Interactions between extracellular matrix and growth factors in wound healing

Authors


Reprint requests:
Gregory S. Schultz, PhD, Department of Obstetrics and Gynecology, University of Florida, 1600 SW Archer Road, Gainesville, FL 32610-0294. Tel: +352 273 7560;
Fax: +352 392 6994;
Email: schultzg@ufl.edu

ABSTRACT

Dynamic interactions between growth factors and extracellular matrix (ECM) are integral to wound healing. These interactions take several forms that may be categorized as direct or indirect. The ECM can directly bind to and release certain growth factors (e.g., heparan sulfate binding to fibroblast growth factor-2), which may serve to sequester and protect growth factors from degradation, and/or enhance their activity. Indirect interactions include binding of cells to ECM via integrins, which enables cells to respond to growth factors (e.g., integrin binding is necessary for vascular endothelial growth factor-induced angiogenesis) and can induce growth factor expression (adherence of monocytes to ECM stimulates synthesis of platelet-derived growth factor). Additionally, matrikines, or subcomponents of ECM molecules, can bind to cell surface receptors in the cytokine, chemokine, or growth factor families and stimulate cellular activities (e.g., tenascin-C and laminin bind to epidermal growth factor receptors, which enhances fibroblast migration). Growth factors such as transforming growth factor-β also regulate the ECM by increasing the production of ECM components or enhancing synthesis of matrix degrading enzymes. Thus, the interactions between growth factors and ECM are bidirectional. This review explores these interactions, discusses how they are altered in difficult to heal or chronic wounds, and briefly considers treatment implications.

The synthesis and deposition of extracellular matrix (ECM) is a critical feature in the healing of chronic wounds, which are characterized by substantial loss of the dermal matrix, and acute wounds such as inguinal hernia where matrix defects result in compromised function. While acute wounds progress through a relatively orderly series of events, chronic wounds exhibit delayed healing where the healing events fail to progress through the sequential phases of healing and are often characterized by an excessive inflammatory response. Cells that synthesize dermal ECM proteins must meet increased production demands, exhibiting prolonged and sustained activity that is needed to successfully accomplish healing. On the other hand, when matrix defects result in the loss of functional integrity, as in the formation of a hernia, the deposition or replacement of a functional matrix is required to accomplish successful and durable healing.

Interactions among the ECM, growth factors, and cells underlie tissue generation and regeneration, including wound healing. These elements interact in an ongoing, mutually influential series of events that has been referred to as dynamic reciprocity1 (Figure 1). Wound healing has been arbitrarily divided into the overlapping phases of inflammation, proliferation, and remodeling—each of which is characterized by dynamic, reciprocal interactions among ECM, growth factors, and cells.2 The events during these phases have been well described in the literature,3–5 and examples of the dynamic reciprocity pervade these accounts. For instance, during the inflammatory phase, fibronectin and other ECM protein fragments in the wound area serve as chemoattractants for monocytes,6 which then bind to ECM proteins. This binding stimulates phagocytosis,7 leading the monocytes/macrophages to further break down ECM fragments and other debris in the area.3 Adherence of monocytes to ECM proteins also stimulates the expression of growth factors8 that can then act on cells to affect the synthesis of ECM components (e.g., proteoglycan synthesis by fibroblasts).9

Figure 1.

 Cells and ECM interact in an ongoing, bidirectional manner that has been referred to as dynamic reciprocity.1 ECM, extracellular matrix.

Interactions between growth factors and ECM in this dynamic reciprocity take several forms. Some are direct, such as the direct binding of growth factors by ECM components, and some are indirect, such as the requirement that cells be adhered to ECM in order to respond to the growth factor signal. In this review we consider various types of ECM-growth factor interactions. Here we focus on the relevance of these interactions to wound healing, although they are broadly applicable to the processes of tissue generation and even homeostasis.1,10 Before considering specific types of ECM-growth factor interactions, we briefly review the composition of the ECM and its roles in the wound healing. We then consider difficult to heal/chronic wounds because they represent situations in which ECM-growth factor interactions are disturbed and where addressing these disruptions may have treatment implications.

