OUP user menu

Molecular and mechanical synergy: cross-talk between integrins and growth factor receptors

Robert S Ross
DOI: http://dx.doi.org/10.1016/j.cardiores.2004.04.027 381-390 First published online: 15 August 2004


The full repertoire of molecules and mechanisms which lead to cardiac hypertrophy are poorly understood. Many studies over the last several decades have shown how various growth factors are involved in the hypertrophic response. It has also been intuitively obvious that mechanical mechanisms which impose hemodynamic loads on the working myocardium must also be involved in this process. Integrins are cell adhesion receptors that are potent bi-directional signaling molecules. They are cellular mechanoreceptors in many cells and are clearly one of the molecules which orchestrate mechano-biochemical coupling in the heart. In recent years they too have been shown to be involved in the hypertrophic response pathway. This review will detail background information on integrins in general, discuss integrins in the myocardium and illustrate how integrin and growth factor signaling pathways might combinatorially function in the heart.

  • Molecular and mechanical synergy
  • Integrins
  • Growth factor receptors

1. Introduction

Postnatal growth of the heart occurs dominantly through hypertrophic mechanisms. While physiological growth in the mature, adult heart may occur as a result of exercise, the bulk of hypertrophic myocardial growth is from pathological stress signals resulting from mechanical triggers and/or endocrine, paracrine, and autocrine factors. Whether orchestrated via mechanical means or by extracellular circulating factors, the hypertrophic response pathway begins at the cell-surface through receptors or channels which activate intracellular signaling cascades, ultimately effecting nuclear cues which culminate in altered gene expression. Numerous signaling pathways have been linked to the hypertrophic response of the myocardium [1,2]. While the growth factor directed hypertrophic pathways have been more extensively studied, the molecular machinery which directs mechanical sensing in the cardiac myocyte is incompletely understood. Candidates for the sensor have included cell surface adhesion receptors termed integrins (below) as well as mechanosensitive ion channels. These sensors, functioning alone or in tandem, transmit information through a complex of proteins at the lateral boundaries of the sarcomere, the Z-disc [3–5]. The goal of this review is to discuss the interaction between growth factor pathways and one of the mechanical sensors, the integrins. I will present information on the structure, function and expression of integrins in the heart, with emphasis on the myocyte compartment. Integrin signaling mechanisms and a summary of studies which have documented general interactions of the integrin and growth factor pathways will follow. Subsequently, I will detail the information on cross-talk between integrins and growth factors in the myocardium.

2. Integrin structure

Extracellular matrix (ECM) substances provide cells with a structural, chemical and mechanical substrate that is essential for normal development and responses to pathophysiological signals. Glycoprotein transmembrane receptors termed integrins are the primary link between extracellular matrix ligands and cytoskeletal structures. They serve both as adhesive receptors and also direct intracellular signaling events [6,7]. Integrins orchestrate multiple functions in the intact organism including organogenesis, regulation of gene expression, cell proliferation, differentiation, migration and death.

Integrins are expressed in most cell types. One cell can express a variety of integrin receptors, thus allowing them to interact with many extracellular matrix ligands. In the cardiovascular system integrins are expressed in cells of the vasculature, blood, neurons, cardiac myocytes and non-muscle cardiac cells. Studies have shown that integrins are involved in heart formation and function, as well as the development of cardiac diseases.

Integrins are a complex family of non-covalently associated heterodimeric transmembrane receptors composed of α and β subunits, with α subunits ranging from 120 to 180 kDa while β subunits are 90–110 kDa. Each subunit contains a large extracellular domain (700–1100 amino acids), a single transmembrane span and short cytoplasmic tails, ranging from 20 to 60 amino acids [8], with the exception of β4 integrin which has a cytoplasmic domain of 1000 residues. The two subunits pair intracellularly as precursors, prior to further carbohydrate processing in the Golgi and transport of the mature αβ heterodimers to the cell surface [9]. About half of the α subunits have a globular head at its N-terminus, which contains seven repeating homologous sequences each folded in a β propeller. In the rest of the α subunits, an I-domain composed of about 200 residues is inserted into the β propeller fold. The I-domain is homologous to the A-domain of the von Willebrand factor protein and therefore is alternatively termed the I/A domain. It is in the I-domain which contains the metal ion-dependent adhesion site or “MIDAS” motif. The I-domain with the MIDAS motif forms the ligand-binding site. Sequences resembling the I-domain of the α chain have also been found in β subunits and are believed to form the ligand-binding site within β chains [10].  It is through the short cytoplasmic domain that integrins signal (see below) and also interact with the cytoskeletal components. Information on integrin structure is displayed in the inset of Fig. 2 and can be explored further in a recent review [8].

