Thrombospondin-4, tumour necrosis factor-like weak inducer of apoptosis (TWEAK) and its receptor Fn14: Novel extracellular matrix modulating factors in cardiac remodelling

Abstract

Cardiac remodelling is defined as changes in the size, shape, and function of the heart, which are most commonly caused by hypertension-induced left ventricular hypertrophy and myocardial infarction. Both neurohumoral and inflammatory factors have critical roles in the regulation of cardiac remodelling. A characteristic feature of cardiac remodelling is modification of the extracellular matrix (ECM), often manifested by fibrosis, a process that has vital consequences for the structure and function of the myocardium. In addition to established modulators of the ECM, the matricellular protein thrombospondin-4 (TSP-4) as well as the tumour necrosis factor-like weak inducer of apoptosis (TWEAK) and its receptor Fn14 has been recently shown to modulate cardiac ECM. TSP-4 null mice develop pronounced cardiac hypertrophy and fibrosis with defects in collagen maturation in response to pressure overload. TWEAK and Fn14 belong to the tumour necrosis factor superfamily of proinflammatory cytokines. Recently it was shown that elevated levels of circulating TWEAK via Fn14 critically affect the cardiac ECM, characterized by increasing fibrosis and cardiomyocyte hypertrophy in mice.

Here we review the literature concerning the role of matricellular proteins and inflammation in cardiac ECM remodelling, with a special focus on TSP-4, TWEAK, and its receptor Fn14.

Key words::

• Cardiac remodelling
• extracellular matrix protein
• Fn14
• left ventricular hypertrophy
• myocardial infarction
• thrombospondin
• TWEAK

Abbreviations
ACTA2	=	α-smooth muscle actin 2
EC	=	endothelial cell
ECM	=	extracellular matrix
HF	=	heart failure
IEG	=	immediate early gene
KO	=	knock-out
LV	=	left ventricular
MI	=	myocardial infarction
MMP	=	matrix metalloproteinase
SMC	=	smooth muscle cell
SPARC	=	secreted protein acidic and rich in cysteine
STEMI	=	ST elevation myocardial infarction
TAC	=	transverse aortic constriction
TGF	=	transforming growth factor
TIMP	=	tissue inhibitor of matrix metalloproteinase
TNF	=	tumour necrosis factor
TSP	=	thrombospondin
TWEAK	=	tumour necrosis factor-like weak inducer of apoptosis

Key messages

• The extracellular matrix is both the target of modifications as well as an active regulator of the cardiac remodelling process, and thus contributes significantly to the development of heart failure.

• Inflammation and proinflammatory cytokines have a critical role in regulating cardiac structure and function during cardiac remodelling.

• Matricellular proteins modulate cellular function and behaviour by regulating cell–cell and cell–matrix interactions in cardiac remodelling.

• Cardiac remodelling-related inflammatory modulators and matricellular proteins are potential targets for therapeutic interventions.

Introduction

Cardiac remodelling is an adaptive process in which the heart changes its size, shape, and function (1). The aim of remodelling is to maintain adequate cardiac output despite an altered cardiac load and is most commonly due to continuous cardiac overload. Conditions that generate cardiac overload include chronic hypertension, volume overload due to mitral valve regurgitation, or the loss of viable myocardium after myocardial infarction (MI) (Figure 1). Cardiac remodelling also occurs under physiological conditions, such as during pregnancy or in response to exercise, but characteristically only the pathological remodelling due to intrinsic or acquired heart diseases can lead to heart failure (HF) if left untreated (1).

Figure 1. Cardiac remodelling plays a crucial role during development of heart failure. Cardiac remodelling occurs most commonly as a result of either prolonged hypertension or myocardial infarction and also in physiological conditions, such as during pregnancy.

