintain the structural integrity of the myocardium after CM loss, although these adaptations lead to increased stiffness of the tissue. The importance of this early reaction is elucidated by an experiment in which the cardiac fibroblast activation was inhibited after injury. This led to impaired wound healing and worsening of cardiac functions (Duan, 2012). Additionally, fibroblasts and myofibroblasts interact with CM and are thought to play a major role in CM proliferation and hypertrophy (Ieda, 2009). Hypertrophy of CM is as a common adaptive response to the loss of CM (as seen after MI) and is associated with a decreased capacity to differentiate. Therefore, hypertrophy of CM after injury impairs the regenerative mechanism of CM proliferation (Zebrowski, 2013).
2.2 The differentiation of cardiac fibroblasts to myofibroblasts
In normal physiological conditions, fibroblasts maintain ECM homeostasis and myofibroblasts are hardly detectable in the myocard (Weber, 2013). In pathological conditions, for example, after a MI, fibroblasts differentiate into myofibroblasts. However, it is arguable that CM are in fact less differentiated than cardiac fibroblasts. This view is strengthened by the fact that CM expresses SMA, which is during embryogenesis an ealier form of actin expressed by CM. The main difference between this differentiated variant and fibroblasts is the ability of myofibroblasts to express SMA. However, an in-‐between differentiated fibroblast exists: the proto-‐myofibroblast, containing actin/myosin-‐based stress fibers, but lacking SMA (Hinz, 2010).
The differentiation of cardiac fibroblasts is dependent on ECM-‐derived TGFβ1 and mechanical stress (Hinz, 2010). These interactions between the ECM and fibroblasts will be further discussed in section 4. It is shown that myofibroblasts also originate from other cell types in pathological conditions, for example by from endothelial and epithelial cells: endothelial (EndMT) or epithelial (EMT) to mesenchymal transdifferentiation, processes which are also seen during embryogenesis (Davis, 2013). Another source of myofibroblasts is the transdifferentiation of circulating fibrocytes (bone marrow derived stem cells). Little literature of lineage sources of myofibroblasts is available, so the relative contibutions of myofibroblast progenitors is unclear (Chang, 2002). However, it is shown that in the kidneys 35% of the myofibroblasts are derived from fibrocytes, 10% from EndMT and 5% from EMT (LeBlue, 2013).
Additionally, the cellular origins of myofibroblasts may be dependent on the degree of inflammation; the contribution of blood-‐derived progenitors like fibrocytes may be more significant in conditions associated with more intense inflammatory responses (Kong, 2014).
Fig. 1 -‐ The interactions between macrophages and myofibroblasts. Macrophages are able to produce angiotensin I, which stimulates the AT1-‐receptor in an autocrine manner. Stimulation of this receptor leads to an upregulation of TGF-‐β1 production, which triggers the appearance of myofibroblasts.
Myofibroblasts also express AT1-‐receptors, which enable them as well to express TGF-‐β1 in response to angiotensin II. The upregulation of TGF-‐β1 production in myofibroblasts leads to the formation and excretion of fibronectin, collagen types I and III and MMPs (Weber, 2013).
3 – Macrophage-‐fibroblast interactions
3.1 Angiotensin signalling
To elucidate the role of TGF-‐β1 in ECM deposition, the role between macrophages and myofibroblasts will be discussed in this section.