OVERVIEW OF THE ECM

ECM is assembled from components synthesized and deposited outside the cell surface that provide structural and functional integrity to connective tissues and organs.2,11 The synthesis and deposition of ECM largely occur in response to growth factors, cytokines, and mechanical signals mediated via cell surface receptors.12 These cell surface receptors provide points of attachment that cells can use to sense mechanical disruptions and to remodel the deposited matrix to render it structurally and functionally viable.13 The ECM can also serve as a reservoir or provisional matrix for growth factors and other proteins deposited upon wounding from degranulating cells and plasma proteins found in the blood.14 More recent studies have indicated that matrix changes can occur before actual injury that appear to predispose individuals to chronic repair processes.15,16

In the context of wound healing, there are at least four major classes of ECM: (1) structural proteins such as the collagens and elastin (although collagen can also have functional properties); (2) multidomain adhesive glycoproteins such as fibronectin, vitronectin, and laminin; (3) glycosaminoglycans (GAGs) such as hyaluronan and proteoglycans such as versican, syndecans, glypicans, and perlecan (that contain GAG side chains, (e.g., chondroitin sulfate and heparan sulfate), often in large amounts; and 4) matricellular proteins such as secreted protein acidic and rich in cysteine (SPARC; also known as osteonectin and BM-40), thrombospondin 1 (TSP1) and 2 (TSP2), tenascin C and X, and osteopontin.

Fibroblasts secrete structural proteins such as various types of collagens, which they can deposit and remodel in the dermal defect to restore tissue integrity.17 Collagen, primarily type I along with a small amount of type III, provides tensile strength to the skin. This protein primarily accounts for the strength and durability of leather. Thus, collagen along with the imbedded and newly formed capillaries forms the granulation tissue in open wounds with tissue defects.2 However, the collagen deposited into the dermal defect in chronic wounds is disorganized and the resulting scar tissue never achieves the tensile strength of unwounded skin. Elastin, responsible for the elastic recoil of the dermal matrix, is absent from the granulation tissue that is initially deposited by resident fibroblasts. Thus, the resulting scar forms a more rigid and inelastic ECM. Restoration of a dermal matrix that mimics the unwounded structure and function, preserving the dermal integrity, would theoretically improve the resulting scar so that the tensile strength and elastic recoil would be closer to that of unwounded, intact dermis.

Fibronectin, an adhesive glycoprotein, is found in blood plasma and can be synthesized locally by resident cells.14 Fibronectin is a multifunctional protein that can act as a structural molecule owing to its fibrillar architecture, a biological glue that mediates interactions between cells and other ECM proteins, or a bridge forming cell to cell contacts.18,19 It is also found in the basement membrane zone where it provides adhesion and reinforces the structural integrity between the dermis and epidermis. Similar to elastin, there is a normal developmentally related secretion of specific fibronectin isoforms that can be induced during wound repair.20 In acute wound repair, fibronectin is initially deposited from blood plasma, whereas in chronic wound repair, it is deposited from plasma leaking across/through porous blood capillaries.21 The cellular isoforms are subsequently secreted by cells present in the wound module. Fibronectin and fibrin create a provisional matrix that promotes cell migration and adhesion, but once dermal /epidermal resurfacing is accomplished, its appearance is largely restricted to the basement membrane zone of the dermal–epidermal junction and of blood vessels. Thus, adhesive proteins like fibronectin have temporally related functions that change depending on wound status or progression to healing.

GAGs and proteoglycans are the proteins that surround or are copolymerized/deposited around other ECM proteins such as collagen and elastin. Proteoglycans are composed of polysaccharide chains attached to a protein backbone.22 These are hydrophilic molecules capable of absorbing up to 1,000 times their volume in water to form a gel like material also referred to as the ground substance. This provides dermal hydration to maintain an appropriate water balance that supports the metabolic needs of the ECM.22 Both proteoglycans and hyaluronic acid or hyaluronan have high viscosity, which leads to enhanced molecular exclusion, changes in tissue osmosis and regulation of flow resistance. Loss of ground substance can lead to tissue adhesion and decreased hydration of the matrix. High levels of hyaluronan are found in fetal ECM and its comparative overabundance in the fetus compared with collagen is responsible, in part, for scarless fetal healing.23