Fig. 2

Diagram of integrins as transmembrane adhesion and signaling receptors. The α subunits (black) and β subunits (gray) bind to extracellular matrix extracellularly, components of the cytoskeleton intracellularly, and orchestrate signaling through numerous pathways which can also in part be activated by various growth factor receptors. In turn, the integrins as adhesive and mechanotransduction receptors can affect cytoskeletal remodeling and nuclear events, both of which could also alter growth factor mediated events. Signaling shown here is of the “outside–in” type (see text). The inset on the right upper portion of the figure shows some detail of integrin structure. A=A-domain/MIDAS (metal ion-dependent adhesion site), Ca2+=divalent cation binding sites termed an EF hand, S–S=disulfide bond, TM=transmembrane region, C=cytoplasmic domain.

3. Integrin family members

Integrins are expressed in both vertebrates and invertebrates [11]. In the worm C. elegans, two α-chains and a single β-chain gene are coded for by its genome, creating two integrin receptors. One of them is a putative RGD-binding integrin and the other recognizes laminin. (RGD is Arg-Gly-Asp, an amino acid sequence which forms one of the canonical integrin recognition sites, for fibronectin, vitronectin and a variety of other adhesive proteins. This tri-peptide sequence is recognized by several integrins such as α5β1, αIIbβ3 and most of the αVβ integrins.) In the fruit fly Drosophila melanogaster at least five integrin heterodimer receptors are formed from five α chains genes and two for β chains. These receptors bind to either RGD or laminin-containing extracellular matrix substances.

Vertebrates have evolved a more complex integrin family. As an example, 18 α and 8 β integrin genes have been identified in mammals. These subunits assemble into over 24 integrin heterodimers in man [6]. Phylogenetic analyses of those sequences found in both vertebrates and invertebrates suggested that the α integrin genes have probably evolved from a single α integrin ancestor by gene duplication. Currently, the evolution of the β integrin genes has not been clearly resolved [12]. Many extracellular matrix proteins serve as ligands for the integrin receptors including fibronectin, collagen, laminin, vascular cell adhesion molecule-1 (VCAM-1), intercellular adhesion molecule-1 (ICAM-1), vitronectin, tenascin, osteopontin, von Willebrand factor and thrombospondin. Each ECM protein can serve as a ligand for several different integrin receptors and, similarly, a majority of integrins can bind several ECM ligands. Moreover, the integrin family is made even more complex by alternatively spliced forms of several integrin subunits. This large integrin repertoire allows for many alternatives for cells to cope with and adapt to their environmental changes.

4. Integrin expression in the myocardium

In heart, expression of α1, α3, α5, α7, α9, α10, β1, β3, and β5 subunits is found in myocytes (Table 1). Expression of most of the α subunits is temporally modified and developmentally regulated. For example, α1β1 and α5β1 are expressed in the embryonic heart, down-regulated postnatally, and like many fetal genes, are up-regulated with hemodynamic stress [4]. β1 isoforms are the dominant myocyte β integrin subunit. β1A and β1D are the two major β1 isoform splice variants found in the myocytes. β1A is expressed in embryonic cardiac myocytes and becomes downregulated at approximately E18 of mouse gestation. β1D integrin is a skeletal and cardiac muscle isoform. In heart, expression of β1D begins during late gestation and becomes the major form in the postnatal heart [13]. We have demonstrated that following hemodynamic loading of the mouse ventricle, β1D integrin expression is upregulated [14] (Fig. 1).

