Integrin Regulation of Epidermal Functions in Wounds

Abstract

Significance: Integrins are bidirectional signaling receptors for extracellular matrix that regulate both inside-out signaling that controls keratinocyte-mediated changes to the wound microenvironment and outside-in signaling that controls keratinocyte responses to microenvironmental changes. As such, integrins represent attractive therapeutic targets for treatment of chronic wounds or general promotion of wound healing. Advances in wound management are particularly important as the elderly and diabetic populations within the United States continue to grow.

Recent Advances: Although integrins are best known for mediating cell adhesion and migration, integrins in wound epidermis also control cell survival, proliferation, matrix remodeling, and paracrine crosstalk to other cellular compartments of the wound. Importantly, the concept of targeting integrins in the clinic has been established for treatment of certain cancers and other diseases, laying the groundwork for similar exploitation of integrins as targets to treat chronic wounds.

Critical Issues: Despite their attractiveness as therapeutic targets, integrins have complex roles in wound healing that are impacted by both their own expression and a highly dynamic wound microenvironment that determines ligand availability. Therefore, identifying relevant integrin ligands in the wound and understanding both distinct and overlapping functions that different integrins play in the epidermis will be critical to determine their precise roles in wound healing.

Future Directions: Future research should focus on gaining a thorough understanding of the highly coordinated functions of different integrins in wound epidermis, and on determining which of these functions go awry in pathological wounds. This focus should facilitate development of integrin-targeting therapeutics for treating chronic wounds.

C. Michael DiPersio, PhD

Scope and Significance

Integrins are bidirectional signaling receptors that serve as an interface between extracellular matrix and the intracellular milieu. While it is well known that integrins are the major receptors for keratinocyte adhesion to the basement membrane that underlies the epidermis in skin, roles for epidermal integrins in wound healing expand far beyond adhesion. This review will address autonomous keratinocyte functions that are regulated by integrins during wound healing, including survival and proliferation. We will also discuss roles for epidermal integrins in modulating the wound microenvironment through local matrix remodeling, as well as paracrine crosstalk to other cells within the wound.

Translational Relevance

Chronic wounds are characterized by defects in skin, including impaired re-epithelialization. In their roles as signaling receptors that regulate keratinocyte functions including proliferation and migration, integrins are attractive targets for more efficacious drug therapies to treat chronic wounds. However, exploiting integrins as therapeutic targets will require a better understanding of how they coordinately regulate epidermal functions during wound healing and how these functions are altered in pathological wounds. Recent studies shed light on the complexity of integrin signaling and extracellular matrix ligation and on the various and far-reaching roles that epidermal integrins play in wound healing.

Clinical Relevance

The ever-growing elderly and diabetic populations in the United States create a demand for more effective wound healing therapies. For these groups in particular, chronic wounds are a major impediment to overall patient quality of life. Moreover, chronic wounds are a major clinical problem that translates to a substantial financial burden to the United States as healthcare costs associated with wound management continue to rise. The development of novel therapies to restore normal epidermal functions to chronic wounds, including those that target integrins expressed on keratinocytes, is crucial to enhance patient quality of life and reduce healthcare costs.

Introduction

Basal keratinocytes in the stratified epidermis of skin are adherent to a specialized form of extracellular matrix (ECM) known as the basement membrane (BM), which physically separates the epidermis from the underlying connective tissue of the dermis. Regeneration of the BM is essential during wound re-epithelialization to restore tissue compartmentalization and provide structural support to the neo-epidermis. In addition, newly deposited BM proteins, together with ECM and matricellular proteins that appear in the wound bed, provide signals to the regenerating epidermis that are essential for normal wound healing. Importantly, altered mechanical properties or composition of the ECM are well known to contribute to the pathogenesis of chronic wounds.¹ Integrins are the major cell surface receptors for adhesion to the ECM, and they can mediate both inside-out and outside-in signal transduction pathways that control a wide variety of cell functions that contribute to both normal and pathological tissue remodeling processes, including proliferation, survival, ECM remodeling, migration, and gene expression.^2–4 In this review, we will provide an overview of what is currently known about the regulatory roles that epidermal integrins play in cutaneous wound healing. We will briefly cover what might be considered as “classical” roles that integrins play in the regulation of cell adhesion, migration, survival, and proliferation. In addition, we will discuss recently discovered roles for certain integrins in controlling the ability of the epidermis to modulate the wound microenvironment, through either effects on ECM or intercellular crosstalk to other cellular compartments. As space limitations preclude an exhaustive discussion of the numerous publications in this field, we direct the reader to several excellent reviews for further coverage of relevant topics.^5–8

Discussion of Findings and Relevant Literature

Integrins: regulators of cell adhesion and signal transduction

All members of the integrin family are heterodimeric, transmembrane glycoproteins that consist of an α and a β subunit, each with a large extracellular domain, a single-pass transmembrane domain, and a cytoplasmic domain. Most α and β subunit cytoplasmic domains are relatively short (∼20–70 amino acids), with the exception of the β₄ cytoplasmic domain (over 1,000 amino acids).³ Eighteen α subunits and eight β subunits can dimerize in different combinations to form 24 different integrins with distinct, although often overlapping ligand-binding specificities. As a group, integrins can interact with a wide variety of ECM proteins, to which they bind via their extracellular domains.³ Simultaneously, integrins interact via their cytoplasmic domains with cytoskeletal proteins to mediate a transmembrane linkage of the ECM to the cytoskeleton, which is critical for controlling cell shape, polarization, and motility.^3,4,9–11

Although integrins lack intrinsic enzymatic activity, they can interact directly or indirectly with a wide variety of signaling effectors, thereby functioning as conduits of bidirectional signal transduction across the cell membrane.^3,9,12,13 Cytoplasmic interactions can regulate the activation state of an integrin, thereby modulating its affinity for extracellular ligands during “inside-out” signaling.^3,14 In addition, integrins can regulate numerous intracellular pathways in response to ECM binding or other extracellular cues during “outside-in” signaling.¹² Indeed, as integrins are clustered at sites of cell adhesion, their cytoplasmic domains function as docking sites to recruit signaling and adaptor proteins that regulate a wide range of downstream pathways. In addition, some integrins undergo lateral interactions with other cell surface proteins, such as growth factor receptors, tetraspanins, urokinase receptor (uPAR), syndecan, or caveolin, which may occur at sites of cell-ECM adhesion or from within specialized membrane microdomains to modulate integrin signaling or adhesive functions.^2,15–23 Thus, signaling functions of an individual integrin may be regulated through subcellular localization that controls its association with distinct binding partners, or vice-versa.