Matricellular proteins are secreted macromolecules that interact with cell-surface receptors, ECM, growth factors, and proteases but do not function as structural molecules per se.24 These glycoproteins regulate interactions between cells and ECM. For instance, SPARC, TSP1, and tenascin C disrupt cell–matrix interactions (i.e., are counteradhesive)—a critical component of angiogenesis and tissue remodeling.25

The importance of ECM to cellular growth and differentiation has been known for many decades. Studies in the 1960s found that the ability of muscle cells to differentiate in vitro depended on the presence of collagen.26 In the early 1980s, Bissell et al.1 postulated a mechanism by which the ECM could exert this influence by altering gene expression, and the later discovery of integrins provided confirmation of an anatomical substrate for these interactions. Subsequent research showed that the integrins consist of an extracellular portion that binds to ECM and an intracellular domain that associates with the cytoskeleton.12 The interaction of integrins with cytoskeletal components regulates cellular shape and architecture,27 which then affects intracellular signaling via MAP kinases and other mediators.12 Many intracellular signaling proteins have been linked to integrin signal transduction, including Rho GTPases, Raf, Ras, FAK, and MAP kinases.12

Degradation and remodeling of the ECM by proteases, particularly matrix metalloproteases (MMPs), is a key feature of leukocyte influx, angiogenesis, reepithelialization, and tissue remodeling. MMPs also degrade growth factors and their receptors, as well as angiogenic factors. Control of these various elements by MMPs, in part, determines whether angiogenesis will be stimulated or inhibited.28 MMPs also play an essential role in liberating growth factors and cleaving ECM proteins to reveal regions that can activate growth factor receptors (see text on matrikines in next section).29 Thus, MMPs act not only to degrade and remodel selected ECM components at appropriate times, but also to reveal selected bioactive ECM segments through targeted cleavage that ultimately influence cellular behavior.29 During wound healing keratinocytes at the wound edge begin producing MMPs as they detach from the basement membrane and migrate across the wound bed.30 Production of MMPs is regulated by cellular interactions with the matrix, as demonstrated by the ability of human keratinocytes grown on native type I collagen, but not denatured collagen or Matrigel, to express high levels of MMPs.31 These data provide another example of the ECM's regulation of the pattern/level of cellular gene expression. Although controlled production of proteases is critical to normal wound healing, chronically elevated levels of certain MMPs can lead to matrix degradation and are associated with impaired wound healing.32,33

The ECM works in conjunction with the entire cellular microenvironment to determine cellular phenotype and behavior.10 ECM interacts with growth factors in many different ways that ultimately result in mutual regulation. In the following section, we consider several major types of interactions between ECM and growth factors, focusing on examples relevant to wound healing.

TYPES OF ECM-GROWTH FACTOR INTERACTIONS

ECM binding of growth factors

ECM can bind to and release certain growth factors, thereby exerting direct control over their activity. In this way, the ECM functions as a sequestration and storage site for growth factors, concentrating their activity in the vicinity of cells and protecting them from degradation.34

An example of this in wound healing is the binding of basic fibroblast growth factor (FGF-2) to heparan sulfate35 (Figure 2). FGF-2 induces the growth of fibroblasts and endothelial cells during wound healing.36 In the early 1990s, two groups demonstrated that heparan sulfate was required for the cellular response to FGF-2.37,38 Later, it was shown that FGF molecules bind to their receptors in the presence of the proteoglycan heparin or heparan sulfate, which stabilizes two FGF receptors, each with an FGF molecule bound as a tetrameric complex39 (Figure 2). Prevention of FGF-2 binding to heparan sulfate not only prevented the ability of FGF-2 to support fibroblast growth, but also reduced binding to its cell-surface receptors.37,38 Binding of FGF-2 to heparan sulfate also imparts stability to the growth factor.40 In fact, active FGF-2–heparan sulfate complexes can be generated through proteolysis, and this release can be positively and negatively regulated by factors that affect cellular proteolytic activity.41 FGF-2 bound to heparan sulfate also exhibits prolonged activity.42 This has been demonstrated in endothelial cells, where a 10-minute exposure of cells to FGF-2 plus heparan sulfate stimulates the production of plasminogen activator to the same extent as cells continuously exposed to FGF-2 alone.42 Levels of selected FGF receptors are upregulated during wound healing43 and exposure to FGF-2 enhances healing in diabetic mice.44

Figure 2.