Fig. 1

β1D integrin expression is increased following hemodynamic loading of the murine left ventricle as shown via immunostaining (panels A and B) as well as Western blotting (panel C) with an isoform-specific antibody. Panel A—Immunostaining of left ventricular tissue from a sham-operated mouse with an anti-β1D integrin antibody. Panel B—Immunostaining of left ventricular tissue from a mouse following 7 days of transverse aortic constriction with an anti-β1D integrin antibody. Panel C—Western blotting using an anti-β1D integrin antibody and whole heart lysate from sham and pressure overloaded ventricular tissue. Duplicate lanes are displayed for each condition. POL=pressure overload. Panels A and B magnifications are both 400 × .

View this table:
Table 1

Cardiac myocyte integrin subunits, primary ligand binding and relative developmental changes in their expression levels

SubunitPrimary ligand(s)EmbryoNeonateAdultHypertrophy
α1LN, Col++
α3BLN, Col, FN++++++NC
β1ALN, Col, FNEarly++++
β3VN, FN, OPN??+
  • LN=laminin; Col=collagen; FN=fibronectin; VN=vitronectin; OPN=osteopontin; late=late period of gestation; early=early stage of gestation; “+”= low/basal expression’ “−”=no expression, “++”= moderately high expression; “+++”=highly expressed; ↑=increased following hypertrophy, NC=no change; ?=undetermined, *=by polymerase chain reaction only.

In addition to muscle cells, the heart of course contains several non-myocyte cell types including fibroblasts, smooth muscle cells, and endothelial cells. Fibroblasts constitute almost 2/3 of the cardiac cell population. Following stressful stimuli such as might occur with hemodynamic loading or as a result of post-infarct remodeling, cardiac myocytes hypertrophy while cardiac fibroblasts increase in number and produce ECM proteins, such as collagens and fibronectin. Cardiac fibroblasts sense mechanical stimuli through integrins and respond by altering ECM production and initiating mechanotransduction events [15–17]. Abnormal production of ECM in combination with exuberant proliferation of cardiac fibroblasts can lead to some of the detrimental consequences of ventricular remodeling, culminating in cardiac dysfunction. The cardiac fibroblasts express a repertoire of α subunits like that of the cardiac myocyte but they do not express α6 and α7 [18]. In contrast, αv, the collagen-specific α2 subunit, and α8 are uniquely expressed by the cardiac fibroblasts but not by cardiac myocytes [19–21].

5. Integrin organization and signaling

As discussed above, integrins provide dynamic links between cells and ECM molecules. In addition to their function as adhesion molecules, integrins also are mechanotransducers [22] meaning they convert mechanical forces to biochemical signals. The myocyte contains a regular array of actin and myosin filaments organized into the sarcomere between two adjacent Z lines. Non-sarcomeric actin microfilaments attach the sarcomere to the cardiomyocyte cell membrane. By electron microscopy, the cytoskeleton appears as a dense and seemingly random array of fibers but it plays a critical structural role, supporting the cell membrane and directing the movements of cell organelles and other elements in the cytosol. In the cardiac myocytes, the cytoplasmic tail of the β1 integrins connects to bundles of actin filaments, via bridging proteins like α-actinin, talin, paxillin and vinculin/metavinculin. Because of their co-localization, these proteins are participants in linking the actin filaments through the membrane to the ECM. This arrangement is shown in Fig. 2. Talin has been shown to bind directly to the β1 integrin cytoplasmic tail, thus linking the integrin receptor to the actin cytoskeleton [23]. Therefore integrins connect the extracellular matrix to the cytoskeleton and to cytoplasmic proteins. Interestingly, the striated muscle-specific β1D integrin has been shown to bind more tightly to talin than the ubiquitously expressed β1A-integrin. β1D may thus provide a firmer attachment to the cytoskeleton, than similar attachments by the ubiquitously expressed β1A integrin in non-muscle cells. This arrangement is clearly advantageous in the continuously contracting cardiac muscle cell.

