Focal Contact and Hemidesmosomal Proteins in Keratinocyte Migration and Wound Repair

Abstract

Significance: During wound healing of the skin, keratinocytes should move over while still adhering to their underlying matrix. Thus, mechanistic insights into the wound-healing process require an understanding of the forms and functions of keratinocyte matrix adhesions, specifically focal contacts and hemidesmosomes, and their components.

Recent Advances: Although the structure and composition of focal contacts and hemidesmosomes are relatively well defined, the functions of their components are only now being delineated using mouse genetic models and knockdown approaches in cell culture systems. Remarkably, both focal contact and hemidesmosomal proteins appear involved in determining the speed and directional migration of epidermal cells by modulating several signal transduction pathways.

Critical Issues: Although many publications are centered on focal contacts, their existence in tissues such as the skin is controversial. Nonetheless, focal contact proteins are central to mechanisms that regulate skin cell motility. Conversely, hemidesmosomes have been identified in intact skin but whether hemidesmosomal components play a positive regulatory function in keratinocyte motility remains debated in the field.

Future Directions: Defective wound healing is a developing problem in the aged, hospitalized and diabetic populations. Hence, deriving new insights into the molecular roles of matrix adhesion proteins in wound healing is a prerequisite to the development of novel therapeutics to enhance tissue repair and regeneration.

Jonathan C.R. Jones, PhD

General Introduction

The skin, comprising an outer epidermal sheet composed of multilayers of keratinocytes and an underlying connective tissue or dermis, is a barrier that protects the body from pathogens and excessive water loss, and which is involved in insulation, temperature regulation, and sensation. In intact skin, keratinocytes in the basal layer of the epidermis abut a basement membrane that is rich in extracellular matrix (ECM) proteins, including laminin, type IV collagen, type VII collagen, nidogens, and perlecan. These matrix components are woven into a complex meshwork that anchors the epidermis to the dermis.

Injuries of the skin, including burns and mechanical trauma, may result in damage to just the keratinocyte layers of the epidermis or damage to both the epidermal and dermal layers. The focus of this review is on the latter type of wounds where the basement membrane is breached and blood vessels are damaged. Under these circumstances, fibronectin is released into the wound bed along with fibrin deposited by platelets.¹ These two molecules crosslink to form a plug that provides a provisional matrix for subsequent cellularization of the wound.

Complete epithelialization of a wound by keratinocytes is essential for the restoration of the barrier function of the skin. The epidermal “component” of wound healing comprises three major stages: migration, proliferation, and differentiation of keratinocytes. These processes are initiated, at least in part, as a response to cues from surrounding cells, such as cytokines released from leukocytes and platelets, and from various environmental signals, such as stretch and electrical fields.^2,3 The keratinocytes immediately adjacent to the wound proliferate and, at the same time, undamaged keratinocytes at the wound margins migrate over a provisional matrix, rich in matrix molecules, including fibronectin and fibrin, covering the wound bed (Fig. 1).¹ Migration of keratinocytes requires that both cell-cell and cell-substratum contacts and the keratinocyte cytoskeleton are remodeled to allow the cells to detach from the intact basement membrane in the unwounded epidermis (Fig. 1).⁴ These cells then migrate onto the provisional matrix of the wound. As the cells move over the wound, they degrade and remodel the provisional matrix while also depositing new matrix proteins, including laminin-332 (Fig. 1).^5,6 After complete epithelialization of the wound, both keratinocytes and fibroblasts contribute to the reformation of the basement membrane. Finally, barrier function is restored as the keratinocytes covering the wound site stratify and differentiate.⁷

Figure 1.

A schematic of the edge of a wound-healing epidermis. On the left, a keratinocyte in an unwounded region adheres to the laminin-332 (LM332) in the basement membrane (BM) in part through a keratin-associated hemidesmosome (HD) and possibly, an actin-interacting focal contact (FC). The HD is targeted by kinases and proteases (arrow) in cells at the wound margin (middle cell). Cells that leave the BM move over a provisional matrix composed of fibronectin (depicted as fibrils) and laminin-332. Focal contact and hemidesmosomal integrins interact with this matrix and likely regulated the directed migration of the motile keratinocytes. The box indicates the key for α₃β₁ and α₆β₄ integrin heterodimers.