Necrosis of CM leads to a local elevation of necrosis-‐associated proteins, which function as danger signals to the innate immune system. The first days after the necrosis, inflammatory cells are recruited, guided by gradient concentrations of tissue chemokines to the site of the injury in order to scavenge the necrotic cells and induce a healing response. Among these inflammatory cells are macrophages expressing angiotensinogen, renin and angiotensin-‐converting enzyme (ACE), which enable the forming of angiotensin II (ATII) (Sun, 2001). This production of ATII fulfils an autocrine function by binding to the AT1 (angiotensin II type I)-‐receptor of macrophages, hereby inducing expression of TGF-‐β1. This growth factor is associated with the formation of myofibroblasts at the site of injury. Angiotensin signalling triggers the myofibroblasts into further TGF-‐β1 production (Sun, 1998). Some of the interactions between macrophages and myofibroblasts are shown in figure 1. Not shown in this figure is the activation of TGFβ receptors on myofibroblast and macrophages cell membranes, which causes upregulation of collagen production via connective tissue growth factor (CTGF) and fibronectin (Kong, 2014; Weber, 2013). Myofibroblasts are able to activate the AT1-‐receptor in an autocrine manner, since they produce ATII and thus, cross-‐activation of the AT1-‐receptor between macrophages and myofibroblasts can take place. This creates a feed-‐forward loop. Additionally, CM are also able to secrete ATII under pathological conditions: Tsai Chia-‐Ti showed that rapid depolarization of atrial CM induced angiotensin II secretion (Tsai Chia-‐Ti, 2011). By these mechanisms, the formation of a 'secretome' is induced: the total deposition of organic molecules by the myofibroblasts, including myofibroblasts and macrophages. As can be seen in figure 1, TGFβ1 signalling leads to inhibition of matrix metalloproteinases (MMP) production in myofibroblasts (Ye, 2011), hereby increasing the turnover of matrix molecules, since MMPs degrade matrix molecules. However, this turnover eventually stabilizes, but remains in scar tissue. This stabilizing is due to the remaining MMPs at the sites of the scarring (Cleutjens, 1995).
3.2 TGFβ signalling
TGFβ exists in three different isoforms (TGFβ1, TGFβ2 and TGFβ3) of which TGFβ1 is found almost ubiquitously in the human body. TGFβ1 is generally released in its latent form, unable to interact with its receptor: the TGFβ type II receptor (TRβII). Activation of TGFβ is dependent on mechanical stress, a mechanism that will be discussed in section 5. Binding of TGFβ1 to its receptor leads to recruitment and phosphorylation of another receptor, TGFβ type I receptor (TRβI). Activation of TRβI leads to downstream intracellular signalling via Smad proteins.
Smad2/3 are able to phosphorylate Smad4, which consequently translocates to the nucleus and influences (ECM-‐ related) gene transcription. Besides Smad-‐mediated transcription, TGFβ is also able to mediate transcription through other pathways such as TAK1 (TGFβ-‐activated kinase 1) (Dobacwewski, 2011). This is called non-‐canonical signalling and involves activation of the TRβII receptor.
Much research is focused on the canonical signalling pathways; however, there is mounting evidence that non-‐canonical signalling plays a more central role: research showed that inhibition of the non-‐canonical pathways led to reduced fibrosis and remodeling in mice in response to cardiac overload (Leask, 2010). An important protein involved in non-‐cannonical TGFβ1 signalling is TAK1 (Zhang, 2000). It has been shown that TAK1-‐overexpression in mice causes cardiac hypertrophy. In addition, dominant negative TAK1 inhibits TGFβ-‐induced hypertrophic events in mouse cardiomyocytes and fibroblasts (Ono, 2003).
4 – Fibroblasts and mechanical stress
4.1 Mechanical stress and TGF-‐β1
TGF-‐β1 in the heart is secreted in an inactive form, a latent complex. It is unable to associate with its receptors. Activation of a relatively small amount TGF-‐β1 is sufficient to induce a maximal cellular response and is dependent on stiffness of the matrix (Annes, 2003).
The modulation of TGF-‐β1 bioactivity is a poorly understood progress in which mechanical stress plays a role, possibly in the following manner: fibroblasts secrete TGF-‐β1 bound to latency-‐associated protein (LAP). This complex can attach to the TGF-‐β1 binding protein-‐1 (LTBP-‐1), forming the Large Latent Complex (LLC). LTBP-‐1 can attach to molecules in the ECM, forming a reservoir of latent TGF-‐β1 (Annes, 2004). The LAP-‐part of this complex provides binding parts (‘RGD’ in figure 2) for myofibroblast integrins, including αvβ5. This integrin is connected to the cytoskeleton. When stress is applied to αvβ5 by stretching the ECM, the latent TGF-‐β1-‐complex shifts, releasing active TGF-‐β1 (Hinz, 2010; Wipff and Hinz, 2008).
However, this mechanism, which is shown in figure 2, has not been fully elucidated yet.
Noteworthy, stiff ECM seems to be required for this process, since the ECM molecules attached to the LLC are only unable to move along with the movement of the LLC under these conditions.