Initially, integrins were thought to function only as cell-matrix adhesion molecules. Now it has become well accepted that they are important signal transducers [7]. (Figs. 2 and 3) It is through their signaling as well as adhesive roles that integrins can cooperatively function with growth factors (see below). ECM-integrin interactions function in a bi-directional manner across cell membranes. Once integrins bind to ECM, their interaction with the actin cytoskeleton leads to redistribution of integrins into specific structures known as focal adhesions. Vinculin, talin, α-actinin and paxillin all co-localize within the focal adhesions [24]. After binding ECM ligands, integrin receptors transmit signals across the membrane to a host of cytoplasmic molecules. Through these extracellular to intracellular events termed “outside–in signaling,” integrins regulate cell attachment, survival, proliferation, cell spreading, differentiation, cytoskeleton reorganization and cell shape [25]. They can even influence cell migration, differentiation, gene expression, suppression of tumorigenicity, changes in intracellular pH or concentration of cytosolic Ca2+.

Fig. 3

Diagram of “inside–out” integrin signaling, a mechanism by which growth factor stimulation can alter integrin function. Growth factors can alter signaling and cytoskeletal organization independent of integrins. Both of these events are potential sites of growth factor-integrin interactions.

Since the integrins do not themselves possess enzymatic activity, to signal they must trigger downstream molecules [6,26]. Examples include activation of tyrosine kinases such as pp125 focal adhesion kinase (FAK), or small GTPases such as Rho or Rac, and regulation of cytoskeletal components such as talin, paxillin or p130CAS. The integrin cytoplasmic domain is essential in this process and has been shown to bind numerous molecules. These include calreticulin and FAK, as well as melusin and muscle integrin binding protein (MIBP), both of which are preferentially expressed in muscle [27–30].

In addition to transmitting signals from ECM to the intracellular environment, integrins can be modified by agonists that bind to non-integrin cellular receptors and in turn modify integrin activation/function; a process termed “inside–out” signaling. This again is a means by which growth factors and integrins can cooperatively signal/function. (Fig. 3) Specifically, these events can lead to both increased binding of integrin to ECM ligand as well as clustering of multiple integrins in close spacing within the cell membrane. Recent work has shown that events initiated by the non-integrin receptors cause alterations in the integrin cytoplasmic domain ultimately effecting a conformational change and conversion of the integrin from a low to a high activation state [31]. The modulation of “activation state” may actually alter two discrete processes, i.e. “affinity” and “avidity [32].” In this scenario, integrin affinity relates conformational changes in the integrin heterodimer to strength of ligand binding, akin to that which can be measured as the strength of interaction between antigen and antibody in solution. Avidity on the other hand may be related to agonist mediated “clustering” of multiple integrin heterodimers in the cell membrane. When clustered receptors are in place in the membrane, the receptor displays a higher “apparent affinity” (termed “functional affinity” or “avidity”) which relates the density of the receptor cluster to the strength of ligand binding. Therefore “inside–out” signaling mechanisms can cause integrins to undergo a switch from a “low affinity/avidity” state to a “high affinity/avidity state.” It is in this manner integrin function can be modulated by substances which do not bind directly to the integrin receptor. Morphologically, this clustering is apparent in the co-localization of integrins in Z-bands of myocytes, with cytoskeletal components and signaling complexes [33]. The Z band localization in vivo can be thought of as analogous to the focal adhesion which is more commonly discussed in studies of integrin localization and signaling in non-myocyte cell studies.