The full range of signal transduction pathways that can be regulated by integrins is too extensive to discuss here and has been reviewed extensively elsewhere.^2–4,12 However, the adhesion-dependent activation of focal adhesion kinase (FAK) serves as an instructive example of outside-in integrin signaling.^24,25 Following integrin-mediated cell adhesion, FAK is auto-phosphorylated on Y397, creating a high-affinity binding site for the SRC-homology 2 (SH2) domain of SRC (or another SRC-family kinase). Once bound, SRC phosphorylates additional FAK tyrosines to create docking sites for other kinases or adaptors, such as GRB2, phosphatidylinositol 3-kinase (PI3-K), and p130CAS. In this way, the FAK/SRC complex can link integrins to downstream signaling effectors that include mitogen-activated protein kinases (e.g., ERK, JNK, and p38), certain Rho family guanosine triphophatases (GTPases; e.g., CDC42, Rho, RAC1), and the serine/threonine kinase AKT.^2,24,25

FAK activation often occurs as a result of cooperative signaling from integrins and growth factors and has been linked to several cell functions with relevance to wound healing, including formation of polarized lamellipodia, migration, proliferation, survival, and expression or activity of ECM-degrading proteases.^26–31 These functional linkages suggest that some epidermal wound functions might be controlled by the temporal appearance during wound healing of ligands for integrins that are known to activate FAK (for example, α₃β₁ and α₅β₁). Interestingly, FAK, RAC1, and integrin α₃β₁ have each been be linked to the maintenance of the epidermal stem cell compartment that resides in the hair follicle bulge,^30,32,33 and α₃β₁ binding to laminin-332 can activate FAK/SRC-to-RAC1 signaling in cultured keratinocytes,^26,34 suggesting that this signaling axis might control expansion of stem cells that contribute to skin tumorigenesis or wound re-epithelialization. However, keratinocyte-specific knockout studies revealed that while α₃β₁ and FAK each contribute to development of squamous cell carcinoma (SCC),^29,33 neither is essential for wound re-epithelialization,^29,30,35,36 indicating that elements of the wound microenvironment that are absent from cell culture may compensate for FAK signaling deficiencies in vivo. Indeed, it should be noted that many other integrin-linked signaling molecules, such as integrin-linked kinase (ILK) and phospholipase C (PLC), have also been shown to play important roles in the epidermis, as reviewed elsewhere.²⁸

Epidermal integrins

The stratified epidermis of the skin is a continually regenerating tissue wherein the loss of dead keratinocytes from the outermost layer is balanced by keratinocyte proliferation in the basal layer.^37–39 In normal resting epidermis, integrin expression is restricted to the basal keratinocytes that are attached to the BM, and it is down-regulated as differentiating keratinocytes detach from the BM and are displaced upwards into the suprabasal layers.³⁷ In the epidermis of a healing wound, keratinocytes acquire new adhesion/migration properties as a result of either changes in integrin expression or altered adhesion/signaling functions of integrins that were already expressed prior to wounding. Constitutively expressed integrins that persist or are upregulated in wound epidermis include α₃β₁ and α₆β₄ (both laminin-332 receptors), α₂β₁ (a collagen receptor), α₉β₁ (a receptor for cellular fibronectin, tenascin, and other ligands), and α_vβ₅ (a vitronectin receptor).^37,40 Integrins that are expressed de novo in healing wounds include α₅β₁ (a fibronectin receptor) and α_vβ₆ (a fibronectin and tenascin receptor). These epidermal integrins can bind numerous ligands present in the provisional wound ECM, including fibronectin (α₅β₁, α₉β₁, α_vβ₆), vitronectin (α_vβ₅), and tenascin (α₉β₁, α_vβ₆), as well as laminin-332 (α₃β₁, α₆β₄) that is deposited by migrating keratinocytes.^8,37,40,41

Numerous studies in both human and rodent models have documented the various expression patterns and functions of individual integrins in the epidermis during skin development, in adult skin, and during wound repair.^8,37 Of course, there is a need for caution when extrapolating results obtained in murine models to human wound healing, as there are species-specific differences that include hair follicle density, skin thickness, and relative importance of wound contraction. Nevertheless, many aspects of human wound healing are recapitulated in murine models.⁴² Moreover, many integrin functions have been evolutionarily conserved in mice and humans, and genetic studies in global or epidermis-specific knockout mice have been instructive in determining the relative importance of individual integrins for specific epidermal functions in both resting skin and during wound healing.⁸ The general importance of β₁ integrins is illustrated by the phenotypes of mice that harbor epidermis-specific null mutations in the Itgb1 gene (encoding the integrin β₁ subunit), which display an array of defects that includes impaired proliferation, loss of sebaceous glands and hair follicles, organizational defects in the BM, and impaired wound re-epithelialization.^43–45 Importantly, knockout of individual β₁ integrins through null mutation of specific α subunits leads to only a subset of the β₁-null phenotypes in each case, indicating unique functions for different integrins. Nevertheless, some integrins have overlapping functions, such as in the regulation of epidermal migration (see Epidermal migration). Interestingly, while early studies in cultured keratinocytes implicated β₁ integrins in the control of cell differentiation, the effects of either epidermis-specific β₁ subunit deletion or knockout of individual integrins (α₃β₁, α₆β₄, α₂β₁, α₉β₁, or α_vβ₅) on epidermal stratification and differentiation were mild or absent,^{43,44,46–51} indicating that adhesion-dependent regulation of keratinocyte differentiation in vivo is not dependent on any particular integrin(s).

In the sections that follow, we will discuss what is currently known about the roles of individual integrins in the regulation of distinct epidermal/keratinocyte functions, many of which have been elucidated through studies utilizing both cell culture and genetic models. Expression patterns, potential ligands, and known functions of individual integrins in the unwounded and wounded epidermis are summarized in Table 1 and described in detail in the corresponding subsections. However, it is important to consider that the repertoire of integrins expressed on wound epidermis suggests complex interactions, where interplay between different integrins with synergistic or opposing effects may be necessary for proper regulation of keratinocyte wound functions. This concept is supported by observations that different integrin ligands present in the wound bed can have either cumulative or dominant effects on keratinocyte function.^{34,41,52–56} Thus, the precise temporal and spatial control of integrin–ECM engagement, perhaps dictated by the temporal appearance or accessibility of ECM ligands in the wound bed, is likely to be critically important for coordinate regulation of different integrins during wound healing.

Table 1.

Epidermal integrins, their known ligands, and summary of their known functions in unwounded or wounded skin

Integrin	ECM Ligands in Skin	Known Functions in Unwounded Epidermis	Known Functions in Wound Epidermis
α3β1	Laminins (LN-332 and LN-511)	Expressed constitutively. Required for BM assembly during skin development and epidermal-dermal adhesion in neonates.	Required for stability of nascent BM and epidermal-dermal adhesion. Promotes wound angiogenesis through crosstalk to vasculature. May negatively control epidermal migration.
α6β4	Laminins (mainly LN-332 in hemidesmosomes)	Expressed constitutively. Function within hemidesmosomes is essential for epidermal-dermal adhesion.	Presumably required for hemidesmosome assembly and epidermal adhesion following re-epithelialization. Roles in epidermal migration are unclear.
α9β1	FN, TN, OPN, VEGF, TSP, ADAMs, EMILIN1, and others	Expressed constitutively at low levels. Nonessential for epidermal development or adhesion.	Required for normal keratinocyte proliferation during wound re-epithelialization.
α2β1	Collagens	Expressed constitutively. Nonessential for epidermal development or adhesion.	No essential roles identified. May contribute to epidermal migration over collagen.
α5β1	FN (via RGD)	Expressed at very low levels. Nonessential for epidermal development or adhesion.	No essential roles reported. May contribute to epidermal migration over fibronectin.
αvβ6	FN, TN, LAP of TGFβ-1 and -3 (via RGD)	Not expressed in interfollicular epidermis. Nonessential for epidermal development or adhesion, although required in juvenile mice for normal hair growth.	No essential roles identified, although may promote keratinocyte-mediated activation of latent TGFβ. Also implicated in ECM proteolysis and keratinocyte survival.
αvβ5	VN (via RGD)	Expressed at very low levels. Nonessential for epidermal development or adhesion.	No essential roles identified; may contribute weakly to keratinocyte-mediated activation of latent TGFβ.