This raises a chicken-‐egg-‐like question: How can TGF-‐β1 cause an increase in stiffness of the ECM, if increased stiffness of the ECM is required to activate TGF-‐β1 from its latent state? An answer may lie in the process of increased early cross-‐linking of collagen, catalysed by Lysyl Oxidase (LOX) enzymes after injury. LOX is showed to be upregulated after injury (López, 2010).
The process of cross-‐linking stiffens the ECM to a sufficient degree to cause elevated activation of the latent TGF-‐β1 (Penelope, 2007). Another possibility is a stiffness elevation induced by integrin-‐mediated contraction of myofibroblasts only. The contraction of multiple myofibroblasts may cause sufficient stiffness of the ECM (stress to the LLC) to release TGF-‐β1
from its latent state (Wipff, 2007). However, this purely mechanical process is only shown by cultured myofibroblasts yet. In summary, stiff ECM is activates latent TGF-‐β1.
4.2 Mechanical stress induces SMA protein expression and control of SMA fibers
The onset of cardiac fibroblast differentiation to myofibroblasts after injury takes a few days.
This is remarkable, since early wound healing attracts TGF-‐β1-‐generating inflammatory cells like macrophages (Blakytny, 2003). Why is it that TGF-‐β1 is not able to induce a differentiation response in cardiac fibroblasts directly after the injury, despite the widely available TGF-‐β1? An
answer may lie in research performed by Goffin and colleagues, whereby was shown that myofibroblasts differentiation is supressed when growing on soft ECM (Goffin, 2006). This suggests a TGFβ1-‐indepent link between mechanical stress and myofibroblast differentiation (Hinz, 2010), a thought strengthened by research in integrin adhesions by Wang and colleagues, who showed that applied force to those adhesions is sufficient to cause an upregulation of SMA promotor activity (Wang, 2002). The mechanism behind this mechanical induction may involve the AT1 receptor, as discussed in the next section. However, research with TGF-‐β1 signaling inhibitors has shown that the onset of SMA expression is not possible in total absence of TGF-‐β1 (Hinz, 2001). Therefore, both mechanical stress and the presence of TGF-‐β1 are necessary for myofibroblast differentiation.
An additional level of mechanical control of myofibroblasts is the relocalization of SMA-‐
fibers in response to a shift of environmental stiffness. Goffin investigated this in the following manner: he moved myofibroblasts cultured on stiff plastic to a soft sillicone-‐based substrate.
This caused dislocation of the SMA and formation of proto-‐myofibroblast associated stress fibers (Goffin, 2006). Thus, not only myofibroblast differentiation is suppressed by soft growth medium, it causes also a disturbance in the formation of SMA of differentiated myofibroblasts.
4.3 Mechanical stress and the AT1-‐receptor
Yasuda and colleagues demonstrated that the AT1 receptor could be activated by mechanical stress, independently of ATII, by the use of an inverse AT1 receptor agonist (Yasuda, 2008). This means an extra mechanism exists by which increased stiffness of the myocard could lead to additional activation of myofibroblasts.
Fig. 2 -‐ The large latent complex (LLC) is bound to the ECM via TGF-‐β1 binding protein-‐1 (LTBP-‐1). TGF-‐β1 is entrapped in this complex and hereby, not able to bind the myofibroblast TGF-‐β1-‐
receptor. The LLC is connected to myofibroblast by myofibroblast integrins, such as αvβ5. Cell contraction of myofibroblasts, or stretching of the ECM (not shown in this figure) leads to disruption of the LLC, whereby TGF-‐β1 is released in its active form (Hinz, 2010).
5 – Fibroblast – cardiomyocyte interactions
5.1 Interactions between fibroblasts and cardiomyocytes
Fibroblasts and myofibroblasts are thought to modulate the structure and function of CM by the release of soluble mediators, indirectly via the ECM and by direct cell-‐cell contact (Kakkar, 2010). These interactions will be briefly discussed in the following sections.
5.2 Interactions via soluble mediators
The precise nature of cardiomyocyte-‐fibroblast interaction through soluble mediators remains unknown (Cardlegde, 2015), mainly because of the complex interaction between soluble mediators between both cell types, since both cells can release and respond to a variety of substances. Nevertheless, the main (most studied) soluble mediators in this process will be discussed.