6. Growth factors and integrins: general concepts

Integrins themselves signal through a host of pathways, but it is rare for any signaling pathway to function in isolation. Extensive studies have been performed to show that amongst other collaborative interactions, integrins and growth factors (GF) form a particularly robust synergy [34]. Receptor tyrosine kinases (RTKs), G-protein coupled receptors (GPCRs) and cytokines have all been linked to this response. As with so many integrin studies, many of these have been performed in either platelet or fibroblast model systems. Spatially, the intersection of these two pathways occurs at both membrane proximal and distal sites (Fig. 4). As an example, aggregation of integrin receptors from fibroblasts leads to an accumulation of over 20 signaling related molecules adjacent to the plasma membrane, including Src, Fyn, RhoA, Rac1, Ras, MEK, ERK and JNK [35]. Direct associations between integrins and RTKs have been identified. In fact, these associations do not appear generic but are subunit specific. As an example, αVβ3, which is one of the integrins expressed in endothelial and smooth muscle cells, was identified to associate with the platelet-derived growth factor (PDGF) receptor (R), the insulin receptor, the insulin-like growth factor-I receptor (IGF-1) and the vascular endothelial growth factor receptor (VEGF-R). Relevant cardiovascular end-points were elucidated from studies which found that inhibitors of αVβ3 integrin decreased both IGF-1 signaling and also reduced atherosclerotic lesions in a pig model [36]. In contrast, the epidermal growth factor (EGF) receptor interacts with α5β1 and α6β4 integrin [37] while α6β4 associates with Erb-B2 [38]. Given this preferential association of one integrin subunit with particular growth factor receptors, one can begin to understand how altered expression of cell-surface integrins could lead to significant changes in cellular signaling events. How cells can be alternatively driven to proliferate or exit from the cell-cycle might be taken as another example of integrin-specific, growth-mediated events. An example of this is the skeletal myoblast proliferation stimulated when the myoblast adheres to fibronectin (mediated through an integrin receptor like α5β1), as compared to differentiation into multinucleated myotubes as would occur when the cells are plated onto laminin (an event mediated via α7β1) [39]. Similarly, absence of skeletal muscle β1 integrin prevents normal fusion of skeletal myocytes in the intact mouse [40].

Fig. 4

Sites and types of integrin and growth factor interaction within the cell. See text for discussion of the various intersection points.

Integrin receptor occupancy and clustering may be required for RTK tyrosine phosphorylation/activation [34]. Further, growth factor receptor occupancy may lead to preferential activation of the subset of RTK receptors which are bound to integrins [41]. In addition to the more traditional synergism found between these two signaling systems, it has been shown that integrin activation of the growth factor receptors can occur even in the absence of growth factor. For example, integrin-mediated clustering of EGFR may lead to EGFR autoactivation [42].

A second level of intersection between the GF receptors and integrins is at the level of plasma membrane lipid rafts, the sub-domain of the membrane rich in cholesterol and sphingolipids which has been linked to cell signaling [43]. The rafts are generally sites of proteins which contain glycosyl phosphatidylinositol (GPI) anchors and transmembrane proteins may associate there as components of protein oligomers. Integrins, as well as growth factor receptors, have been identified in the lipid raft [44] and shown to interact. This is illustrated from a study evaluating differentiation of the myelin-forming oligodendrocyte [45]. In this cell type, PDGF is required for both proliferation and later, cell survival. The authors found that as the functional role of PDGF shifts from proliferation to survival, its receptor becomes sequestered in lipid rafts. Since integrins (specifically α6β1, a laminin receptor) accentuate the “survival signaling” from PDGF, the study also evaluated the sub-cellular localization of the integrin and found this specific integrin receptor dominantly localized in the exact subset of lipid rafts where the PDGF-R was detected. It did so only when the cells were attached to laminin. Finally, the study detailed that not only did laminin-plated oligodendrocytes harbor α6β1 and the PDGF-R in the same raft subsets, but appropriate downstream signaling constituents which are involved in oligodendrocyte survival—including activated phosphoinositide 3-kinase (PI3K), protein kinase B (PKB)/AKT, and focal adhesion kinase (FAK)—were also localized in the rafts. This work is representative of how lipid rafts may be a cellular platform to allow for integrin/growth factor interactions.