The focus of this review is how specific matrix molecules, their receptors, and associated proteins in keratinocytes are involved in epidermal repair after skin insult. In keratinocytes, matrix proteins and their receptors are clustered in two distinct protein complexes: focal contacts (adhesions) and hemidesmosomes (Fig. 2).⁸ In vivo, at the ultrastructural level, hemidesmosomes link basal keratinocytes to the basement membrane zone and are identified by their characteristic subplasma membrane, cytoplasmic electron dense plaques to which keratin intermediate filaments attach (reviewed in Jones et al.⁹). In sharp contrast, in vitro, cultured keratinocytes do not assemble bona fide hemidesmosomes, at least as judged electron microscopically. In such cells, hemidesmosomal proteins aggregate into what some have termed stable anchoring complexes (Figs. 3 and 4).^10–13 In culture, ex vivo, keratinocytes assemble focal contact structures that serve as the cell surface anchorage site for the actin cytoskeleton (Fig. 4). The tissue equivalent of the focal contact is not obvious in a skin cell at the ultrastructural level although keratinocytes in the basal layer of the epidermis express a wide range of focal contact proteins. That focal contacts and their proteins play important roles in cell migration is undisputed. In contrast, whether hemidesmosomal proteins positively modulate migration is controversial, at least with regard to skin wound healing. We will begin with a brief description of the focal contact and its components. This is not a comprehensive review of focal contact proteins, as reviewing the more than 100 proteins found in focal contacts would be an overwhelming task. Rather, we will focus on a select number of focal contact proteins with known functions in regulating motility.

Figure 2.

Schematics of the protein components of the focal contact (left) and hemidesmosome (right).

Figure 3.

Immortalized human keratinocytes were allowed to attain confluence in vitro and were then scratched wounded. At 6 h after wounding, the cells were prepared for confocal immunofluorescence using antibodies against β₄ integrin (A). Cells at the wound edge have begun to move onto the wound site. White arrows indicate β₄ integrin staining along the leading lamellipodia of cells at the wound margin. These same areas are marked by arrows in the phase image of the cells in (B). Scale bar, 20 μm.

Figure 4.

Immortalized human keratinocytes maintained at sub-confluence in vitro were prepared for triple label immunofluorescence using antibodies against α₆ integrin, laminin-332, and paxillin as indicated. The merge of the three images indicates that α₆ integrin-rich hemidesmosome protein complexes co-distribute with laminin-332, while paxillin focal contacts localize at the very edge of the cells and fail to show clear co-localization with laminin-332. The phase image shows the cell outline. Scale bar, 10 μm.

Discussion of findings and relevant literature

Focal contacts

In vitro, as an individual cell migrates, focal complexes at the cell edge mature into focal contacts, which, in concert with the actin cytoskeleton, generate the traction forces needed to pull the cell forward.⁴ Focal contacts then disassemble at the cell rear to allow the latter to release from the underlying matrix as the cell advances over its substrate. The migration of cells in tissues is more complex, however, as keratinocytes advance over a wound as a sheet via a process termed collective cell migration.¹⁴ Recent studies have begun to uncover the precise mechanisms that regulate collective migration, but they are still incompletely understood.¹⁴ With regard to wound healing in the skin, one assumes that focal contacts at the leading edge of the “leader” cells along the migrating tip of the epidermis define sheet polarity, generate traction forces, and guide the migration of “follower” cells.¹⁴ In the skin, there is also the complication that there is some multi-layering of cells, one or two cells back from the leading edge of the migrating epidermal tip. Whether these suprabasal layers simply tag along for the ride or play an important role in the wound-healing process remains unknown.