TGFβ1 -‐ Probably the most important soluble mediator secreted by fibroblasts is TGFβ1. However, since the functions of this substance are already discussed in previous sections, it is only briefly mentioned in this section.
FGF2 and FGF16 -‐ Two soluble mediator signals, dependent on TGF-‐β1 signalling, that play major roles in the communication between CM and cardiac fibroblasts are Fibroblast Growth Factor 2 and 16 (FGF2 and FGF16, respectively). Cardiac fibroblasts are the main source of FGF2 in the heart (Santiago, 2010). FGF2 is an important factor in the process of cardiac hypertrophy (Siyun, 2009). FGF16 is thought to possess the opposed function (Matsumoto, 2013). FGF2 and FGF16 compete for the same binding sites on their primary receptor, the Fibroblast Growth Factor Receptor (FGFR1c) (Lu, 2008). A schematic overview of cardiomyocyte-‐fibroblast communication is shown in figure 3. Cardiac fibroblasts are able to secrete FGF2 in response to TGFβ1, which induces TGFβ1 secretion in CM. Increased TGFβ1-‐levels cause hypertrophy in CM, are associated with the differentiation of fibroblasts into myofibroblasts and increase stiffness of the ECM, as discussed in earlier sections. FGF2 causes an upregulation of the production of FGF16 in CM. FGF16 has an inhibiting function on the TGFβ1 production in both CM and cardiac fibroblasts. Under pathological conditions, the availability of TGFβ1 increases, which causes hypertrophy of CM. However the pathophysiological role of FGF16 in humans has not been analyzed yet, research has been performed on the role of FGF16 in mice with angiotensin II-‐
induced hypertrophy (Matsumoto, 2013). Macrophages play a major role in inducing this hypertrophy by producing ATII, as described in section 3.1. ATII causes induced TGFβ1-‐
production in fibroblasts, while FGF16 causes the opposite effect. Therefore, FGF16 might be a useful approach in counteracting adverse remodeling processes in the heart.
Notch signalling -‐ A factor in embryonic cardiac signalling is Notch-‐signalling, which plays a role in cell fate regulation. This form of signalling is also activated in the injured heart but absent in the adult heart in mice (Gude, 2008). Evidence for Notch-‐signalling between fibroblasts and CM is present, especially between the Notch ligand Jagged-‐1 on the surface of CM and its Notch1 receptor on the surface of cardiac fibroblasts. The Notch1 receptor is activated during pressure overload in mice and is associated with inducing CM hypertrophy, although the precise mechanisms of this interaction are not understood yet (Fujiu, 2014). Inhibition of Notch-‐
signalling in mice led to worsened remodeling and hypertrophic response after injury (Croquelois, 2008). This conclusion is supported by a study conducted by Nemir and colleagues, who found that CTGF expression was downregulated by Jagged-‐1-‐induced Notch signalling in mice (Nemir, 2014). Additionally, Notch-‐signalling negatively regulated cardiac fibroblast to myofibroblast differentiation (Fan, 2011).
5.3 Indirect cardiomyocyte-‐fibroblast interactions via the ECM
Ieda and colleagues have conducted research on the differences in proliferative abilities of embryonic fibroblasts and adult fibroblasts on CM in mice. They found a proliferative effect of embryonic fibroblasts on CM, but a hypertrophic effect of adult fibroblasts on cardiomyoctes.
Additionally, adult fibroblasts caused a proliferative effect on embryonic CM, but to a lesser degree than embryonic fibroblasts. The mechanism of proliferation was regulated via the ECM components fibronectin and collagen, and via heparin-‐binding EGF-‐like growth factor (HBEGF).
HBEGF is a mitogen (a cell-‐division triggering substance) secreted by cardiac fibroblasts. In this research it has been shown that fibronectin and collagen promoted HBEGF-‐induced CM mitotic activity through integrin signalling. α5β1 integrin (fibronectin-‐specific receptor) is a receptor that directly associates with the HBEGF-‐receptor, which is necessary for optimal activation of growth signalling. Embryonic CM express more collagen-‐ and fibronectin-‐specific receptors, while adult CM express more laminin-‐ receptors. This suggests that the CM expression of integrins might be responsible for CM proliferation on fibronectin and collagen (Ieda, 2009).