 Heparin-FGF interactions. FGF molecules bind to their receptors in the presence of the proteoglycans heparin or heparan sulfate. Two FGF receptors, each with an FGF molecule bound, are linked and stabilized by heparin in a tetrameric complex.39 FGF, fibroblast growth factor.

Another example of ECM-growth factor binding is the interaction between transforming growth factor-β (TGF-β) and the protein components of decorin and betaglycan. Different genes code for three TGF-β isoforms in humans: TGF-β1, 2, and 3. TGF-β1 and 2 are likely involved in scar formation, as their inhibition with neutralizing antibodies leads to reduced scar formation in animals.45 In contrast, TGF-β3 promotes reorganization of matrix molecules, resulting in an improved dermal architecture and reduced scarring.46 TGF- β1 induces the synthesis of decorin and biglycan.47 Conversely, binding of TGF-β1 to decorin, betaglycan, and biglycan inhibits its activity, suggesting a negative feedback loop.48 All three of the TGF isoforms regulate synthesis of the ECM, as well as cell growth, proliferation, and death.14,49 TGF-βs are secreted in latent forms that require activation before they can exert their activity.49 Binding to TSP-1 is one of the events that can activate latent TGF-βs.50

A third example of direct physical interactions between the ECM and growth factors is the binding of vascular endothelial growth factor (VEGF) to heparan sulfate.51 VEGF is critical for angiogenesis, controlling blood vessel formation and growth. Two of the three most widely expressed isoforms of VEGF, V145 and V189, bind to heparan sulfate, although the third widely expressed isoform, V121, does not.52 Plasmin, a natural component of wound fluid, releases V145 and V189 from their bound state, thereby enabling their endothelial cell mitogenic activity and their enhancement of vascular permeability.53

The sequestration and release of growth factors by the ECM has several consequences, including prolongation of growth factor action, localization of growth factor activity to the immediate environment, and variations in the intensity of growth factor activity (enhanced or inhibited).34 In the case of FGF-2, binding to the ECM is necessary for it to exert cellular responses. Sequestration may also protect growth factors from degradation, which may be an important role in wounds that exhibit increased protease activity.

Integrin-mediated interactions

Cell attachment to ECM required for growth factor response

With few exceptions, cells must be bound to the ECM in order to survive and grow. It has been known for at least 30 years that a key property separating tumorigenic cells from normal cells is the requirement for anchorage to ECM for survival and growth.54 Research in the late 1970s showed that incubation of cells in the absence of substratum (plastic or ECM) led to the inhibition of mRNA and protein synthesis55; it was later found that this effect was due the influence of adhesion on cell shape, which was required for the proliferation of anchorage-dependent cells.56

Integrins are the primary family of cell-surface receptors that mediate attachment to the ECM.57 Integrins are critically involved in cell survival, apoptosis, growth, and development, with phosphorylation events playing a key role in early signal transduction57 (Figure 3). Various integrin signaling pathways have been identified, involving tyrosine kinases, GTPases, and other proteins.57 Survival signals appear to involve focal adhesion kinase, with research demonstrating that if focal adhesion kinase or the correct ECM is absent, cells enter apoptosis through a p53-dependent pathway.58 Selected integrins are expressed transiently in wounds and play critical roles in healing.59 For instance, integrin binding is necessary for keratinocyte motility, as was first demonstrated with adhesion blocking antibodies.59

Figure 3.

 Fibrillar collagen binding to integrin and DDR2 sites on a fibroblast. Collagen binding stimulates production of MMP2, ECM remodeling, cellular differentiation, and migration of fibroblasts through basement membranes,76 in addition to exerting control over the cell cycle. ECM, extracellular matrix; MMP, matrix metalloproteases; DDR, discoidin domain receptor.