Common features in the process of pathological cardiac remodelling regardless of the aetiology are molecular, cellular, and interstitial changes of the myocardium such as myocyte hypertrophy, disorganization of the sarcomere, cell loss, and deposition of extracellular matrix (ECM) (Figure 2). Cellular and molecular changes of the myocyte include e.g. changes in gene expression patterns such as activation of immediate early genes (IEGs) (c-jun, c-fos) followed by the reactivation of a fetal gene programme (atrial natriuretic peptide (ANP), β-myosin heavy chain), a shift from fatty acid oxidation to using carbohydrates as energy substrates, and alterations in calcium handling. There is also extensive myofibrillar remodelling that leads to altered contractile properties of the myocyte and subsequently to altered contractile function of the whole myocardium (Figure 3) (1,2). The response of cardiac fibroblasts to stress or injury includes alterations in proliferation, migration, ECM metabolism, and release of autocrine/paracrine mediators (3). A key event in fibroblast activation is their transformation into myofibroblasts, which have smooth muscle cell (SMC)-like features and exhibit increased migration, proliferation, and secretion (Figure 3) (4). Myofibroblasts have a contractile apparatus and express proteins distinct from those of undifferentiated fibroblasts, e.g. α-smooth muscle actin (ACTA2), α5β1 and αvβ3 integrins, tensin, and focal adhesion kinase (5). Myofibroblasts are not normally found in the healthy heart muscle (6) but can be detected in hypertrophic and fibrotic hearts and in the scar tissue following MI (4,7). In response to specific hormones, growth factors, and cytokines, the myofibroblasts become highly proliferative. They also secrete ECM proteins, predominantly type I and III collagens, and matrix metalloproteinases (MMPs) as well as bioactive molecules and therefore are a major contributor to the cardiac remodelling process. These fibroblast-stimulating factors are released from inflammatory cells, cardiac myocytes, and fibroblasts themselves and include angiotensin II (Ang II), transforming growth factor (TGF)-β1, and endothelin-1 (ET-1). Mechanical stretch can also directly or indirectly modulate fibroblast phenotype. Fibroblasts are capable of secreting a wide array of bioactive molecules themselves, such as TGF-β1, interleukin (IL)-1β, Ang II, as well as ANP and B-type natriuretic peptide (BNP). The initiating stimulus determines which factors are secreted, whereas the response depends on the receptors expressed in the target cells. The ECM produced by myofibroblasts also differs from that of undifferentiated fibroblasts; modification of the balance between MMPs and their inhibitors causes fibrosis (4,7). In addition to cardiac cells, the ECM plays an important role in cardiac remodelling, as it is both the target of modifications and an active regulator of the process. ECM modifications have severe consequences on the structure and function of the myocardium and contribute significantly to the development of HF (8,9).

Figure 2. Schematic representation of cellular and molecular changes in the heart during hypertensive hypertrophy and after myocardial infarction. In the hypertensive heart, the hypertrophy of cardiomyocytes and the transition of fibroblasts to myofibroblasts are accompanied by an increase in the amount of ECM, which leads to fibrosis. Inflammation, cell death, and scarring are integral parts of post-MI remodelling associated with subsequent hypertrophy and fibrosis of the non-infarcted area.

Figure 3. Subcellular targets in myocytes and fibroblasts during cardiac remodelling.

Recently, non-structural ECM proteins such as matricellular proteins have gained increasing attention in the cardiac remodelling process. They modulate cellular function and behaviour by regulating cell–cell and cell–matrix interactions, and many of these proteins appear to affect ECM composition (10). Inflammation and proinflammatory cytokines also have critical roles in regulating cardiac remodelling and in the development of HF (11). This interaction between inflammation and the ECM is reciprocal, in that while inflammation modulates the structure of the ECM, e.g. by affecting matrix turn-over, the ECM also regulates the inflammatory response, for instance by releasing fragments that can affect cytokine synthesis (12). There are also connections between inflammation and matricellular proteins, some of which have been shown to regulate inflammation (12). The following review of the literature concerning the role of matricellular proteins and inflammation in cardiac ECM remodelling will focus on novel ECM-modulating factors, i.e. the matricellular protein thrombospondin (TSP)-4 and the proinflammatory cytokine tumour necrosis factor-like weak inducer of apoptosis (TWEAK) and its receptor Fn14.

Cardiac ECM remodelling

In cardiac ECM, structural proteins such as collagen and elastin provide support for the cells and participate in transmission of the mechanical force generated by myocytes (13). Adhesive proteins like fibronectin and laminin as well as proteoglycans modulate cardiac cell growth, proliferation, and migration as well as regulate growth factor activity and collagen network organization (14). In addition, a wide array of bioactive molecules participates in the complex signalling between the cells and the ECM (15).