Another indirect interaction between the ECM and CM under pathological conditions is in the form of increased deposition of collagen fibers by myofibroblasts. These fibers are shown to entrap CM, hereby reducing their ability to contract periodically. By making it unable for CM to contract, these fibers decrease the workload of CM, which causes atrophy. In other words, the matrix is too stiff to allow proper contration of CM. It is thought that this is one of the reasons why scar formation contributes to the progressive nature of heart failure (Fidziańska, 2010;
Weber, 2013).
Summarized, integrin signalling is required for CM to form ECM-‐substance recognizing receptors, of which collagen and fibronectin are associated with embryonic cardiomyocyte proliferation and laminin is associated with adult cardiomyocyte hypertrophy. Another form of indirect ECM-‐cardiomyocyte interaction is the entrapment of CM in excessive collagen fibres during pathological conditions, leading to atrophy of CM.
Fig. 3 -‐ A schematic overview of the communication between CM and cardiac fibroblasts through some main soluble mediators. Cardiac fibroblasts produce FGF2 in response to hypertrophied conditions. FGF2 activates the FGFR1c receptor on CM and, via an autocrine manner, on cardiac fibroblasts. This stimulation causes an upregulation of TGF-‐β1 production and secretion, responsible for cardiomyocyte hypertrophy. Besides the upregulation of TGF-‐β1, stimulation of the FGFR1c receptor causes an increase in FGF16 expression and secretion by CM. FGF16 competes with FGF2 for the same receptor, hereby inhibiting the action of FGF2 and fulfilling a role in a negative-‐feedback loop (Fujiu, 2014).
Fig. 4 – A schematic view of a gap junction between a cardiac fibroblast and a cardiac myocyte. Connexin 43 and 45 are involved in the formation of the gap junction.
Cardiac fibroblasts are able to sense the stiffness of the surrounding matrix via β1-‐
integrin signalling. The exact pathway of this signalling is not elucidated yet.
However, is is thought that β1-‐integrin signalling plays a role in the regulation of Na2+-‐ion channels. By altering the intra-‐
cellular ion concentration of cardiac fibroblasts, and thereby, via gap junctions, the ion concentration of cardiac myocytes, the automaticity of cardiac myocytes can be changed in response to mechanical stress (Kakkar, 2010).
5.4 Interactions via cell-‐cell contact
The electrical conductance system of the heart relies on the syncytium-‐forming gap junctions between CM. Recent research indicates that communication between cardiac fibroblasts and myocytes may occur in a similar manner (Kakkar, 2010). Gap junction proteins associated with connections between fibroblasts and myocytes are connexin-‐43 and -‐45. Fibroblasts may play a role in the process of cardiomyocyte depolarization (Goshima, 2004). This process is shown in figure 4. Cardiac fibroblasts form gap junction with cardiac myocytes via connexin 43 and 45.
Additionally, cardiac fibroblasts are able to sense the density or stiffness of their surroundings by β1-‐integrins. These receptors influence the Na2+-‐ion channels of fibroblasts by a mechanism yet to be uncovered. The altered intracellular ion concentrations of cardiac fibroblasts influence the intracellular concentrations of cardiac myocytes through gap junction transport (Rook, 1989). By this mechanism, the automaticity of cardiac myocytes is influenced by the mechanical properties of the surroundings of the cardiac fibroblasts. However, the implications of this mechanism to CM are not elucidated yet (Kakkar, 2010).
In vitro studies on myocyte-‐fibroblast interactions showed another way by which myocyte-‐
fibroblast cell-‐cell interaction can alter cardiomyocyte function. Namely, by causing a state called ‘hibernation’ in CM in response to hypoperfusion in order to prevent cell death. This state is associated with sarcomere depletion and loss of cytoplasmic structure of CM. It has been suggested that this hibernation mimics a more embryonic-‐like state by causing dedifferentiation of CM (Vanoverschelde, 1997). GATA4 is a zinc-‐finger transcription signal that plays a role in this process. During embryogenesis, GATA4 is a critical regulator of the cardiac differentiation-‐
specific gene program. Zaglia and colleagues showed that rat cardiac fibroblast still express GATA4 after birth, in contrast to other important embryonic transcription signals like Isl1 and Nkx2.5. In this research it was also shown that GATA4-‐induced dedifferentiation of CM is associated with cell-‐cycle re-‐entry (Zaglia, 2009). However, since the lack of evidence in in vivo models yet, more research has to be performed on this subject.