While not completely distinct from the above interaction, another intersection between the growth factors and integrins can reside at more downstream signaling molecules (Figs. 2 and 4). In a general sense, activation of particular signaling cascades directly by integrins could lead to growth factor dimerization and phosphorylation/activation. Alternatively, the integrin signaling event could synergize with one independently initiated by the growth factor. One example might be FAK which is one of the canonical integrin signaling molecules, demonstrated to bind directly to the cytoplasmic tail of β integrin subunits [46]. FAK activation leads to the binding of a number of SH2 domain containing Src-family kinases, Cas, paxillin, the 85-kDa subunit of PI3K and the adapter proteins Grb7 and Grb2-SOS complex. In addition, FAK activation leads to further assembly of a host of signaling complexes which can regulate small GTPases of the Rho and Arf families, recruits Rac and its downstream effector PAK to the cell membrane and ultimately influences mitogen activated protein kinase (MAPK)/extracellular signal-regulated kinase (ERK) activation, a key intersection point with growth factor signaling [47]. FAK itself may also directly (or perhaps indirectly) interact with the EGFR and PDGFR [48]. An additional link to the ERK/ MAPK pathway intersection point is through Ras/Raf/MEK in which src-family kinases (SFK), adapter proteins such as Grb2/SOS and Rac, may also be involved. As an example, fibroblasts in suspension have impaired activation of Raf-1 and downstream MEK/ERK [49]. Integrin adhesion/signaling may be required for recruitment of Rac to the plasma membrane, allowing coupling to PAK, PAK activation, Raf and MEK phosphorylation/activation [50]. This level of cooperativity between GFs and integrins may in part be regulated by the ability of integrins to organize the cytoskeleton (perhaps specifically cortical actin filaments) and also the ability to direct translocation of signaling components of both pathways (e.g. ERK) from cytoplasm to nucleus [51]. How the translocation of signaling molecules is regulated by integrins is intriguing but poorly understood.

Before leaving this area, one must also realize that integrins can activate c-Jun NH2-terminal kinases (JNKs) which are involved in hypertrophic signaling [52]. A relevant example stems from the work of MacKenna, et al. [52] who showed that a 4% static biaxial stretch of rat cardiac fibroblasts resulted in both ERK2 and JNK1 activation. The investigators found that ERK2 was only activated when the cells were plated on fibronectin, while JNK1 was activated when the cells were plated on fibronectin, vitronectin, or laminin. Interestingly, plating cells on collagen before stretching did not activate either kinase. The results could be blocked by adding specific integrin inhibitors: ERK2 activation could be modified by simultaneously blocking α4 and α5 integrins while activation of JNK1 could not be blocked with the α4 and α5 integrins inhibitors, suggesting that other integrin subunits were involved. These results are potentially important in consideration of the changes that occur in hypertrophic and post-infarct remodeling.

The bulk of the above detail has centered on the intersection between integrin and receptor tyrosine kinase signaling. It is worth noting that important studies have also detailed connections between integrins and GPCRs. A host of G-protein coupled agonists (including many relevant to cardiac myocyte hypertrophy) have been shown to phosphorylate FAK. This includes bombesin, endothelin, gastrin, lysophosphatidic acid (LPA), and angiotensin II [53–55]. A FAK-related kinase termed Pyk-2 has also been shown to be involved in both integrin and histamine GPCR response pathways [56]. Likewise, thrombin and P2Y receptor signaling have been linked to ECM anchorage and integrins [57,58]. Further, cytokine and integrin signaling have been linked. Examples include interleukin (IL)-1 receptors that have been shown to localize in focal contact sites and IL-8 stimulated neturophils, which only activated NF-kappa B when the cells were attached to substrate [59]. Transforming growth factor β (TGF-β) is an important cytokine in fibrotic responses of many organs, including the heart. Integrins are in part responsible for the activation of TGF-β [60].

As mentioned initially, the integrins signal bi-directionally. The bulk of this section has centered on typical “outside–in” signaling, i.e. from the cell membrane towards the nucleus. Cross-talk between these signaling systems can also act via an “inside–out” manner. One example previously mentioned was that integrins activate the EGF receptor in the absence of EGF ligand [42]. Further, VEGF can alter cellular adhesion of endothelial cells via αVβ3, αVβ5 and β1 integrins [61]. TGF-β stimulation of fibroblasts can cause increased expression of α5β1 integrins [62] while, conversely, overexpression of this integrin heterodimer increased TGF-β expression [63]. In sum, it is clear that bi-directional cross-talk exists between the growth-factor and integrin pathways.