The balance between ECM synthesis and degradation determines the amount of cardiac ECM, which is produced mainly by (myo)fibroblasts but additionally by myocytes, endothelial cells (ECs), and SMCs (16). The synthesis of cardiac ECM is influenced by a variety of mechanical and biochemical stimuli such as stretch, circulating hormones, as well as autocrine and paracrine factors generated in the myocardium (4,15). Major effectors of ECM synthesis are Ang II and TGF-β that in addition to stimulating its production can also inhibit its degradation (4). The ECM is degraded mainly by MMPs. MMPs are secreted by all the major cardiac cell types as inactive precursors, which can be activated by proteases or membrane-bound MMPs (17). MMP expression is increased during cardiac remodelling in response to cytokine, vasoactive peptide, and growth factor activation, e.g. to tumour necrosis factor (TNF)-α, one of the most important cytokines affecting ECM composition (4,17). MMPs not only regulate the degradation of the ECM but can also stimulate matrix synthesis by releasing matrix-bound growth factors and other profibrotic molecules from the ECM (18). MMPs are inhibited by tissue inhibitors of MMPs (TIMPs), and ultimately it is the balance between MMPs and TIMPs together with ECM synthesis that determines the rate of matrix turn-over (17). The rate of ECM turn-over in the heart is normally low, but in pathological situations both the synthesis and degradation increase (8).

The most common pathological situations that result in cardiac remodelling are hypertensive hypertrophy and MI (Figure 2). In the former the net effect is an increase in the amount of ECM, which leads to fibrosis, a process characteristic of pathological remodelling since it is absent in physiological hypertrophy (19). Fibrosis is detrimental because the accumulation of excessive ECM impairs diastolic function, diminishes coronary flow reserve, and predisposes to arrhythmias (19). In the reverse situation, when cardiac ECM degradation exceeds its synthesis, such as after MI, left ventricular (LV) dilatation, sphericalization, and increased compliance ensue (20). In addition to thinning and scarring of the infarcted wall, post-MI remodelling causes complex time- and region-dependent changes including also subsequent hypertrophy and fibrosis of the non-infarcted area (21).

In addition to the quantity of the ECM, its quality also affects the physical properties of the myocardium, in particular its compliance. The size and alignment of the collagen fibres as well as altered ratios of diverse collagen types may affect myocardial stiffness, and especially the post-translational modifications of collagen fibrils contribute to the compliance of the myocardium (22,23). An important type of collagen modification is the formation of intermolecular cross-links that make the collagen fibres less soluble and resistant to degradation. Consequently, myocardial collagen cross-linking stiffens the myocardium and is accelerated in situations with increased collagen deposition, such as in LV hypertrophy (23,24). In contrast, poorly cross-linked collagen is more rapidly degraded, which predisposes to LV dilatation (22) as reported in association with dilated myocardium in HF (25,26).

Inflammation and proinflammatory cytokines in cardiac ECM remodelling

As a classical biological response to injury and a mechanism for adaptation and tissue repair (11), inflammation is an integral part of post-MI remodelling (27). A transient myocardial inflammatory reaction with release of inflammatory mediators can also be seen in response to pressure overload in vivo (28,29). A well characterized mechanism for the activation of the inflammatory response is the cell destruction after MI that exposes antigens capable of triggering the immune response, but inflammation can also be provoked by myocardial cytokine release in response to haemodynamic stress (30). Since the initial observation that circulating levels of a classical proinflammatory cytokine TNF-α are increased in HF patients (31), a number of studies have established a critical role for proinflammatory cytokines in the development of HF as well as in regulating cardiac structure and function in other forms of cardiac remodelling. This has led to a theory called the ‘cytokine hypothesis’ which claims that HF progresses because cytokine cascades that are activated following myocardial injury exert deleterious effects on the heart (32).

Cytokines are a group of relatively small molecular weight proteins that are secreted by a broad variety of cells in response to different types of stimuli such as environmental stress. They tend to act over short distances in a juxtacrine, paracrine, or autocrine manner, but in some cases they can also have endocrine effects (32). In the heart they are rapidly up-regulated in response to myocardial injury and act as part of the intrinsic cardiac stress response system. This up-regulation has been shown to be beneficial when short-term and limited but has deleterious effects when sustained (33); e.g. TNF-α modulates fibroblast function by stimulating MMP expression and decreasing collagen synthesis, although long-term stimulation may promote accumulation of collagen and fibrosis (4,33).