See text for expanded discussions and supporting literature.

LN, laminin; BM, basement membrane; FN, fibronectin; TN, tenascin, OPN, osteopontin; VEGF, vascular endothelial cell growth factor; TSP, thrombospondin; ADAM, a disintegrin and metalloproteinase; EMILIN, elastin microfibril interfacer; RGD, arginylglycylaspartic acid; LAP, latency-associated proteins; TGFβ, transforming growth factor beta; ECM, extracellular matrix; VN, vitronectin.

Finally, it is worth noting there are compelling similarities between wound healing and development of squamous cell carcinoma (SCC),⁵⁷ and that some integrin expression patterns in SCC mirror those that occur in cutaneous wound healing.^37,40 Consistently, integrins regulate a number of epithelial cell functions that are important in both processes, including migration, proliferation, survival, matrix remodeling, and the production of pro-angiogenic factors. Therefore, we will occasionally refer to studies of certain integrins in SCC or other carcinomas, where results from these models may offer clues regarding integrin expression or function during wound healing.

The laminin-binding integrins (α₃β₁ and α₆β₄)

Laminin-332, the main adhesive ligand in the resting epidermis, is composed of three distinct chains designated α₃, β₃, and γ₂.^58,59 The effects of laminin-332 on keratinocyte behavior are mediated mainly through its two integrin receptors, α₃β₁ and α₆β₄,^41,54–56 although other receptors such as syndecan-1 also contribute.⁶⁰ Integrin α₆β₄ is a major transmembrane component of hemidesmosomes, which are the intermediate filament-associated structures on the basal surfaces of keratinocytes that anchor the epidermis to the dermis.^10,61 In this context, α₆β₄ is thought to play primarily a structural role with little or no signal transduction function.⁸ In contrast, α₃β₁-mediated adhesions are associated with the actin cytoskeleton and manifest as focal adhesions in cultured keratinocytes.^10,61 Moreover, α₃β₁ can initiate adhesion-dependent signal transduction in keratinocytes through activation of FAK, SRC, and other signaling molecules.^26,34,62 As receptors for a common BM ligand, it seems likely that α₃β₁ and α₆β₄ function coordinately (either cooperatively or in opposition) with regard to some keratinocyte processes. However, since these two integrins differ considerably with regard to both cytoskeletal interactions and signaling capacity, their combined effects on keratinocyte function are likely to occur through convergence of their distinct functions, rather than as redundant functions. Correspondingly, there is evidence that distinct proteolysis of laminin-332 (LN-332) may differentiate between α₃β₁ and α₆β₄ binding. For example, an N-terminal proteolytic cleavage fragment of the full-length laminin α₃B chain can preferentially ligate integrin α₃β₁.⁶³ Moreover, laminin-α₃ chain processing within the C-terminal globular domains appears to modulate whether LN-332 localizes to α₆β₄-containing hemidesmosomes or α₃β₁-mediated adhesions, suggesting differential integrin interaction.⁶⁴

Although the epidermis stratifies normally in mice that lack α₃β₁ and α₆β₄, either alone or in combination, these integrins play essential but distinguishable roles in maintaining the epidermal–dermal junction through their interactions with laminin-332 in the BM.^{50,51,65–67} Indeed, human gene mutations in the relevant integrin subunits, or in the individual subunits of laminin-332 itself, lead to variants of the human blistering skin disease junctional epidermolysis bullosa (JEB). However, differences in the details of these JEB variants, which are mirrored in the corresponding knockout mice, illustrate the very different roles that α₃β₁ and α₆β₄ play in epidermal–dermal adhesion. Indeed, the essential adhesion role for α₆β₄ in hemidesmosomes is revealed by the extensive epidermal blistering that occurs in mice that harbor a null mutation in either the Itga6 or Itgb4 gene (encoding the α₆ or β₄ subunit, respectively),^65–67 and in human patients with loss-of-function mutations in α₆β₄.^68,69 In contrast, the absence of α₃β₁ (through null, or loss-of-function, mutation of the Itga3 gene encoding the α₃ subunit) causes much less severe epidermal blistering in newborn mice⁵⁰ or young patients,⁷⁰ and hemidesmosomes are left intact in non-blistered regions of α₃β₁-deficient skin. Moreover, blisters in α₃β₁-deficient skin form via a mechanism of BM rupture that is distinct from loss of epidermal attachment to laminin-332 per se, which persists in these mice via α₆β₄.^50,51

α₆β₄

Hemidesmosomes are essential for stable epidermal adhesion, and their dissolution is necessary for epidermal migration during wound healing.¹⁰ Hemidesmosome disassembly involves the SRC-mediated phosphorylation of the β₄ cytoplasmic domain, possibly triggered by endothelial growth factor (EGF) or other signals from the wound microenvironment.⁷¹ In this sense, the function of α₆β₄ from within hemidesmosomes can be viewed as a brake to epidermal migration that must be released for wound re-epithelialization to proceed. Following wound healing, hemidesmosomes are re-assembled to restore stable adhesion of the neo-epidermis.¹⁰

While the hemidesmosomal role for α₆β₄ in stable epidermal adhesion is well established, whether this integrin also regulates signal transduction pathways that control keratinocyte functions (e.g., survival, proliferation, migration) is less clear.⁸ Indeed, pro-migratory/pro-invasive roles for integrin α₆β₄ have been described for carcinoma cells, where this integrin may cooperate with receptor tyrosine kinases such as EGFR, RON, and MET to activate migration-promoting signals, as reviewed elsewhere.⁵ However, the extent to which these roles extend to keratinocyte migration during wound re-epithelialization remains uncertain. Noteably, some studies do support a pro-migratory role for α₆β₄ in keratinocytes, possibly through antagonism of α₃β₁⁷² and/or through regulation of laminin-332 matrix deposition that directs keratinocyte migration.⁷³

α₃β₁

In vivo and in vitro studies have identified several roles for integrin α₃β₁ in regulating keratinocyte functions that may contribute to wound healing, including ECM deposition and organization,^74,75 cell polarization and migration,^34,53 cell survival,⁶² cell proliferation,⁷⁶ and secretion of ECM proteases and pro-angiogenic factors.^36,77–79 Many of these processes depend on α₃β₁ binding to laminin-332, although some may involve ECM-independent functions of α₃β₁ that are modulated by lateral interactions (direct or indirect) with other cell membrane proteins, such as tetraspanins,^77,80 uPAR,⁸¹ or adherens junction proteins.^82,83 For example, α₃β₁ binding to the tetraspanin CD151⁸⁴ can regulate α₃β₁-mediated signaling⁸⁵ and epithelial cell motility.⁸⁶ Interestingly, CD151 is upregulated in wound epidermis, and deletion of CD151 leads to wound healing defects,⁸⁷ suggesting that CD151 might modulate α₃β₁ function during wound repair.