Angiogenesis is an example of a process that occurs during wound healing that requires adhesion of cells to the ECM in order for them to respond to growth factors. VEGF has been found to increase expression of the collagen binding integrins α1β1 and α1β2 in dermal microvasculature.60 Antibodies that block α1 and α2 integrin subunits substantially inhibit VEGF-induced angiongenesis without affecting the preexisting vasculature.60 This indicates that these integrins are essential to VEGF-induced angiogenesis. Related research has shown that the integrin αvβ3 is expressed on blood vessels in human wound granulation tissue but not in normal skin, and that antibodies against this integrin block angiogenesis induced by FGF and TNF-α without affecting preexisting blood vessels.61 Additionally, a temporal relationship between αvβ3 expression and wound angiogenesis has been identified, with this receptor initially expressed on hypertrophied microvessels and then on capillary sprouts that invade the fibrin clot; antibodies against this receptor also transiently inhibit granulation tissue formation.62 The critical nature of integrin binding to angiogenesis is also noted in other situations such as embryogenesis, where inhibition of β1 integrins interfere with the development of the embryonic vasculature63 and in oncology, where integrin inhibitors are employed to inhibit tumor angiogenesis.64

Cellular adhesion to ECM induces growth factor expression

Adhesion of cells to ECM alters gene expression.1 The maintenance of growth of stratified epidermis serves as an example of the ECM's effect on cellular proliferation in vivo: cells in contact with the basal lamina divide, whereas cells that are no longer attached migrate upward, change shape, and differentiate into the upper epidermal layers.1 Adherence-dependent effects on the expression of growth factors also characterize wound healing. For instance, monocytes must attach to ECM in order to differentiate into macrophages. This attachment leads to the upregulation of PDGF mRNA.8,65

Matrikine ligand presentation by ECM components

Matrikines have been defined as signaling elements that exist as subcomponents of ECM proteins and bind to cell surface receptors that belong to the cytokine, chemokine, ion channel, or growth factor receptor family.66 Certain matrikines, sometimes referred to as matricryptins, are hidden or inaccessible, but become manifest following conformational changes or proteolysis.67 Matrikines bind to receptors at lower affinity than growth factors, chemokines, or cytokines.68 Unlike soluble ligands, matricryptins are constrained to a surface and their influence is limited to a local region.68 In this way, only cells in the vicinity of the ligand are affected, and, in the case of cryptic matrikines, cells in the vicinity of the ligand are only affected after the relevant ECM has been degraded.68,69

The epidermal growth factor (EGF)-like repeats of tenascin-C and laminin are examples of matrikines. At these sites, tenascin-C and laminin bind to EGF receptors where they act to enhance fibroblast migration.68,70 In the case of laminin-332 (formerly laminin 5), this activity appears to occur only following cleavage by MMPs71 (Figure 4). During wound healing, tenascin C and laminin-332 are expressed by keratinocytes at the leading edge of the dermal–epidermal junction.72,73 This correlates in time with keratinocyte migration and MMP-2 expression.74 Based on this and other evidence, the binding of laminin-332 to the EGF receptor during wound healing is believed to provide promigratory tracks within the wound healing edges for migrating and proliferating keratinocytes, whereas binding of tenascin-C may enhance migration as well as deadhesion of cells.69

Figure 4.

 Laminin peptide binding to the EGF receptor. Cleavage of laminin by MMPs liberates a portion of the protein and exposes a cryptic site that can bind to the EGF receptor, where it enhances cell motility.70,68,71 MMP, matrix metalloproteases; EGF, epidermal growth factor.

Another example of ECM signaling through growth factor receptors during wound healing is the interaction of collagen with the discoidin domain receptor (DDR)75 (Figure 3). The DDR2, which binds fibrillar collagens and is expressed strongly in the dermis, has been found to regulate fibroblast proliferation and migration through the ECM.76 These events occur in conjunction with transcriptional activation of MMP-2.76 These findings indicate that collagen binding to DDR2 stimulates growth and MMP-2 transcription, which are likely critical to wound healing.

Growth factor regulation of ECM

Growth factors also regulate the ECM by stimulating cells to increase the production of ECM components or enhance synthesis of MMPs that break down ECM. In this way, the interactions between growth factors and ECM are bidirectional.

An example of growth factor regulation of ECM in wound healing is TGF-β control of ECM production and degradation77 (Figure 5). TGF-β stimulates the synthesis of collagen and fibronectin in a variety of cell lines,78,79 as well as many other ECM components including hyaluronic acid, TSP, tenascin, and others.77 TGF-β also reduces the proteolytic degradation of ECM components by reducing synthesis and secretion of proteases80 and stimulating synthesis of protease inhibitors.81 Furthermore, TGF-β has been found to increase expression of integrins that bind collagen, fibronectin, and vitronectin.82,83

Figure 5.