Cardiac injury activates immunomodulatory mechanisms initiating an inflammatory reaction; e.g. the complement cascade up-regulates chemokine and cytokine synthesis in the infarcted heart (34). Complement activation may play an important role in mediating neutrophil and monocyte recruitment in the injured myocardium (35). Furthermore, cytokines like TNF-α promote fibroblast proliferation and transformation into myofibroblasts (13). In addition to cardiac fibroblasts, myofibroblasts can also be derived from monocytes, bone-marrow-derived cells, and ECs (4).

Matricellular proteins in cardiac ECM remodelling

Matricellular proteins are non-structural ECM proteins that modulate cellular function and behaviour by regulating cell–cell and cell–matrix interactions. Characteristically matricellular proteins are minimally expressed in normal adult tissue but are up-regulated during development and in tissue injury (10). The matricellular proteins involved in cardiac remodelling include e.g. secreted protein acidic and rich in cysteine (SPARC), tenascin-C, osteopontin, periostin, and TSPs. The importance of these proteins in cardiac remodelling has been demonstrated in knock-out (KO) mice which have a mild cardiovascular phenotype but have increased adverse remodelling and decreased survival in response to MI or pressure overload (36,37). Based on these studies, many of the matricellular proteins seem to have a role in myocardial collagen deposition since e.g. osteopontin, SPARC, and periostin deficiency result in reduced accumulation or maturation of collagen after MI (12).

Thrombospondins

TSPs are a family of five multimeric multidomain matricellular proteins (Figure 4). They are capable of interacting with a large number of proteins and proteoglycans which enables their various actions and tissue and cell type-specific functions depending on the bioavailability of TSP receptors and binding partners in the given environment (38). Additionally, despite the structural relatedness, their functions have diverged due to their different transcriptional regulation and functional capabilities (38,39). In heart, increased expression of TSPs-1–4 has been shown during cardiac remodelling (37) (Table I).

Figure 4. Schematic structures of TSP proteins, TWEAK, and Fn14. A: TSP-1 and TSP-2; B: TSP-3 and TSP-4; C: TSP-5; D: TWEAK; and E: Fn14 protein. OM = oligomerization domain; VWC = von Willebrand factor C type domain; TSR = thrombospondin repeats; EGF = EGF-like repeats; Cyt = cytoplasmic domain; TM = transmembrane domain; EC = extracellular domain; RB = receptor binding domain; Sig = signal peptide. Arrow indicates a cleavage site. Modified from (88–90).

Table I. Expression and localization of TSPs in adult heart in vivo (modified from (37)).

	Normal, adult heart	Cardiac remodelling
		Pressure overload	MI	Localization
TSP-1	mRNA: + (53,56,91) protein: + (53)	↑ (53,92)	↑ (52,56)	cardiac myocytes and intimal SMCs in human cardiac allografts (55) ECs and mononuclear cells after ischaemia–reperfusion (40) cardiac fibroblasts/myofibroblasts and mononuclear cells after MI (56) ECs in diabetic cardiomyopathy (93)
TSP-2	mRNA: ND protein: + (44,94)	↑ (in hypertrophied hearts prone to HF) (43)	ND	ND
TSP-3	mRNA: + (91) protein: + (44)	ND	ND	ND
TSP-4	mRNA: + (53,91) protein: + (53)	↑(47,53)	↑ (53)	ECs in hypertrophied heart (53)
TSP-5/COMP	mRNA: + (58) protein: ND	ND	ND	ND

+= expression; ↑ = increased expression; ND = not defined.

TSP-1 is a major activator of TGF-β, and a suggested function of TSP-1 in cardiac remodelling is limitation of the inflammatory reaction after MI (40). TSP-2, on the other hand, has been shown to regulate myocardial matrix integrity during cardiac remodelling and ageing (41). The absence of TSP-2 results in cardiac rupture after MI (42) or in response to pressure overload (43) probably through increased MMP activity, which also contributes to the age-dependent dilated cardiomyopathy in TSP-2 KO mice (41). In addition, TSP-2 has been shown to inhibit the ageing-associated local myocardial inflammatory response (41). There is hardly any evidence regarding TSP-3 expression in the heart. In a DNA microarray study of Schroen et al., they show that TSP-3 mRNA levels are increased in hearts from hypertensive homozygous renin over-expressing (Ren-2) rats that had progressed to early HF compared to Ren-2 rats that had remained compensated (43). In another study, very low expression levels of TSP-3 in human heart were reported (44). In conclusion, whether TSP-3 has a role with respect to fibrosis or myocardial remodelling remains to be studied.