Roles for α₃β₁ in keratinocyte migration have been controversial. On one hand, studies using either human or murine keratinocytes support a role for α₃β₁ in regulating front–back polarization and promoting processive migration on laminin-332,^34,53 which occurs in part through regulation of a FAK/SRC–RAC1 signaling axis.^26,34 In contrast, other studies report that α₃β₁-deficient keratinocytes display enhanced motility and directional migration, although in some cases this appears to involve compensatory upregulation of other integrins such as α₆β₁.^35,74 These discordant results may stem in part from differences in the cell culture models used, including species from which cells were derived and the differential deposition of ECM ligands for other integrins.^{41,52,54–56} Indeed, α₃β₁ may been have trans-dominant inhibitory effects on other keratinocyte integrins.^35,36,88,89 In any case, in vivo studies performed by separate groups using similar epidermis-specific α₃-knockout models have shown that α₃β₁ is not essential for efficient wound closure^35,36 and may even slightly inhibit re-epithelialization, at least in adult mice.^35,36 Interestingly, absence of α₃β₁ reduced efficient wound closure in skin grafts from neonatal α₃-null mice,⁹⁰ which might suggest a role in wound re-epithelialization at earlier developmental stages, or that compensatory mechanisms present in adult skin are absent from neonatal skin. Alternatively, absence of α₃β₁ from other cellular compartments may have caused reduced wound closure in the latter study, since skin grafts were derived from mice with a global α₃-null mutation.⁹⁰ These questions might be resolved in future studies using mice with compound, epidermis-specific deletion of multiple integrins.

While the dispensability of α₃β₁ for epidermal migration was unexpected, in vivo and cell culture studies have revealed other roles for this integrin in allowing the epidermis to modulate the tissue microenvironment; for example, through alterations of the ECM or paracrine crosstalk to other cellular compartments. Indeed, epidermis-specific deletion of α₃β₁ was associated with reduced wound angiogenesis, and α₃-null keratinocytes showed reduced secretion of factors that stimulate endothelial cell migration.³⁶ In addition, our group recently identified a novel role for α₃β₁ in promoting the stability of nascent BM that forms during wound healing, as we discuss later (see Local basement membrane deposition and assembly).⁹¹ Moreover, α₃β₁ has been shown to control laminin-332 deposition and organization in cultured keratinocytes,⁷⁴ and we recently observed that proteolytic processing of the laminin-γ₂ chain is delayed in wounds of α₃β₁-deficient epidermis and impaired in α₃-null keratinocytes.⁹¹

Other β₁ integrins (α₉β₁, α₂β₁, and α₅β₁)

Integrin α₉β₁ is expressed constitutively in the epidermis and is upregulated during wound healing, where it may bind to several ECM proteins or other ligands in the wound, including the EIIIA/EDA segment in cellular fibronectin, tenascin, EMILIN1, osteopontin, thrombospondin, certain ADAM (a disintegrin and metalloprotease domain) family members, and VEGF (vascular endothelial growth factor).^8,92,93 Distinct binding motifs for α₉β₁ have been identified, including the AEIDGIEL in tenascin-C, the PEDGIHE motif that occurs within an exposed loop of the fibronectin EIIIA domain, and distinct motifs within ADAM family members.^93,94 While α₉β₁ has been studied less extensively than other integrins, its roles are likely to be complex given the wide range of potential ligands for this integrin that are present in wounds. Genetic studies have revealed an important role for α₉β₁ in wound re-epithelialization. Indeed, keratinocyte proliferation was significantly impaired in wounds of mice with epidermis-specific deletion of α₉β₁, while the rate of re-epithelialization was unaffected, resulting in diminished thickness of the neo-epidermis that was presumably due to a reduced number of migrating keratinocytes.⁴⁹ Thus, α₉β₁ contributes to establishing integrity of the neoepidermis by promoting proliferation of wound keratinocytes, rather than by driving epidermal migration per se.

The collagen-binding integrin, α₂β₁, is also expressed constitutively in epidermis and upregulated in wounds. Remarkably, however, mice with a global α₂-null mutation display normal skin development with no obvious defects in wound re-epithelialization or contraction, indicating that this integrin is not essential for wound closure.^47,95 However, α₂β₁-deficient mice did display enhanced angiogenesis, reduced infiltration of mast cells, and (in one study) reduced tensile strength of healed wounds, indicating complex roles in wound healing.^47,48 While many of the wound healing defects in the α₂-null mice are probably caused by absence of α₂β₁ from the involved cells, it remains to be determined whether some of these defects reflect the loss of integrin-dependent crosstalk between different cellular compartments of the wound, as mentioned above for α₃β₁. Moreover, there is the possibility that loss of α₂β₁–collagen interactions that promote keratinocyte migration are compensated by other integrin–ECM interactions. Future studies in which α₂β₁ is deleted from epidermis, alone or in combination with other integrins, may address these questions.

Integrin α₅β₁ is largely absent from resting epidermis and expressed de novo in wound epidermis, where it presumably contributes to epidermal migration over fibronectin in the provisional matrix of the wound.⁹⁶ However, in vivo roles for epidermal α₅β₁ during wound healing are not defined, since the α₅-null mutation is embryonic lethal,⁹⁷ and studies in epidermis-specific α₅ knockout mice have not been reported. Thus, it remains possible that loss of α₅β₁ accounts for much of the epidermal migration defect that was reported in mice lacking β₁,⁴⁵ since other α subunit knockouts did not phenocopy this defect.

The α_v integrins (α_vβ₅ and α_vβ₆)

The integrins α_vβ₅ and α_vβ₈ are expressed at low levels in resting epidermis, with the latter being expressed suprabasally, while α_vβ₆ is restricted to hair follicle stem cells and is not normally expressed in the interfollicular epidermis of unwounded skin.^8,98–100 However, both α_vβ₆ and α_vβ₅ are upregulated during wound healing,^8,98–100 where each may bind to several ECM ligands through the arginine-glycine-aspartate (RGD) motif (e.g., fibronectin, tenascin, vitronectin).^3,40 Given the diversity of ligands that α_v integrins can bind, their upregulation in wound epidermis may be important for modulating keratinocyte motility over the complex matrix of the wound bed. Moreover, α_vβ₆ has been shown to regulate the expression of extracellular proteases, including matrix metalloprotease (MMP)-9,¹⁰¹ MMP-3,¹⁰² and uPA,¹⁰³ suggesting roles in ECM remodeling. Finally, as discussed below (in epidermal survival section), α_vβ₆ might promote cell survival and prevent keratinocytes from differentiating.

The de novo induction of α_vβ₆ during wound healing might also provide temporal control over local transforming growth factor-β (TGFβ) activation, since this integrin can activate the ECM-bound pool of latent TGFβ.^40,104,105 Each of the mammalian TGFβ isoforms (TGFβ-1, 2, and 3) is secreted as an inactive complex consisting of the latency-associated protein (LAP) and the latent TGFβ binding protein (LTBP), which is covalently linked to the ECM through fibronectin.^104,106 Integrin α_vβ₆ (and perhaps to a lesser extent α_vβ₅) can bind to the RGD motif within the LAP, resulting in a conformational change in the complex that activates latent TGFβ₁ or TGFβ₃.^104,105,107 Although this mechanism has been best characterized in carcinoma cells, it also occurs in keratinocytes and may initiate α_vβ₆-dependent TGFβ signaling pathways that are important for wound healing.¹⁰⁵

Functions of wound epidermis that are regulated by integrins

A number of diverse keratinocyte functions are regulated by adhesion to the ECM, and epidermis-specific gene knockout studies have revealed roles for individual integrins in distinct keratinocyte functions that are important for normal wound healing, including migration, proliferation, ECM remodeling/BM regeneration, stable attachment of the neo-epidermis, and induction of angiogenesis. In some cases, however, there is considerable discordance between results from cell culture studies and those from in vivo studies. For example, studies in cultured keratinocytes have clearly implicated certain integrins in the control of cell migration, while studies in knockout mice have so far failed to identify individual integrin–ECM interactions that are essential for wound re-epithelialization in adult skin (discussed below).