6 -‐ Recent developments in research on cardiac regeneration
In order to fit in the role of cardiac fibroblasts in cardiac regeneration, it is important to get an overview of recent developments in cardiac regeneration eventhough cardiac fibroblasts are not the main subject of these researchers.
6.1 The proliferative window of CM
Porrello and colleagues performed much research to cardiomyocyte regeneration and showed in 2011 that hearts of 1-‐day-‐old mice (P1) were able to regenerate to their normal anatomy and function after a ventricular resection. Yet, mice lost this ability gradually in their first week after birth (Porrello, 2011). This experiment was repeated succesfully in rats (Zogbi, 2014). However, Darehzereski and colleagues showed that cardiomyocyt proliferation does not necessarily increase during these conditions compared to normal neonatal rodents, but that the adequate growth of myocardial tissue after injury was sufficient to induce a recovery from an MI-‐
simulating injury (Darehzereski, 2015). These researches show, besides the fact that sufficient proliferation of CM occurs after injury; no chronic scar tissue is formed. This implicates a different role for fibroblasts during neonatal compared to adult development in rodents.
Additionally, although not as extreme, cardiomyoyte regeneration in humans is also significantly elevated in young humans compared to adults (Mollova, 2013; Bergmann, 2009), implicating that this different role of fibroblasts in cardiomyocyte proliferation, in some way, could be extended to humans. These findings are consistent with research performed by Ieda and colleagues, who showed that embryonic rat fibroblasts stimulate embryonic and, to a lesser extent, adult rat CM to proliferate, and that adult fibroblasts stimulate hypertrophy in adult CM.
Differences in gene expression of HBEGF and β1-‐integrin receptors played a role in this shift of CM function (Ieda, 2009).
6.2 Neuregulin signalling
Additionally, recent research on Neuregulin (Nrg)–signalling in the heart by Polizzotti and colleagues provided promising results. Nrgs are a family of growth factors of which Nrg-‐1 and its tyrosine kinase receptors ErbB2 and ErbB4 are the most common in the myocardium. Nrg-‐1 is secreted by, and located on the surface of endothelial cells and fibroblasts. The tyrosine kinase receptors are located on the surface of CM (Parodi and Kühn, 2014). Administration of Nrg-‐1 to mice after cardiac injury caused cell cycle re-‐entry of CM, even when the ‘proliferative window’
of seven days was expired. This proliferative effect of Nrgs is succesfully repeated in humans, however, cell cycle re-‐entry of CM was not significantly elevated in patients older than 6 months.
Yet, this limited time frame of cell cycle re-‐entry could be explained by the fact that these patients suffered from cardiac diseases, in contrast to the investigated mice. The presence of such diseases may drive CM out of the cell cycle into a ‘quiescent state’ (Polizzotti, 2015).
Summarized, local administration of Nrg might be a useful therapeutic approach to induce CM proliferation.
Fig. 5 – The interactions between myofibroblasts, macrophages, CM and the ECM during pathological conditions.
Necrosis-‐associated proteins attract macrophages to the site of injury. Macrophages produce ATII, which activates myofibroblasts and macrofages to produce TGF-‐β1. Myofibroblasts differentiate from fibroblasts in pathological conditions and are also able to produce ATII. TGF-‐β1 signalling causes myofibroblasts to upregulate the production of matrix components and to inhibite MMPs. This leads to increased matrix formation and increased stiffness of the ECM. Increased stiffness of the ECM leads to activation of latent TGF-‐β1, which helps to maintain the activation and differentiation of myofibroblasts. Components of the ECM are able to act on integrin-‐receptors on CM, causing hypertrophy and on myofibroblasts, causing altered ion concentrations. This shift in myofibroblast intracellular ion concentration spreads to CM, facilitated by gap junctions between the two cell-‐types. Additionally, the increased amount of collagen entraps CM, causing atrophy. The increased stiffness of the ECM leads to angiotensin-‐independent activation of AT1-‐receptors, which also leads to activation of myofibroblasts (created with Lucidchard).