Transmembrane proteins of the focal contact

Integrins

At the core of the focal contact, matrix receptors called integrins link the actin cytoskeleton to various components of the ECM, albeit indirectly (Fig. 2).¹⁵ They also mediate both inside-out and outside-in signaling which is central to the role of focal contacts in regulating pathways that determine both the directionality and speed of motile cells.¹⁶ Integrins exist as heterodimers of one α and one β subunit. To date, 18 α and 8 β subunits have been identified; they can assemble in 24 specific combinations to form heterodimers in mammals.¹⁵ Both subunits contain a large N-terminal extracellular globular head domain (∼1,000 amino acids for α subunits and ∼700 amino acids for β subunits), a transmembrane segment, and a C-terminal cytoplasmic tail that has affinity for cytoskeletal as well as scaffold and signaling proteins.¹⁵

Many α subunits contain sites for post-translational processing, such as cleavage into two polypeptides. Cleavage occurs intracellularly and results in two α integrin chains (one heavy and one light) that are covalently linked by a disulfide bridge.¹⁷ The heavy chain remains completely extracellular in the mature integrin molecule, whereas the light chain encompasses a single transmembrane region and a short cytoplasmic tail. The β integrin subunits are not cleaved, and their cytoplasmic tails are typically short, around 50 amino acids in length. The one exception to this is the β₄ integrin subunit, which has a cytoplasmic tail of more than 1,000 amino acids.¹⁵

Each integrin heterodimer binds to a distinct ECM ligand, such as collagen, fibronectin, or laminin. The respective N-terminal head domains of the α and β integrin subunits function cooperatively by binding to specific ligands together.¹⁵ Integrins contain multiple sites for interacting with various extracellular ligands; the classic example is the short tripeptide amino-acid sequence of Arg-Gly-Asp (RGD).¹⁵ Although this recognition sequence was originally identified as the minimal essential peptide sequence for fibronectin adhesion, it is a common motif that integrins utilize to bind to other cell-adhesive proteins.¹⁸ However, laminin-binding integrins, such as α₆, utilize different recognition sequences.

During migration and wound healing, cells extend actin-stabilized filopodia to “sample” the terrain.¹⁹ Integrins for the ECM substrate become enriched at the filopodia, thereby enhancing matrix adhesion.²⁰ Integrin-ligand interactions also induce conformational changes in the receptor that allow recruitment of various signaling intermediates (see below), which, in turn, activate Rho GTPAses, including Rac. The latter direct cytoskeletal reorganization and the extension of actin-rich lamellipodia.²⁰ Integrin activation of Rac leads to a signaling cascade, resulting in phosphorylation of myosin light chain kinase, which activates myosin II to create actin-myosin contraction forces to pull the cell forward.²¹ During cell migration, these steps are repeated to generate traction as cells extend lamellipodia and form new points of contact with the ECM.²⁰

A detailed discussion of integrin functions in the intact and wounded epidermis is provided in an excellent recent review.²² For that reason, our discussion on the topic will be brief. In intact skin, keratinocytes express the focal contact-associated integrins α₂β₁, α₉β₁, and α₃β₁.^23,24 The functions of these integrins have been examined extensively by the generation of knockout mice. Although the results from such analyses have provided important new mechanistic clues to the molecular pathways regulating wound healing, there are some caveats. Mouse wounds heal primarily by contracture, as their skin is relatively loose. In contrast, human skin is tight and heals primarily by epithelialization. Thus, translating results from in vivo mouse wound healing studies to the human condition may not be valid in every case.

Intriguingly, despite its expression in many tissues, α₂β₁ integrin-deficient mice are not only viable but display no obvious defects in wound healing relative to their wild-type litter mates.²⁵ Analyses of α₉β₁ integrin conditional knockout mice demonstrate poor re-epithelialization after wounding. However, this is most likely the result of decreased keratinocyte proliferation, rather than defects in cell migration.²⁴