The complex process of wound healing is further regulated by growth factors, some of which work in concert or collaboration with integrins to control epidermal functions. Two growth factors of particular importance in wound healing, VEGF and TGFβ,¹⁰⁸ have each been functionally linked to integrins that are expressed in wound epidermis. For instance, α₆β₄ was shown to enhance VEGF translation in carcinoma cells,¹⁰⁹ and VEGF expression can upregulate α_vβ₆ in some carcinoma cells,¹¹⁰ raising the intriguing possibility of dynamic integrin–growth factor feedback loops that might extend to wound keratinocytes. Additionally, α_vβ₆-mediated activation of latent TGFβ (see previous section) might stimulate the expression of α₅β₁ and α_vβ₅ in keratinocytes.¹¹¹ However, while TGFβ stimulates pro-migratory integrins, it has also been shown to inhibit keratinocyte proliferation, so there are conflicting reports with regard to the net effect of TGFβ signaling on re-epithelialization.¹¹² In any case, it is clear that the repertoire of integrins expressed in wound epidermis, in combination with growth factors present in the wound microenvironment, is important for coordinating diverse keratinocyte functions that collectively ensure efficient wound repair and epidermal regeneration.

In the following sections, we will discuss diverse functions of the wound epidermis that are regulated by keratinocyte integrins (summarized in Fig. 1). We will begin with a discussion of the roles that integrins play in autonomous keratinocyte functions that promote wound healing, followed by a discussion of more recently appreciated roles for certain epidermal integrins in the regulation of paracrine crosstalk to other cell types in the wound.

Figure 1.

A model depicting functions of wound epidermis that may be controlled by keratinocyte integrins. Arrow (1) indicates autonomous regulation of keratinocyte functions such as migration, proliferation, local matrix remodeling, and stable adhesion of the neo-epidermis. Arrows (2) and (3) indicate regulation of paracrine crosstalk from the epidermis to other cellular compartments, including myofibroblasts to control wound contraction (2) and vascular endothelial cells to promote wound angiogenesis (3). Components of the wound microenvironment are indicated. BM, basement membrane; HF, hair follicle; ECM, extracellular matrix. To see this illustration in color, the reader is referred to the web version of this article at www.liebertpub.com/wound

Epidermal migration

Given their global importance in cell adhesion and motility, it seems intuitive that a major function of integrins during wound healing is the regulation of epidermal migration. Consistently, deletion of β₁ integrins from the epidermis resulted in severely compromised wound re-epithelialization,⁴⁵ and numerous cell culture studies have shown that individual integrins can regulate keratinocyte adhesion and migration on their respective ECM ligands.^{34,45,53,61,73,113,114} Nevertheless, results from in vivo wound healing studies in integrin knockout mice are surprisingly discordant with findings from cell culture models, and the roles for individual integrins in epidermal migration remain unclear. For example, α_vβ₆ and α_vβ₅ can each mediate keratinocyte migration on their respective ligands in vitro, yet absence of either or both integrins in vivo did not cause impaired wound healing in young adult mice (although deletion of α_vβ₆ caused delayed wound healing in older mice).^46,115,116 Similarly, integrins α₃β₁, α₂β₁, and α₉β₁ can each mediate keratinocyte migration on relevant ligands in vitro, but their ablation in vivo had remarkably mild or no effects on wound re-epithelialization in adult mice.^35,47–49 As already mentioned, in a striking example of discordance, α₃β₁ has been reported to promote migration of cultured keratinocytes on laminin-332,^34,53 yet it is not essential for wound closure in vivo, at least in adult mice.³⁵ In fact, deletion of α₃β₁ from the epidermis led to a slightly increased rate of wound closure,³⁵ consistent with other studies using anti-α₃β₁ blocking antibodies that indicated a suppressive role in migration.⁸⁹ Together, these disparate findings indicate a complex role for α₃β₁ in epidermal migration that is sensitive to ECM composition and may include modulation of the rate of re-epithelialization.

In some cases, these discrepancies between cell culture models and in vivo wound models may reflect the greater complexity of ECM in vivo, where multiple ligands in the wound bed might provide opportunity for compensation between integrins with overlapping migration functions, as discussed at length elsewhere.⁸ Such overlap makes sense from an evolutionary standpoint, given the importance of rapidly restoring epidermal barrier function after wounding. Consistent with this idea, mice with epidermis-specific deletion of all β₁ integrins (i.e., β₁-null) showed impaired wound re-epithelialization,⁴⁵ while deletion of individual integrins did not,^8,35,47–49 suggesting that there is considerable overlap in the ability of distinct β₁ integrins to promote epidermal migration.

Although α₆β₄ can regulate the motility of cultured keratinocytes,^72,73 it remains uncertain whether this integrin has important pro-migratory roles in wound re-epithelialization. It is clear that α₆β₄ is not sufficient for normal wound re-epithelialization in the absence of β₁ integrins, since β₁-null epidermis shows severe migration defects.⁴⁵ Nevertheless, it remains possible that some pro-migratory/pro-invasive roles for α₆β₄ that have been described in carcinoma cells (mentioned previously) may extend to wound keratinocytes. In any case, a role for α₆β₄ in epidermal migration cannot yet be ruled out, since the severe epidermal blistering that occurs in α₆β₄-deficient mice presents a formidable challenge to wound healing experiments. More detailed discussions of potential roles for α₆β₄ in keratinocyte migration can be found in separate reviews⁸ (also, see the review by Hopkinson et al.¹¹⁷ in this issue).

Epidermal proliferation

Epidermal homeostasis is maintained by a resident population of stem cells within the basal cell layer that gives rise to committed progenitor cells, or transit-amplifying cells, which in turn replenish the differentiated keratinocytes that are eventually shed from the outer layer of the skin.^37–39 While most of the proliferative cells that maintain homeostasis are thought to reside in the interfollicular epidermis, stem cells in the bulge of the hair follicle make the most substantial contribution to the replenishment of keratinocytes that accompanies wound re-epithelialization.^118,119 β₁ integrins as well as α₆β₄ are expressed at higher levels in epidermal stem cells, where they are thought to help control the balance between stem cell renewal and terminal differentiation.^{5,37,39,120–123} It follows that changes in integrin signaling or adhesion functions that shift this balance are likely to be important for wound healing. Consistently, integrin signaling through mitogen-activated protein kinases (MAPKs) or the Rho family GTPase, RAC1, has been linked to the maintenance of epidermal stem cells.^32,33,123 As discussed in the section on other β₁ integrins, integrin α₉β₁ is essential to maintain proper levels of keratinocyte proliferation during wound healing.⁴⁹ Interestingly, a recent study in an SCC model revealed a novel role for α₃β₁ in the retention of the slow-cycling cells in the hair follicle bulge, where deletion of this integrin allowed these cells to detach from the BM and emigrate into the suprabasal layers where they terminally differentiated. This loss of slow-cycling cells resulted in reduced skin tumor formation, presumably because these are the same cell that would otherwise accumulate oncogenic or tumor suppressor mutations that eventually give rise to tumors.¹²⁴ These findings might reflect similar functions for α₃β₁ or other integrins within the hair follicle stem cells that participate in epidermal regeneration during wound healing.