7 – Discussion
7.1 Summary of interactions between cell types
The interactions between myofibroblasts, CM, macrophages and the ECM under pathological conditions are summarized in figure 5. This figure shows the basic principles of the discussed interactions between different cell types in pathological conditions. Most interactions eventually lead to impaired cardiomyocyte function or decreased ability to proliferate.
Necrotis-‐associated proteins attract macrophages to the site of injury. Macrophages produce ATII and are able to activate their membrane-‐bound AT1 receptor, in an autocrine way.
Activation of this receptor leads to increased TGF-‐β1 expression.
Myofibroblasts differentiate from fibroblasts during pathological conditions and also express the AT1-‐receptor and are able to produce ATII. Therefore, besides autocrine ATII signalling, cross-‐
activation of the AT1 receptor take place between these cell types. TGF-‐β1 signalling in CM leads to an increased production of ECM components like several collagen types, fibronectin and laminin. Additionally, the production of MMPs is inhibited. These changes lead to an increased formation of ECM-‐proteins, which leads to an increased stiffness of the ECM, whereby latent TGFβ1 is activated. An increased stiffness of the ECM also leads to ATII-‐independent activation of the AT1-‐receptor on the cell surface of myofibroblasts, hereby stimulation the production of TGFβ1 in an additional manner. Components of the stiffened ECM also activate integrin receptors on the surface of myofibroblasts and CM. In CM, laminin activates integrin receptors, which causes hypertrophy of CM. Integrin receptors on the surface of cardiomyocyte are able to alter the ion flow by regulating ion channels in response to mechanical stress. This leads to altered intracellular ion concentrations in myofibroblasts, which spreads to CM by gap junctions, hereby altering the membrane potential. Finally, collagen fibers can entrap CM, which causes atrophy.
Not shown in figure 6, are some other important signalling factors including FGF2-‐FGF16 signalling, which plays a role in the regulation of TGFβ1 expression of cardiac fibroblasts and CM and Notch signalling, which plays a role in CM hypertrophy in pathological conditions.
7.2 Future perspectives
This general picture shows interactions involved in pathological cardiac remodeling, and could help to elucidate processes that inverse the unfavorable outcomes associated with this remodeling. Together with recent developments on cardiac regeneration described in section 5, some interesting approaches to induce regeneration can be deducted from this review and will be briefly summed up below.
FGF 16 -‐ As described in section 5.2, FGF16 inhibits TGFβ1 production in fibroblasts and CM, which leads to a reduced production of ECM-‐components. Therefore, finding a way to increase FGF16 production in cardiac tissue after injury might diminish the formation of scar tissue.
Additionally, research has shown that FGF16 might play a role in CM proliferation in mice (Lavine, 2005). However more research has to be performed on the role of FGF16 in humans, this proliferative effect could provide an additional beneficial factor for FGF16.
Integrin signalling -‐ Integrin signalling plays, through HBEGF, a role in CM proliferation and hypertrophy. Fibronectin-‐specific integrin signalling in embryonic rats induces CM proliferation, while the diminished fibronectin-‐specific integrin (α5β1) signalling in adult rats is associated with CM hypertrophy (Ieda, 2009). More research on the differences between adult and embryonic integrin signalling in humans could provide additional insight in CM proliferation.
Neuregulin -‐ Administration of Nrg stimulates CM to re-‐entry the cell cycle. However, since pathological conditions in patients made CM impassible to Nrg-‐therapy by inducing a quiescent state, more research has to be performed on the exact mechanisms (Polizzotti, 2015).
‘Hibernation’ of CM -‐ Finally, the hibernation state or ‘dedifferentiation’ in reaction to
hypoperfusion of the heart as shown by Vanoverschelde and colleagues could also be interesting since differentiation is associated with loss of proliferative abilities (Vanoverschelde, 1997). If this type of dedifferentiation could induce increased proliferation of CM, this research may provide a starting point for the development of a new therapy. However, therefore, more research to the abilities and gene expressions of hibernated CM has to be performed.
Notch-‐signalling – Fan and colleagues found that Notch-‐signalling inhibits the differentiation from cardiac fibroblasts to myofibroblasts (Fan, 2011). Additionally, Notch signalling is negatively correlated to the formation of CTGF and is involved in regulating the hypertrophic response of CM after injury (Nemir, 2014). Therefore, Notch-‐signalling is worth investigating.