The role of the laminin-332 receptor α₃β₁ in cell migration and wound healing is controversial. The wounds of α₃β₁ integrin-null mouse epidermis exhibit faster wound closure when compared with wild-type skin.²⁶ Moreover, using keratinocytes isolated from the α₃β₁ integrin null mice, migration assays reveal that the null keratinocytes migrate faster and with greater directional persistence.^26,27 These results imply that α₃β₁ integrin functions as a brake on migration. However, other data argue against this. For example, Choma et al. also analyzed the migratory behavior of α₃ integrin-null mouse keratinocytes and have presented evidence that these cells lack persistent migration, stable lamellipodia, and polarity.²⁸ These findings are consistent with studies indicating that α₃β₁ integrin-laminin-332 ligation promotes keratinocyte polarization and persistent migration.²⁹ Indeed, our recent data, using cultured human keratinocytes which exhibit a loss of α₃ integrin expression, indicate that such cells move slower than their wild-type counterparts on laminin-332 rich matrices.³⁰ Thus, together, these studies suggest that α₃β₁ integrin is a positive regulator of the motility rate of a keratinocyte.^28–30 One possible explanation for the discrepancy of these studies is that there are differences in integrin profiles of the α₃ integrin null mouse keratinocytes derived in different laboratories. Specifically, Margadant and co-workers have reported an up-regulation of α₆ integrin in their α₃ integrin-null mouse skin cells.²⁶ Thus, α₆β₁ integrin could compensate for α₃β₁ integrin loss in such cells and support migration on laminin-332 matrices.²⁶ However, such up-regulation has not been observed by others and may explain why the latter workers observe motility defects in α₃ integrin-null keratinocytes maintained on laminin-332.³¹ Another possibility comes from recent analyses of the composition of the matrix deposited by mouse and human keratinocytes in vitro. The former assemble a matrix enriched in both fibronectin and laminin-332, whereas the latter deposit a matrix primarily composed of laminin-332 with little, if any, fibronectin.³² Moreover, Hamill et al. demonstrated that fibronectin acts as a brake to laminin-332 induced migration of both mouse and human keratinocytes and that the matrix of α₃ integrin-null mouse epidermal cells contains less fibronectin than controls.³² Hence, the differences observed in migration speed of mouse keratinocytes lacking α₃ integrin and wild type cells may result from a defect in the deposition of fibronectin by the former cells rather than α₃ integrin functioning as a direct brake on migration in the latter cells. Of course, this raises the issue as to how α₃ integrin regulates fibronectin deposition and why human and mouse skin cells deposit matrices of a different composition. Interestingly, α₃ integrin is found between cells at the leading margin of wounds (Fig. 5A). What function it plays in such sites is unclear. Finally, to complicate the analyses of the functions of α₃β₁ integrin in wound healing in vivo is the finding that loss of α₃β₁ integrin in the epidermis of mice impacts wound healing in vivo in animal models via crosstalk with endothelial cells and angiogenic mechanisms.³³

Figure 5.

The images show the migrating tongue of keratinocytes in an incisional human wound at 30 h after wounding. The cells were moving toward the right [large arrows in (A) and (B) indicate the direction of movement]. (A) The section was prepared for immunofluorescence using antibodies against α₃ integrin (red) and BPAG1e (green). The nuclei are stained with DAPI (blue). α₃ integrin (red) is predominantly at sites of cell-to-cell contact with no clear co-distribution with BPAG1e (green) located along the advancing wound margin juxtaposed to the wound matrix (arrows). (B) The section was costained with antibodies against β₄ integrin (red) and BPAG1e (green). The nuclei are stained with DAPI (blue). Note that β₄ integrin and BPAG1e co-distribute as evidenced by the yellow color along the wound margin (arrows). M, matrix of wound bed. These images were presented in Underwood et al. as Figs. 8C and 5C.¹⁰⁴ Scale bar, 10 μm.

On wounding, expression of the fibronectin receptor α₅β₁ integrin as well as α_vβ₅ and α_vβ₆ integrins is upregulated.³⁴ α₅β₁ integrin likely regulates the migration of keratinocytes over the fibronectin-rich provisional matrix during early stages of wound healing.³⁵ Ablation of α_vβ₅ or α_vβ₆ integrin in the epidermis of the mouse has no apparent impact on re-epithelialization of wounds with the possible exception of chronic wounds.^36,37 However, whether α_vβ₆ integrin plays important roles in human skin wound healing remains an open question. Since it is involved in activation of TGF-β and, hence, triggers pathways leading to up-regulation of collagen gene expression, there is considerable interest in its role in the development of scarring and fibrosis in the skin after wounding.³⁸