Epidermal survival

Presumably, wound keratinocytes must somehow maintain appropriate survival signals that prevent them from undergoing anoikis (i.e., apoptosis that is induced by reduced or inappropriate adhesion) as they lose contact with damaged BM and encounter the dramatically altered ECM of the wound bed. As mentioned above, the de novo expression of α_vβ₆ may prevent wound keratinocytes from undergoing anoikis, as this integrin is known to activate AKT survival pathways in SCC cells.¹²⁵ In addition, α₃β₁-mediated adhesion has been shown to promote the survival of keratinocyte cell lines through pathways that involve FAK and ERK, suggesting a possible role in survival of wound keratinocytes.⁶² Adhesion through α₆β₄ has also been linked to epidermal cell survival,⁶⁷ and it is conceivable that this integrin contributes to adhesion-dependent survival of the neo-epidermis when hemidesmosomes are assembled following wound closure. However, mice lacking both α₃β₁ and α₆β₄ (e.g., through combined α₃-null and β₄-null mutations) displayed apoptotic basal keratinocytes only in detached regions of epidermis, but not in non-blistered epidermis,⁵¹ indicating that other adhesion mechanisms are sufficient to protect from anoikis.

Local basement membrane deposition and assembly

The cutaneous BM not only provides a physical separation between the epidermal and dermal compartments of the skin, but it also provides cues for signaling pathways that regulate a variety of epidermal cell functions including polarization, differentiation, survival, tissue structure, and migration. Indeed, many of the keratinocyte functions discussed above are regulated by keratinocyte-mediated changes to the local ECM, including the BM, which as discussed here may occur through either deposition of ECM proteins or matrix proteolysis. Early studies in β₁ integrin-deficient mice or embryoid bodies have revealed essential roles for certain β₁integrins in proper BM formation in embryonic or adult tissues.^50,126,127 In vivo and in vitro studies from our group and others collectively support roles for integrin α₃β₁ in several steps that may be important for the regeneration of a stable BM during wound healing, providing an example of how some integrins may regulate the ability of the epidermis to regulate ECM organization and assembly (Fig. 2). Indeed, as already mentioned, α₃β₁ regulates laminin-332 organization in cultured keratinocytes.⁷⁴ Moreover, we recently demonstrated that mice lacking epidermal α₃β₁ display blisters in newly regenerated wound epidermis,⁹¹ which are characterized by BM rupture that phenocopies the epidermal blisters in neonatal α₃-null mice,⁵⁰ indicating that the developmental role for α₃β₁ in maintaining BM integrity is recapitulated during adult wound healing. Interestingly, this blistering phenotype was linked to reduced fibulin-2 levels in both neonatal skin and early wounds, and genetic deletion of fibulin-2 (i.e., in the presence of α₃β₁) was sufficient to generate neonatal skin blisters.⁹¹ Since fibulin-2 binds to the laminin-γ₂ chain within the L4 module near the N-terminus,¹²⁸ and this interaction has been implicated in stable incorporation of laminin-332 into the developing BM,¹²⁹ these findings suggest an important role for α₃β₁-dependent fibulin-2 expression in BM maturation (Fig. 2).

Figure 2.

A model depicting possible roles for epidermal integrin α₃β₁ in regeneration of BM during wound healing. In vivo and cell culture studies have implicated α₃β₁ in several different steps that may impact the stable assembly of laminin-332 (LN-332) into BM, including (1) LN-332 deposition, (2) expression of the LN-332-binding protein fibulin-2 (Fbln2), and (3) proteolytic processing of the LN-332 γ₂ chain (see text for detailed discussion and supporting references). Other epidermal integrins may have similar roles in the incorporation of LN-332 or other BM proteins. The α₃, β₃, and γ₂ chains of LN-332 trimer, as well as the L4 module on the γ₂ chain, are indicated. To see this illustration in color, the reader is referred to the web version of this article at www.liebertpub.com/wound

Integrins can also control matrix proteolysis by the epidermis, as supported by the ability of several integrins to control the expression or functions of MMPs or other ECM-degrading extracellular proteases (discussed below). While a variety of ECM proteins present in the provisional wound matrix undergo proteolysis, we will focus on laminin-332 as an example since both its α₃ and γ₂ chains undergo extensive proteolytic processing.^130–133 Although the importance of these processing events in vivo are not yet fully understood, some have been shown to influence the migratory behaviors of cultured keratinocytes. For example, differential processing of the laminin α₃ chain, regulated by tissue-type plasminogen activator and plasmin, was shown to alter epithelial cell motility.^64,134 Laminin-332 proteolysis is also likely to be an important mode of regulating BM architecture through modulation of key linkages with other ECM proteins, such as the interaction of fibulin-2 with laminin-332 described above.⁵⁸ Indeed, the L4 module to which fibulin-2 binds is lost upon proteolytic processing of γ₂, which might be an important step for BM maturation.^128,129,132 Interestingly, recent data from our laboratory indicates a role for epidermal α₃β₁ in the regulation of laminin-γ₂ chain processing, both in vivo during wound healing and in keratinocytes cultured under conditions of high calcium.⁹¹ Thus, α₃β₁ may have a dual role in BM assembly that involves both matrix protein deposition and proteolytic processing of matrix proteins (Fig. 2).

Epidermis-mediated modulation of the wound microenvironment

Most studies of keratinocyte integrins have focused on their “classical” roles in regulating cell-autonomous processes such as motility, proliferation, local matrix assembly/remodeling, and survival, as discussed above. However, epidermal integrins may have far-reaching effects on other aspects of the wound microenvironment, including modification of the ECM in distal parts of the wound, and crosstalk to distal cells with essential roles in wound healing (i.e., endothelial cells, fibroblasts). Consistently, some keratinocyte integrins can modulate the expression of genes that are implicated in paracrine crosstalk to other cells.^{36,47,135,136} Indeed, our own studies have identified α₃β₁-dependent regulation of a number of genes in keratinocyte cell lines that encode growth factors, extracellular proteases, or ECM/matricellular proteins with known roles in modulating the microenvironment of wounds or tumors^36,79,137 (Missan and DiPersio, unpublished observation). In this section, we will discuss these more recently appreciated roles for certain epidermal integrins in modification of the wound microenvironment, including paracrine functions that influence extra-epidermal compartments of the wound.