TSP-4 gene expression has been shown to be increased in DNA microarray studies of ischaemic and failing human myocardium (45,46) and in rats with hypertensive HF (47). According to recent reports, TSP-4 null mice develop pronounced cardiac hypertrophy and fibrosis as well as LV dilatation and depressed systolic function in response to transverse aortic constriction (TAC) (48,49). Defects in collagen maturation were also observed, which suggests that TSP-4 is involved in the post-translational modification of collagen (48). This is supported by the fact that TSP-4 binds to the fibrillar collagens I and III in vitro (50) and has previously been shown to be associated with collagen fibrils in vivo (51). Indeed, it has been suggested that TSP-4 assists in collagen fibril assembly (50,51) possibly in the same way that TSP-5 (i.e. cartilage oligomeric matrix protein (COMP)) has been shown to facilitate collagen–collagen interactions and fibrillogenesis (52).

LV TSP-4 is rapidly up-regulated in rats in response to acute pressure overload well before the development of LV hypertrophy and fibrosis (53), but this induction is attenuated in hypertrophied heart (53). In heart, localization of TSP-4 is restricted to ECs (53,54) and SMCs derived from coronary arteries (54). This distinguishes TSP-4 from e.g. TSP-1, which is expressed in several cell types including myocytes, ECs, SMCs, and fibroblasts (40,55,56). In addition, TSP-4 expression correlated with LV remodelling in hypertensive hypertrophy as well as after MI, when the beta-blocker metoprolol elicited beneficial structural remodelling in parallel with a reduction in TSP-4 gene expression (57). In contrast, induction of TSP-1 gene expression was linked exclusively to the LV remodelling process after MI but not in hypertensive hypertrophy (57), indicating (together with a different expression pattern) that TSPs may have unique roles in the ECM remodelling process. Although a lack of TSP-4 in cardiac remodelling seems to be detrimental, increased TSP-4 expression during hypertensive hypertrophy and ageing might also be harmful due to e.g. increased collagen cross-linking and LV stiffness (47,53).

The data describing TSP-5 expression in the heart are contradictory. Fang et al. reported that TSP-5 is expressed in the adult mouse heart (58), but TSP-5 was not detected in the hearts of newborn mice (59). Moreover, the hearts of COMP-deficient mice showed normal cardiac morphology (60). Altogether, the role of TSP-5, if there is any, is not well defined in the heart, and, to our knowledge, there are no studies describing TSP-5 during cardiac remodelling.

TSP-1 and -2 have been implicated in tissue injury and inflammation such as in wound healing and foreign body reaction (61). Although TSP-1 and TSP-2 have been associated with inhibition of the cardiac inflammatory response (40,41), TSP-4 has been shown to be associated only with vascular inflammation in cardiovascular pathologies (62). In vasculature, TSP-4 influences the recruitment of macrophages by activating endothelial cells and directly interacting with macrophages to increase their adhesion and migration (62).

In conclusion, the recently reported role of TSP-4 in limiting fibrosis and LV dilatation after TAC in mice (48,49) agrees with the positive correlations of TSP-4 with both remodelling after MI and in hypertensive hypertrophy and provides interesting insights for the development of anti-fibrotic therapy targeting TSP-4. Further studies are needed to evaluate whether TSP-4 expression correlates directly with fibrosis in cardiac remodelling. TSP-4 may have a unique role in the ECM remodelling process, which differs from that of other thrombospondins such as TSP-1. Moreover, TSP-4 might be an endothelial-specific marker of pressure overload. Whether TSP-4 has an effect on the myocardial inflammatory reaction also needs to be investigated.

TWEAK and Fn14

Similar to the other TNF family members, TWEAK is initially synthesized as a transmembrane protein and cleaved intracellularly to a soluble form that can trimerize (63). Cells can express both membrane-bound and soluble TWEAK, and recently it was shown that the membrane-bound form can also bind and activate its receptor Fn14 and thus have juxtacrine effects (64). Fn14 was first discovered as a fibroblast growth factor (FGF)-inducible gene (65) then later classified as a member of the TNF receptor family and identified as the receptor for TWEAK (66). Fn14 is the smallest member of the TNF receptor family and lacks the characteristic death domain but has binding sites for TNF receptor-associated factors that activate intracellular signalling pathways (66). The structures of TWEAK and Fn14 are depicted in Figure 4. As a widely expressed cytokine, TWEAK can also be detected in the heart (67), where it is expressed by various cell types, including myocytes, as is its receptor Fn14 (67,68). Basal Fn14 expression is usually relatively low, but is induced by many substances and tissue injury (66,69,70).