Syndecans and focal contacts

Cell interaction with matrix at the focal contact is also mediated by members of the syndecan family (Fig. 2).³⁹ Indeed, these transmembrane heparan sulfate proteoglycans often function as integrin co-receptors.⁴⁰ Each syndecan consists of a conserved transmembrane core and a highly divergent extracellular domain. The ectodomain of a syndecan can interact with a variety of growth factors, cytokines, and ECM proteins, including laminin-332.⁴¹ There are four syndecan members expressed in mammals with syndecan-1 and -4 being highly expressed in epithelial tissues, including keratinocytes.⁴¹ The heparan sulfate chains of the syndecans bind cytokines and growth factors, which, in turn, influence signaling cascades and matrix protease function that are necessary for wound healing.^42,43

That syndecan-1 is involved in wound healing derives from studies of syndecan-1 null mice.⁴⁴ Closure of wounds in their corneas is impeded as a consequence of epithelial cell migration and proliferation defects.⁴⁴ The latter may be due to perturbation in the expression and localization of α₉ integrin.^24,44,45 In addition, Stepp et al. have demonstrated that syndecan-1 null keratinocytes exhibit enhanced adhesion with a corresponding decrease in the rate of their migration.⁴⁶ This appears due to changes in either the organization of their laminin-332 matrix and/or by an indirect increased activation of adhesion-promoting α₆β₄ integrin complexes.⁴⁶

Syndecan-4 binds the G domain of the α₃ subunit of laminin-332.⁴⁷ Addition of a synthetic peptide corresponding to the syndecan-4 binding region to cultured primary human keratinocytes promotes their migration apparently by activating the β₁ integrin subunit.⁴⁷ A role for syndecan-4 in regulating wound healing is also indicated in studies of syndecan-4 null mice that exhibit delayed wound repair. However, this wound-healing defect may be due to motility defects in fibroblasts rather than due to defects in keratinocyte migration.⁴⁸

Interestingly, the extracellular domains of syndecans are constitutively shed, a process that is up-regulated in inflammation, wound healing, and some tumors.⁴⁰ Specifically, the shed ectodomains of syndecan-1 and -4 are elevated in injured tissue as well as wound fluid where they likely influence signal transduction, protease activities and modulate the functions of intact syndecans by competing for interactions with growth factors or cytokines.^49–51

Cytoplasmic proteins of the focal contact

Kindlins and talin

Proteins of the kindlin family are found in focal contacts where they regulate the function of β integrin subunits (Fig. 2). There are three members of the kindlin family of focal contact proteins: kindlin-1, mainly in epithelial cells; kindlin-2, expressed in most tissues; and kindlin-3, which is expressed in cells of the hematopoietic system.⁵²

In humans, mutations in kindlin-1 are known to cause Kindler syndrome, a rare skin disease that is characterized by fragile skin with trauma-induced microblisters, poikiloderma, cutaneous atrophy, and photosensitivity in adulthood.⁵³

All three kindlin proteins share a high degree of sequence similarity and domain structure. Each contains a pleckstrin homology domain and a 4.1 protein, ezrin, radixin, and moesin (FERM) domain. The FERM domain consists of three subdomains, one of which encompasses an integrin binding site.⁵² The FERM domain of the kindlins is similar to that found in talin, especially within the integrin-binding region.⁵⁴ Indeed, it has been shown the both kindlin-1 and -2 directly bind the cytoplasmic tails of β₁ and β₃ integrin, where they serve as adaptor proteins.⁵⁵

To distinguish the functions of kindlin-1 and -2, He et al. developed keratinocytes deficient in one or both of the kindlins. Loss of both kindlin-1 and -2 results in more severe changes in cell morphology than loss of one or the other.⁵⁶ In addition, these proteins compensate for each other with regard to activation of β₁ integrin and in keratinocyte adhesion to a substrate. In a scratch wound assay, keratinocytes deficient in both kindlins are unable to migrate. Kindlin-1 deficient cells move but heal a wound more slowly than their wild-type counterparts.⁵⁶ In contrast, kindlin-2 deficient keratinocytes demonstrate an even more marked defect in migration in a scratch assay compared with kindlin-1 knockdown cells. Both kindlin-2 and the double-deficient cells also exhibit a loss in an ability to migrate directionally.⁵⁶