Proteolysis and remodeling of the wound extracellular matrix

While the effects of local matrix remodeling on epidermal functions are clear, long-range effects of altered wound ECM are also important. Indeed, controlled ECM remodeling is an essential feature of normal wound healing,¹³⁸ and defects in ECM organization are associated with chronic wounds.¹³⁹ It has long been known that changes in matrix composition or structure can act upstream of integrins through ligation that turns on intracellular signaling pathways. However, integrin–ECM signals should more accurately be considered as bidirectional, or even as signaling loops, as many integrins have been shown to alter ECM composition/structure at the levels of both gene expression and matrix assembly. For example, integrin α₂β₁ has been shown to be a positive regulator of type-1 collagen gene expression,¹⁴⁰ while α₁β₁ suppresses collagen synthesis in the dermis.¹⁴¹ Furthermore, both fibronectin-binding integrins and ILK have been shown to support fibronectin matrix assembly and fibrillogenesis.^142,143

The dynamic regulation of MMPs and other extracellular proteases can also contribute to changes in ECM that are crucial for successful wound healing. Indeed, MMPs are involved in all stages of wound resolution, from early removal of damaged ECM to late stage scar remodeling,¹⁴⁴ and defects in MMP expression or function contribute to the pathogenesis of chronic wounds.^139,145 Furthermore, MMPs are increasingly credited with the proteolytic release of important growth factors from the ECM or cell surface,^144,146 thereby supplying the wound microenvironment with the appropriate effector molecules to promote healing. Not surprisingly, several integrins that are expressed during wound healing (Table 1) have been implicated in modulating MMP expression in keratinocytes or other cells, suggesting similar roles in the wound microenvironment. Examples include α₃β₁-mediated induction of MMP-9 or uPA in keratinocytes,^79,137 α₂β₁-mediated induction of MMP-1 in osteogenic cells,¹⁴⁰ and α₅β₁-mediated regulation of MMP-3 and MMP-9 in some cells.¹⁴⁷

Paracrine crosstalk to other cellular compartments

It is well known that a complex network of communication exists between the various cell types that reside in the wound microenvironment. The epidermis can send paracrine signals to other cellular compartments that contribute to wound healing, including the vasculature and the stroma (as depicted in Fig. 1), and recent evidence suggests that some of these signals are regulated by keratinocyte integrins.^{32,62,122,123} Such intercellular signals may be propagated from the epidermis to other cells through physical changes in the ECM (i.e., mechanical signaling), through diffusible growth factors that are secreted by the epidermis, or through generation of bioactive fragments following ECM proteolysis.

An important mode of intercellular crosstalk within the wound microenvironment is growth factor–mediated communication. Although keratinocytes reside in the epidermal compartment of skin and are spatially separated from the stromal compartment, it has long been known that growth factors and cytokines produced de novo by the epidermis can diffuse to other cellular compartments of the wound.¹⁴⁸ Thus, the epidermis is able to influence other cell types (i.e., endothelial cells, fibroblasts, immune cells) by regulating the availability of growth factors such as VEGF, TGFβ, and KGF. Some integrins can regulate the expression of growth factors by keratinocytes, which can then mediate paracrine stimulation of other cell types. For instance, deletion of α₃β₁ from cultured keratinocytes or epidermis reduced the expression of the pro-angiogenic factor mitogen-regulated protein 3 (MRP-3), which contributed to reduced stimulation of endothelial cell migration in vitro and impaired wound angiogenesis in vivo.³⁶ Conversely, it has been shown in α₂-null mice that ablation of α₂β₁ (albeit in all cell types at once) results in neovascular enhancement in both wounds and sponge implants, indicating an anti-angiogenic role for this integrin that might involve intercellular crosstalk.⁴⁷ These findings demonstrate that epidermal integrins may regulate keratinocyte-to-endothelial cell crosstalk to modulate wound angiogenesis, and that coordinated activities of different integrins may be important for proper outcome. As integrin–ECM interactions are likely to be regulated temporally during wound healing, we speculate that integrins might regulate the ability of epidermis to influence not only angiogenic growth, but also blood vessel regression and vascular normalization at later stages of wound healing.

There is also published evidence to support a role for crosstalk from keratinocytes to mesenchymal fibroblasts during wound healing. Indeed, delays in wound re-epithelialization were associated with enhanced wound fibrosis, possibly due to a lack of signaling from epidermis, which may play a role in hypertrophic scar development.^135,149 Studies in co-culture models have demonstrated that many fibroblast genes are regulated by keratinocyte-derived factors, including genes coding for growth factors, ECM components, and MMPs,¹³⁶ all vital constituents for successful wound healing. Integrin α₃β₁ is an intriguing candidate for influencing keratinocyte-to-fibroblast crosstalk in cutaneous wounds, given its above described role in promoting keratinocyte-to-endothelial cell crosstalk.³⁶ For example, in a murine model of lung fibrosis, ablation of α₃β₁ in lung epithelial cells resulted in reduced β-catenin/Smad signaling, accompanied by decreased accumulation of lung myofibroblasts.⁸³

Some epidermal integrins may control the bioavailability of ECM-bound growth factors or bioactive ECM fragments, either directly or indirectly, that influence the behaviors of other cell types. A direct role for certain α_v integrins in the local activation of the ECM-associated latent TGFβ complex has already been discussed in an earlier section.¹⁰⁴ While the extent to which the latter mechanism might render an activated growth factor available to a distal cellular compartment remains unclear, there are also examples of integrin-mediated liberation and diffusion of ECM-bound growth factors. For example, certain integrins (i.e., α_vβ₆ and α₃β₁) can induce expression of MMP-9, uPA, or other extracellular proteases^{79,101,137,150} that can degrade ECM and release reservoirs of ECM-associated growth factors (i.e., VEGF) to promote angiogenesis.^151,152 In addition, degradation of laminins, collagens, or other ECM proteins by some of these extracellular proteases may lead to the generation of bioactive matrix fragments that can directly stimulate cell growth or motility.^152,153

Finally, changes in ECM composition or structure that alter the physical properties (i.e., stiffness) of a tissue can influence cell function through mechanical signals. It follows that integrin-dependent changes in ECM, brought about by deposition of matrix proteins or matrix proteolysis, might mediate intercellular crosstalk from the epidermis to other wound cells through mechanical signaling. Mechanisms whereby different cell types use mechanical signals to communicate with one another are still poorly understood. However, mounting evidence supports an important role for alterations in matrix compliance and mechanical stress in the pathogenesis of chronic wound healing,¹ and the current use of topical negative pressure devices as an adjuvant therapy for chronic wounds may promote healing through increased mechanical tension in the wound.^154,155 Moreover, defects in ECM deposition or proteolytic processing that are likely to abberantly alter the mechanical properties of the ECM are thought to contribute to the pathogenesis of chronic wounds.^156,157

Future Directions

Exploiting integrins as therapeutic targets in wound healing

The ever-growing elderly and diabetic populations in the United States create a continuously increasing demand for effective wound healing therapies, which have yet to be identified. Impaired re-epithelialization is a hallmark of chronic wounds and poses a major clinical concern due to increased susceptibility of patients to infection. Prolonged treatment of chronic wounds, as well as infection due to slow healing, places a substantial financial burden on the U.S., and negatively impacts quality of life for the patient. Mounting evidence supports the idea that changes or defects in integrin-dependent keratinocyte functions contribute to the pathogenesis of chronic wounds, providing a strong rationale for exploiting epidermal integrins as therapeutic targets for their treatment.