The most established role of TWEAK is the promotion of the inflammatory reaction since it is able to up-regulate, also in myocytes (67), several proinflammatory cytokines, chemokines, cell adhesion molecules, and matrix-degrading enzymes (71). It has been suggested that infiltrating inflammatory cells produce TWEAK in the context of tissue injury, and the selective up-regulation of Fn14 limits its actions to the injured tissue. The TWEAK/Fn14 pathway probably has a physiological role in the facilitation of tissue repair after injury, but its sustained activation may be pathological (71).

The TWEAK-Fn14 axis has previously been implicated in autoimmune diseases, cancer, and ischaemic stroke (70), but recent reports also associate it with cardiac remodelling (67,68,72). Serum levels of TWEAK (sTWEAK) have been measured in human MI and HF patients, and low sTWEAK levels were reported to predict an adverse prognosis in HF patients (73,74). On the other hand, decreased (73) or elevated (68) sTWEAK levels have been measured in HF patients compared with healthy controls. Increased sTWEAK levels have been reported also in patients with acute ST elevation MI (STEMI) in which the sTWEAK levels correlated with an adverse outcome (75). Since acute STEMI is a thromboembolic process including platelet activation, increased sTWEAK levels might be explained by the observation that sTWEAK is released by activated platelets, at least in vitro (76). Altogether, these studies demonstrate that serum levels of TWEAK are modulated in situations requiring cardiac remodelling, although its causes, sources, and consequences remain to be studied.

TWEAK and Fn14 KO mice are viable and lack any significant abnormalities (77,78). However, transgenic mice over-expressing TWEAK in the liver with elevated levels of circulating TWEAK develop dilated cardiomyopathy with subsequent HF characterized by increased fibrosis and eccentric cardiomyocyte hypertrophy. These effects were shown to be mediated via Fn14 (68). Interestingly, increased plasma levels of TWEAK did not result in an increased myocardial inflammatory response in mice (68). In our experiments, TWEAK protein was mainly localized to rat cardiomyocytes and ECs and, notably, not to inflammatory cells (72). This is in contrast to a previous report that TWEAK was mainly expressed by infiltrating inflammatory cells in response to various forms of tissue injury (71). Furthermore, Fn14 immunoreactive cells were mainly fibroblasts in the inflammatory area, and Fn14 immunoreactivity increased as fibrosis progressed (72).

Our experimental studies show that cardiac Fn14 expression was increased both acutely and chronically in rats in response to pressure overload and MI, while TWEAK expression remained relatively constant (72), which distinguishes the expression of these cytokines from TNF-α and its receptors. Similarly, Fn14 expression was also induced in myocytes in vitro by hypertrophic agonists and other chemical stimuli as well as by mechanical stretch, in contrast to stable or decreased expression of TWEAK (72). Interestingly, in the study by Mittal et al. of denervation-induced skeletal muscle atrophy, TWEAK expression did not change while Fn14 was up-regulated (79). This is similar to our results in the heart. Yet, the absence of TWEAK or treatment with TWEAK antibody was able to rescue the muscle loss (79), which supports the hypothesis that a reduction in Fn14 expression at the base-line limits the actions of TWEAK in healthy tissue and the up-regulation of Fn14 renders tissues more sensitive to TWEAK (80). Moreover in renal ischaemia–reperfusion injury, although TWEAK expression was up-regulated only at one time point, Fn14 antibody attenuated the inflammatory reaction and the development of fibrosis (81). The results from these studies demonstrate a profoundly disparate regulation of TWEAK and Fn14 expressions in the heart and emphasize the importance of Fn14 as a mediator of TWEAK/Fn14 signalling and as a potential target of therapeutic interventions.