 TGF-β regulation of ECM. TGF-β stimulates the synthesis of collagen (shown here) and fibronectin,78,79 as well as many other ECM components including hyaluronic acid, thrombospondin, tenascin, and others.77 ECM, extracellular matrix; TGF, transforming growth factor.

Another example of growth factor regulation of ECM occurs early in the wound healing process, where release of PDGF from platelets at the wound site acts as a chemoattractant for fibroblasts.84,85 PDGF then increases the deposition of collagen by these fibroblasts in a collagen matrix.85

The ECM-growth factor interactions just described are summarized in Table 1 and Figure 6.

Table 1.   Summary and examples of ECM-growth factors interactions
DescriptionExample
  1. ECM, extracellular matrix; TGF, transforming growth factor; FGF, fibroblast growth factor; VEGF, vascular endothelial growth factor.

Growth factor binding to ECM componentsFGF-2 must be bound to heparan sulfate chains of proteoglycan to act as a mitogen37,38
Integrin-mediated interactionsIntegrins (α1α2) necessary for VEGF-induced angiongenesis60
Matrikine ligand presentation by
ECM components
Tenascin-C and laminin bind to epidermal growth factor receptors where they act to enhance fibroblast migration70,68
Growth factor regulation of ECMTGF-β controls ECM production and degradation77
Figure 6.

 Examples of ECM-growth factor interactions in a cutaneous wound. 1. Monocytes migrate to the wound site and bind to fibronectin (FN), which leads to their differentiation into macrophages that secrete multiple growth factors. 2. TGF-β binds to fibroblasts, stimulating the production of ECM components such as collagen, FN, and hyaluronic acid. 3. FGF complexed with heparan sulfate (HSP) binds to FGF receptors on endothelial cells (ECs) to induce their migration 4. Matrix metalloproteases (MMPs) cleave laminin to reveal a matrikine site that binds to epidermal growth factor receptors on fibroblasts and stimulates fibroblast migration. ECM, extracellular matrix; TGF, transforming growth factor; FGF, fibroblast growth factor.

IMPLICATIONS FOR DIFFICULT TO HEAL AND CHRONIC WOUNDS

In acute wounds, ECM-growth factor interactions appear to progress smoothly, with each phase transitioning into the next and eventually resulting in stable wound closure. However, in difficult to heal or chronic wounds, ECM-growth factor interactions are disrupted, most often due to systemic abnormalities such as diabetes or venous insufficiency. In fact, chronic wounds are characterized by fundamental biochemical abnormalities of the ECM, as well as increased levels of proteases like MMPs and neutrophil elastase21,32,86,87 (Figure 7). Diabetic and venous ulcers also exhibit abnormalities of growth factor expression,88,89 with some growth factors showing increased expression but not increased functional outcomes (e.g., increased VEGF without increased angiogenesis).90 In venous ulcers, growth factors such as TGF-β may also be trapped or scavenged by extravasated plasma proteins.91 These disruptions preclude the progression of ECM-growth factor interactions that are needed to heal wounds.

Figure 7.

 Comparison of a chronic wound in which repair is arrested and an acute wound in which repair proceeds in an orderly, sequential fashion. Differences between these wounds are seen in clot formation, inflammation, capillary migration, granulation tissue, extracellular matrix, keratinocyte migration, scar formation, bacterial colonization/infection, and biofilm formation.