Analyses of keratinocytes from patients suffering from Kindler syndrome have demonstrated that they have an abnormal, irregular shape, appearing more fibroblastic in nature, and have lost the ability to form colonies.⁵⁷ When grown in culture, these cells show aberrant localization and decreased expression of α₆β₄ integrin and collagen XVII, but surprisingly, an increase in the level of laminin-332.⁵⁷ These results, as well as immunoprecipitation data and structural modeling, have led to the speculation that kindlin-1 may associate with the hemidesmosome.⁵⁸ Indeed, kindlins are candidate mediators of the crosstalk between hemidesmosomes and focal contacts (see below).⁸

Similar to the kindlins, talin is an integrin-activating and actin-binding protein. The precise functions of talin in wound healing of the skin and the migration of keratinocytes have not been precisely elucidated. Nonetheless, since talin is expressed by cultured keratinocytes and localizes to the epidermal-dermal junction, it likely plays roles in wound repair and merits some discussion.⁵⁹ Talin is composed of an N-terminal globular head and a larger C-terminal rod. The head region contains an FERM domain that is composed of three subdomains, F1, F2, and F3.⁵⁴ The F3 subdomain binds to and activates the β₃ integrin tail, while a more extensive head fragment is necessary for activation of β₁ integrins.⁵⁴ The talin rod domain contains binding sites for actin, vinculin, and β integrin tails. Thus, talin regulates both integrin affinity for its extracellular matrix ligand and serves as a linker between ligand bound integrins and the actin cytoskeleton.⁶⁰

The talin molecule is cleaved between the head and tail domain by calpain.⁶¹ The cleavage of talin plays an important role in focal contact dynamics and, thus, cell migration. Indeed, loss of expression of calpain has been implicated in defects in fibroblast migration.⁶² In Chinese hamster ovary cells, it has been shown that the talin head is phosphorylated by Cdk5, blocking its degradation and leading to stabilization of focal contacts.⁶³ If talin is not phosphorylated, it binds Smurf1 and is ubiquitinated, and marked for degradation, thus promoting focal contact disassembly.⁶³ In migrating cells, talin is a regulator that is involved in the balance between maintenance of stable focal contacts which facilitate development of traction forces and the disassembly of adhesion structures in order to permit the cell to move forward. Intriguingly, talin interacts directly with the Rac exchange factor Tiam 1, thereby regulating Rac1 activity and polarized cell migration in certain cells.⁶⁴ In this regard, Tiam1 has been reported to be involved in laminin-332 deposition that is regulated by α₃β₁ integrin and keratinocyte migration.⁶⁵

Paxillin

Paxillin is an adaptor protein that is localized to focal contacts and has roles in cell adhesion and cell migration (Fig. 2).⁶⁶ It contains multiple docking sites for a variety of proteins, including β integrins, actopaxin, GTPase activating proteins, and vinculin, as well as tyrosine kinases such as focal adhesion kinase (FAK) and integrin-linked kinase (ILK).⁶⁶ Paxillin localizes near the leading edge of migrating cells and is believed to be a very early component of the assembling focal contact.⁶⁷ In addition to recruiting various kinases to focal contacts, the phosphorylation of paxillin by FAK, c-Jun amino-terminal kinase (JNK), extracellular signal-regulated kinase, and p38 mitogen-activated protein kinase (MAPK) regulates its activity and localization.^68,69

In nonepithelial cells and tumor cells, paxillin regulates motility by its association with various signaling pathways as well as its effect on focal contact remodeling and lamellipodial stability.^66,70 In wounded epithelial tissues, epidermal growth factor (EGF) activates JNK, which, in turn, phosphorylates paxillin on residue serine 178.⁷¹ This phosphorylation event is required for additional FAK-mediated phosphorylation at residues tyrosine 31 and 118, which is necessary for corneal keratinocyte motility.⁷²