In the study of Chorianopoulos et al., increased expression of Fn14 was noted in the border zone and the remote non-ischaemic part of infarcted myocardium in mice at day 28 post-MI, when the myocardium was cleared from an inflammatory cell infiltrate (67). Moreover, this remote non-ischaemic myocardium is subjected to a remodelling process with fibrosis (and myocyte hypertrophy), which supports the hypothesis that TWEAK and Fn14 may also modulate ECM factors and affect ECM remodelling beyond inflammation (Figure 5). On the other hand, studies in murine NIH 3T3 fibroblasts indicated that Fn14 may play a role in cell–matrix interactions (65). In these experiments, Fn14 expression decreased cellular adhesion to the extracellular matrix proteins fibronectin and vitronectin. Whether this also happens in the heart remains to be studied. Furthermore, TWEAK can induce MMP-9 production in skeletal muscle (82), and this increased production of MMP-9 contributed to the loss of skeletal muscle mass in vivo. This suggests that MMP-9 is involved in the degradation of the components of extracellular matrix in skeletal muscle.

Figure 5. Schematic view of selected factors and pathways involved in the regulation of cardiac ECM remodelling and the potential roles of TSP and TWEAK/Fn14 in the process. The amount of cardiac ECM is determined by the balance between ECM synthesis, deposited mainly by myofibroblasts, and degradation, and regulated by factors like MMPs and TIMPs. The quality of the myocardium is affected by several factors such as intermolecular cross-links between collagen fibres. In addition to established ECM modulating factors, the matricellular protein TSP-4 and the TWEAK/Fn14 pathway seem to have important effects on cardiac ECM structure and homeostasis.

However, further investigation of TWEAK–Fn14 signalling mechanisms is needed, since it is not known whether TWEAK exerts its function on the heart in an autocrine–paracrine fashion like cytokines usually do or in an endocrine manner as proposed by Jain et al. (68). In the former case, it is not known whether the uninducible, basal levels of TWEAK are capable of mediating its effects in the heart or, as is hypothesized, increased Fn14 expression results in ligand-independent Fn14 signalling (70). Neither is it certain if TWEAK is the only ligand of Fn14 (83). TWEAK does not bind to any other TNF receptor superfamily member, including DR3 (70), but it can bind to CD163, a cysteine-rich scavenger receptor that is expressed exclusively on the monocyte/macrophage lineage (84). Although Fn14 is the only recognized receptor mediating the actions of TWEAK, there is evidence that TWEAK can exert its functions without interacting with Fn14 (85,86). Due to the distinct expression of TWEAK and Fn14, it would be interesting to compare the difference between the absence of TWEAK and the absence of Fn14 during cardiac remodelling in order to evaluate the function of this cytokine receptor pair in the heart.

In conclusion, cardiac remodelling influences the TWEAK/Fn14 pathway and thereby justifies it as a potential molecular target in the prevention of pathological remodelling of the heart. Further studies are needed to test whether inhibition of TWEAK signalling is sufficient to block actions of the pathway in the heart or whether direct inhibition of Fn14 downstream signalling is essential to prevent the potentially unfavourable consequences of Fn14 receptor activation.

Conclusions

In addition to the already established ECM-modulating factors, like TSP-1 and TNF-α, recently recognized novel factors TSP-4 and TWEAK/Fn14 provide new insights into the cardiac remodelling process (Figure 5). A common feature of the matricellular protein TSP-4 and TNF-receptor family member Fn14 is that they seem to have important effects on cardiac ECM structure and homeostasis especially in response to increased load (41,42,68). Another feature these molecules share is that their basal expression is low but they are up-regulated upon cardiac injury. This would make at least Fn14 inhibition an attractive candidate for therapy due to potentially minor harmful effects in extracardiac tissues. TSP-4 augmentation, on the other hand, might potentially be beneficial in the setting of acute heart failure (87).

The modifications of ECM affect cardiac structure and function profoundly, and therefore ECM remodelling is an essential component of cardiac remodelling. Understanding the mechanisms controlling ECM remodelling is important for the development of novel therapeutic interventions or diagnostic and prognostic markers for cardiovascular diseases.

Acknowledgements

This study was financially supported by the Academy of Finland (Center of Excellence funding), Aarne Koskelo Foundation, Aarne and Aili Taponen Foundation, Astra-Zeneca Research Foundation, Biocenter Oulu, Emil Aaltonen Foundation, the Finnish Foundation for Cardiovascular Research, the Finnish Medical Foundation, Ida Montin Foundation, Instrumentarium Research Foundation, Maire Taponen Foundation, Oulu University Scholarship Foundation, Paavo Ilmari Ahvenainen Foundation, the Research and Science Foundation of Farmos, and the Sigrid Jusélius Foundation.

Declaration of interest: The authors state no conflict of interest and have received no payment in preparation of this manuscript.