7. Growth factors and integrins: cardiac hypertrophy

The interactions between growth factors and integrins in the myocardium have not been as extensively studied as in other tissues but synergy between these two signaling systems has begun to be documented in the cardiac myocyte and non-myocyte cardiac cells [4]. ECM is an essential component of the developing heart and it is altered in amount and distribution following pathophysiological or growth-factor mediated events such as pressure overload, myocardial infarction or following elaboration of substances such as endothelin-1 or angiotensin II. Terracio et al. [19] were some of the first investigators to document the importance of integrins in the myocardium and how their expression was changed with hypertrophy. Gullberg et al. [64] noted that integrin and PDGF-mediated signaling intersected in that PDGF-stimulated collagen gel contraction orchestrated by cardiac fibroblasts could be partly inhibited by β1 integrin antibodies. Sadoshima and Izumo [3] subsequently utilized mechanical loading of cultured myocytes to simulate hemodynamic loading of the intact myocardium. They showed how the mechanical stimulus induced a variety of intracellular signaling pathways (many of which are known to be involved in growth factor and integrin signaling cascades) and also provoked paracrine release of factors such as angiotensin II (AII) [65]. AII-stimulation of cardiac fibroblasts has been shown in several studies to modulate integrin localization and expression [21,66] Importantly, osteopontin (OPN), a cytokine which binds to several cell surface receptors, including integrins αVβ3, αVβ5, αVβ1, α4β1, α8β1, and α9β1, as well as the hyaluronan receptor, was shown to be produced by both cardiac fibroblasts and myocytes and to be upregulated in the hypertrophic and failing ventricle of humans and animals [16,67]. OPN production is upregulated by AII and recent work from our group has shown that mice deficient in OPN have reduced AngII-mediated cardiac fibrosis and hypertrophy likely due to the decreased adhesion to ECM and reduced cell proliferation by the OPN−/− cardiac fibroblasts [68] (Fig. 5). Similarly, other groups have shown that the OPN−/− mice have reduced collagen production post-myocardial infarction (MI), developed increased post-MI dilation and that AII increased OPN production by cardiac endothelial cells via a MAPK mechanism [69,70].

Fig. 5

Ablation of osteopontin expression leads to reduced fibrosis following angiotensin II stimulation. Histology of murine myocardium following AII stimulation is shown. Panel A—Wild-type mouse. Panel B—osteopontin knockout mouse (OPN−/−). Magnification—20 × .

Several groups, including our own, have shown that the hypertrophic response of cultured myocytes and the intact ventricle can effect increased expression of integrins, and that integrins and ECM ligands for integrin receptors can invoke the hypertrophic response pathway independent of growth factors [14,71,72]. Further, it has become clear that there is synergy between growth factor and integrin mediated stimulation of hypertrophy. Much of this overlap may function through FAK, FAK related kinases (such as Pyk2) and may ultimately signal through protein kinase C (PKC) and ERK related pathways [13,73–79]. In separate studies, integrins have been shown to utilize FAK, Src, Grb2 and Ras to induce hypertrophic signals through to p38 MAPK [80]. Further, mice haploinsufficient for Grb2 had reduced ventricular hypertrophy and fibrosis following aortic banding [81]. Associated with these phenotypic results was defective p38 MAPK and JNK activation, though ERK activation was preserved. These studies have shown that agents known to induce cardiac myocyte hypertrophy including endothelin-1and phenylephrine signal through pathways concomitantly and perhaps synergistically with integrins. Further, mechanical loading of the intact heart also evokes responses related to integrin signaling but also the ones more typically associated with growth factors. Thus like the non-cardiac studies discussed above, one level of interaction between growth factor and integrin signaling may reside in cytoplasmic tyrosine kinases and downstream related molecules such as the p38 MAPK, JNK and ERKs. In addition, there may be synergy orchestrated by molecules that bind directly to integrin cytoplasmic tails. One example is melusin, which was recently shown to be a molecule involved in the mechanotransductive response linked to the hypertrophic pathway through activation of glycogen synthase kinase-3 (GSK-3) [82].