Chronic wounds are often characterized by persistent inflammation, unhealthy granulation tissue, and impaired re-epithelialization.¹³⁹ The mechanisms that prevent the healing of chronic wounds are unknown. However, it is clear that the ECM is compromised in chronic wounds, and aberrant changes in ECM or its receptors are implicated in the pathology of chronic wounds, as reviewed elsewhere.¹³⁹ For example, it has been postulated that elevated levels of pro-inflammatory cytokines in chronic wounds results in atypical secretion/activation of extracellular proteases, which in turn may lead to “suboptimal ECM composition” and reduced integrin ligation that does not support normal growth factor activity.¹³⁹ In addition, diabetic foot ulcers often display up-regulation of MMPs 1, 2, 8, and 9, and concurrently reduced levels of tissue inhibitor of metalloprotease-2 (TIMP-2) compared to traumatic wounds.^139,145 Moreover, deregulation of syndecan-1 and syndecan-4 have been detected in venous leg ulcers.¹⁵⁸ Fibronectin deficiency, most likely resulting from enhanced ECM degradation, has been reported to occur in chronic wounds where it may impair cell migration.^139,159,160 Moreover, fibronectin fragments (generated by fibronectin degradation) have been shown to modulate levels of MMPs and TIMPs in some models,^139,161 and have been reported to occur in chronic wound fluids.^162–164

As bidirectional signaling receptors that regulate both keratinocyte-mediated changes to the wound microenvironment, and keratinocyte responses to those microenvironmental changes, integrins are attractive targets for therapeutic strategies to promote wound healing or to treat chronic wounds. The general concept of therapeutically targeting integrin function is already well established. For example, the integrin-blocking agent Cilengitide, an RGD mimetic, has shown promise in clinical trials involving the treatment of recurrent glioma.^165,166 It is well known that wound healing and skin carcinogenesis share similarities regarding both epidermal cell function and microenvironmental factors that drive each process,⁵⁷ and integrin-targeting therapies are relevant to the treatment of chronic wounds for many of the same reasons that they are relevant to the treatment of cancer. For example, we have already discussed how roles for some integrins within the stem cell compartment may be important during both wound re-epithelialization and skin carcinogenesis (see Epidermal proliferation).

Integrins that are upregulated in chronic wounds might serve as particularly attractive therapeutic targets. For example, integrin α_vβ₆ is not normally expressed in epidermis but is induced during normal wound healing and has been shown to be strongly upregulated in chronic wounds of human patients.¹⁶⁷ In the same study, transgenic mice that constitutively over-expressed β₆ integrin in the epithelium showed no change in wound closure rate but spontaneously developed chronic fibrotic ulcers,¹⁶⁷ while a separate study showed that diabetic β₆-null mice showed delays in early wound closure.¹⁶⁸ These findings indicate complex roles for integrin α_vβ₆ in wound healing that are likely to be precisely timed and may require the correct extracellular milieu. More work in this area must be done to tease apart and fully understand the precise functions of particular integrins in normal wound healing, then identify which of these functions are deficient in chronic wounds, in order to fully exploit integrins as therapeutic targets.

In addition to direct targeting of integrins, rational wound healing therapies might exploit integrins as a means of targeted therapeutic delivery. For instance, one group recently generated an injectable complex containing platelet derived growth factor B (PDGF-B) plasmid DNA and an integrin-selective RGDK-lipopeptide, towards the goal of delivering PDGF-B directly to relevant integrin-decorated cells within the wound bed, thereby circumventing the modest efficacy achieved through repeated topical application.¹⁶⁹ A single, subcutaneous injection of this complex was demonstrated to promote wound healing in a streptozotocin-induced diabetic rat model of chronic wounds, resulting in enhanced re-epithelialization, collagen fibrillogenesis, and blood vessel formation.¹⁶⁹ Thus, it is possible that similar strategies can be developed to direct therapeutic agents to wound epidermis using lipopeptides that recognize keratinocyte integrins.

In summary, while the field has made good progress in identifying functions of individual keratinocyte integrins, and understanding how their coordinated activities might control the range of epidermal functions that are required for normal wound healing, some formidable challenges lie ahead as we attempt to translate this knowledge into therapeutic approaches for wound healing deficiencies in human patients. Considering that several epidermal integrins are required for different epidermal functions, one challenge will be to develop multicombinatorial strategies to target several integrins. An additional challenge is that the relevant ligands in the wound microenvironment have not yet been identified for all epidermal integrins. Finally, as already mentioned, it is likely that the control of diverse epidermal functions by individual integrins is precisely timed, in order to maintain properly coordinated regulation of epidermal wound functions.

Take-Home Messages.

• • Integrins are bidirectional signaling receptors that regulate both keratinocyte-mediated changes to the wound microenvironment and keratinocyte responses to microenvironmental changes.

• • In addition to controlling many cell-autonomous keratinocyte functions such as migration, proliferation, and survival, some integrins can regulate the ability of epidermis to cross-talk to other cellular compartments, such as endothelial cells and fibroblasts, in a paracrine fashion.

• • The dynamic regulation of ECM composition during wound healing is critical to coordinate integrin activity in a spatiotemporal manner in order to successfully promote wound healing. Alternatively, inappropriate ECM composition/mechanics are known to contribute to the pathogenesis of chronic wounds.

• • Chronic wounds display altered integrin expression, as well as changes in the extracellular milieu, which can both be caused by, and contribute to, inappropriate integrin signaling or aberrant MMP activity.

• • Integrins expressed on wound epidermis represent potential therapeutic targets for the treatment of chronic wounds, although some formidable challenges lie ahead before we can fully exploit integrins as clinical targets.

Abbreviations and Acronyms

BM

basement membrane

ECM

extracellular matrix

EGF

epidermal growth factor

FAK

focal adhesion kinase

GTPase

guanosine triphosphatase

ILK

integrin-linked kinase

JEB

junctional epidermolysis bullosa

LAP

latency-associated protein

LTBP

latent TGFβ binding protein

MAPK

mitogen-activated protein kinases

MMP

matrix metalloprotease

PDGF-B

platelet derived growth factor B

PI3-K

phosphatidylinositol 3-kinase

PLC

phospholipase C

SCC

squamous cell carcinoma

SH2

SRC-homolog 2

TGFβ

transforming growth factor beta

TIMP

tissue inhibitor of metalloprotease

uPAR

urokinase receptor

VEGF

vascular endothelial growth factor

Acknowledgments

The authors are grateful to Dr. Livingston Van De Water for critical reading of the manuscript, as well as to other colleagues at Albany Medical College and members of the DiPersio laboratory for valuable discussions and insights. Research in the DiPersio laboratory is supported by NIH grants from NCI (R01CA129637 to C.M. DiPersio) and NIAMS (R01AR063778 to C.M. DiPersio and L. Van De Water). Whitney Longmate is supported by an NRSA predoctoral fellowship from the National Institutes of Health (F31CA174198). We offer our apologies to the many researchers whose valuable contributions to the field could not be cited due to space constraints.

Author Disclosure and Ghostwriting Statement

The authors declare no competing financial interests or other conflicts of interest. No ghostwriters were involved in the preparation of this manuscript.

About the Authors

The authors are members of the Center for Cell Biology and Cancer Research (CCBCR) at Albany Medical College, Albany, New York. Michael DiPersio is a Professor in CCBCR whose research program is focused on investigating molecular mechanisms whereby integrins regulate extracellular matrix remodeling and gene expression in epidermal keratinocytes and carcinoma cells. Whitney Longmate is a predoctoral student in the DiPersio laboratory.