ECM and growth factor abnormalities in diabetic, venous, and other difficult to heal or chronic wounds may result from underlying pathophysiological processes. Ulcers are distinguished from acute wounds by the loss of tissue from the epidermis and dermis, including both cells and ECM. Venous ulcers are caused by circulatory deficits such as microangiopathy and reduced skin perfusion.92 The skin of individuals with venous insufficiency can be fibrotic and edematous, and minor trauma or infection can result in dermal fibrosis and ulcer formation.93 The loss of ECM that characterizes these ulcers may occur due to an imbalance between matrix degrading enzymes and their inhibitors, mechanical disruption of the tissue such as physical debridement, and/or bacterial overgrowth. At the structural level, venous ulcers are characterized by fibrin cuffs, focally extravasated red blood cells, and deposits of hemosiderin.94 The fibrin cuffs are organized structures around blood vessels containing the ECM proteins fibrin, laminin, fibronectin, collagen, and tenascin, as well as trapped monocytes, macrophages, and polymorphonuclear leukocytes.94 It has been suggested that the composition and organization of fibrin cuffs is consistent with their active assembly as opposed to the passive accumulation of materials.94 Active tissue remodeling in venous ulcers is also suggested by the finding of increased matrix MMPs in these wounds.32,95

Diabetic ulcers are believed to be caused by neuropathy and subsequent trauma, impaired microvascular circulation, and/or peripheral arterial disease.96 Levels of MMP-2 and MMP-9 are elevated in the plasma of individuals with diabetes, as are levels of TIMP-1 and TIMP-2, indicating potential abnormalities in ECM metabolism.97 In a preclinical model of type-2 diabetes, the expression of MMPs and ECM proteins is increased, with diabetic rats showing differential regulation of 27 ECM genes compared with normal rats.98 Diabetes is also associated with the glycation of several ECM proteins including collagen and fibronectin, which reduces their ability to adhere cells.99 These findings indicate abnormalities in the ECM that appear to result from the disease process itself.

The ECM abnormalities that characterize difficult to heal and chronic wounds affect growth factors and vice-versa. This results in a disruption of the give-and-take interactions between these two entities and contributes to stalled healing. These interactions also have treatment implications. Wounds that are unable to synthesize a functional ECM may not demonstrate optimal benefit from the application of growth factors. Conversely, wounds with inadequate growth factor activity, due to inadequate or ill-timed synthesis, excessive degradation, or inhibition, may not benefit from the addition of ECM.

In the treatment of chronic corneal wounds, a relatively new approach that has been very successful in some of the most difficult to heal, persistent ulcers is the use of human amniotic membrane sheets that are sutured over the corneal defect.100 The amniotic membrane is thought to act like an intact, functional basement membrane on which the corneal epithelial cells can migrate and attach, provide growth factors that are present in processed basement membrane, and act as a substrate for elevated proteases in tear film that reduces protease damage to the surface of the corneal ECM.100–102 The success of this strategy in chronic corneal ulcers suggests that comparable treatments that provide both matrix and growth factors may be useful in other wound types.

The example of corneal wounds illustrates that ECM-growth factor–cell interactions are not limited to dermal wounds. Moreover, chronic corneal epithelial wounds also exhibit ECM deficits in the form of dysfunctional proteins in the basement membrane and ECM that are not well recognized by the integrins of epithelial cells.103 As a result, epithelial cells that proliferate and migrate over the corneal defect slough off following slight shearing forces by the eyelids. This repeated sloughing perpetuates the wound.

In conclusion, ECM-growth factor interactions are fundamental to all phases of wound healing and are not limited to dermal wounds. Direct physicochemical interactions with the ECM enhance or inhibit the activity of many growth factors and are required for the activity of others, including FGF-2. Indirect interactions between the ECM and growth factors occur via integrins —receptors responsible for cellular adherence. This adherence is a requisite for cellular responses to growth factors. Additionally, subcomponents of ECM proteins called matrikines can interact directly with growth factor receptors on the cell surface. Growth factors regulate the ECM by stimulating cells to increase production of ECM components or regulating production of matrix-degrading proteases and their inhibitors. Thus, ECM-growth factor interactions are bidirectional and interdependent.

Difficult to heal or chronic wounds exhibit ECM deficits and growth factor abnormalities that likely contribute to their stalled progression. Wound healing strategies that incorporate both ECM and growth factors may be beneficial for these wound types and, indeed, therapies of only one type or the other have generally proved disappointing in the clinic.104 At the very least, it seems important to recognize that interactions between these components characterize normal wound healing and these interactions have been disrupted in difficult to heal and chronic wounds.

ACKNOWLEDGMENTS

The authors wish to acknowledge the assistance of Mary Ann Chapman, PhD, in the writing of this document and Michael Schenk for creating the graphics.

Ancillary