A particularly interesting more proximal relationship of growth factor and integrin interaction has been documented for adrenergic and cholinergic receptors [83–85]. In this regard, Wang's group has used an atrial myocyte model to show that ligand binding to β1 integrin could (1) inhibit adenylyl cyclase and cause inhibition of l-type Ca2+ current effected by acetylcholine withdrawal and (2) reduce signaling through α1 adrenergic receptors and increase α2 adrenergic signaling which also regulate l-type Ca2+ current [84,85]. Studies have also documented abnormal muscarinic regulation in cardiac-like cells derived from β-1 integrin deficient embryonic stem cells [86]. In related work, integrin expression was increased in adult rat ventricular myocytes by α-1 adrenergic agents (not β adrenergic agonists) via ERK-mediated signaling. In this study, integrin ligation was found to reduce β-adrenergically stimulated apoptosis [83].

An important concept perhaps underemphasized earlier in this review is that modulation of growth factor function by integrins is in part related to their ability to change the actin-based cytoskeleton of cells. This was shown in some of the above work relating to l-type Ca2+ channels [84,85]. This concept was further illustrated in studies by Wei's [87] group who determined that β1 integrins influenced transcriptional activity of well-accepted hypertrophic marker genes such as α-skeletal actin working in tandem with RhoA, FAK and PI3-K mechanisms, but importantly requiring a well-organized myocyte cytoskeleton.

8. Conclusion

In summary, it is clear that integrins are complex molecules that can interact with growth factors. Extensive work is necessary to further clarify the mechanisms by which they function in the myocardium. Modulation of the levels of integrin protein on the cell surface, variation in subunit types expressed on a particular cell, and strength of cell adhesion to particular ECM ligands are all poorly understood concepts in the myocardium. All could impact on how the heart responds to stressful stimuli, particularly those imposed by mechanical forces. Information on how integrins signal in the heart is also incomplete. Future studies need to detail mechanisms of integrin function in all cell types in the myocardium and how this large family of molecules interacts with other signaling molecules, including growth factors.


Thanks to Drs. Alan Collins and Willa Hsueh for assistance with Fig. 5.


  • Supported by grants from the National Institutes of Health (HL57872 and HL73393).

  • Time for primary review 26 days


  1. [1]
  2. [2]
  3. [3]
  4. [4]
  5. [5]
  6. [6]
  7. [7]
  8. [8]
  9. [9]
  10. [10]
  11. [11]
  12. [12]
  13. [13]
  14. [14]
  15. [15]
  16. [16]
  17. [17]
  18. [18]
  19. [19]
  20. [20]
  21. [21]
  22. [22]
  23. [23]
  24. [24]
  25. [25]
  26. [26]
  27. [27]
  28. [28]
  29. [29]
  30. [30]
  31. [31]
  32. [32]
  33. [33]
  34. [34]
  35. [35]
  36. [36]
  37. [37]
  38. [38]
  39. [39]
  40. [40]
  41. [41]
  42. [42]
  43. [43]
  44. [44]
  45. [45]
  46. [46]
  47. [47]
  48. [48]
  49. [49]
  50. [50]
  51. [51]
  52. [52]
  53. [53]
  54. [54]
  55. [55]
  56. [56]
  57. [57]
  58. [58]
  59. [59]
  60. [60]
  61. [61]
  62. [62]
  63. [63]
  64. [64]
  65. [65]
  66. [66]
  67. [67]
  68. [68]
  69. [69]
  70. [70]
  71. [71]
  72. [72]
  73. [73]
  74. [74]
  75. [75]
  76. [76]
  77. [77]
  78. [78]
  79. [79]
  80. [80]
  81. [81]
  82. [82]
  83. [83]
  84. [84]
  85. [85]
  86. [86]
  87. [87]
View Abstract