FAK and ILK

FAK is a nonreceptor tyrosine kinase that is localized to focal contacts and which is involved in multiple signaling cascades and integrin-mediated signaling pathways (Fig. 2).⁷³ FAK is composed of a kinase domain that is flanked on the N-terminal side by a FERM domain and on the C-terminal end by two proline-rich docking sites and a focal contact targeting sequence.⁷⁴ Integrin/ECM ligation activates FAK, inducing a conformational change and resulting in autophosphorylation and interaction with and activation of Src family kinases.⁷⁴ This process then leads to phosphorylation of multiple downstream targets that drive cell migration. The importance of FAK in keratinocyte migration is indicated by its localization at the epidermal-dermal junction in partial-thickness burn wounds, a site corresponding to actively migrating keratinocytes covering a wound.⁷⁵ Moreover, studies in cultured keratinocytes have demonstrated an increase in phosphorylation of FAK in migrating cells compared with nonmotile cells.⁷⁶ Finally, when FAK expression is knocked down or ablated, keratinocyte migration is significantly reduced compared with controls, with focal contacts in the knockdown cells being less dynamic and exhibiting defective disassembly properties.^76,77 Support for the importance of FAK in keratinocyte migration comes from analyses of human and mouse skin explants. FAK is up-regulated at the front of migrating keratinocytes, and surprisingly, also within the explant itself.⁷⁸ Addition of a tyrosine kinase inhibitor results in decreased keratinocyte migration that is relieved by removal of the inhibitor.⁷⁸ Moreover, keratinocytes emerging from FAK-null mouse skin explants show loss of directed migration that is due, most likely, to their aberrant actin cytoskeleton organizations when compared with wild-type cells.⁷⁷

ILK binds the cytoplasmic tails of β₁, β₂, and β₃ integrins, linking them to the actin cytoskeleton and serves as an adaptor/scaffolding protein for signal transduction.⁷⁹ ILK has been linked to phosphorylation of Akt and it is activated by substrate adhesion in a PI3 kinase-dependent manner.⁸⁰ Akt phosphorylation is known to regulate the motility of a variety of cells.⁸¹ Therefore, one might assume that Akt phosphorylation plays a role in keratinocyte migration. In support of this, ILK knockout mice display impaired migration of keratinocytes in the hair follicle.⁸² In addition, keratinocytes isolated from these mice exhibit unstable lamellipodia, a lack of persistent migration, and failure to close scratch wounds efficiently.⁸²

Alpha actinins

The actinins are members of a large family of actin-bundling proteins, including fimbrin, spectrin, and dystrophin, which share similar actin-binding domains and an EF hand calcium regulatory domain.⁸³ There are four major isoforms of actinin, including two muscle-specific isoforms, actinin-2 and actinin-3, and two nonmuscle isoforms, actinin-1 and actinin-4.^84–86 Though actinin-1 and actinin-4 share about 87% protein sequence homology, they show differences in subcellular localization and function, depending on cell type. Actinin-1 localizes exclusively in the cytoplasm, where it is found along actin stress fibers and is required to crosslink linear dorsal stress fibers.^87,88 It associates with the cytoplasmic domain of β₁ integrin and, when phosphorylated, it can interact with both FAK and Src, where it enhances signaling that is mediated by focal contacts and increases integrin-mediated adhesion (Fig. 2).

In contrast to actinin-1, actinin-4 is concentrated at the leading edge of migrating cells as well as localizing with actin stress fibers and cell-cell contacts.^85,89,90 A role for actinin-4 in skin cell migration comes from recent studies using a knockdown approach. Actinin-4 deficient cultured human keratinocytes have defects in their lamellipodial dynamics, an inability to migrate progressively, and a decrease in the activation state of the actin severing protein cofilin.⁹¹ These defects are rescued by plating the actinin-4 knockdown keratinocytes onto the laminin-332-rich matrix laid down by wild-type keratinocytes, implying a role for actinin-4 in both signaling required for directed migration and the deposition of a laminin matrix.⁹¹

Flii

Flii, a member of the gelsolin family, is an actin remodeling protein that interacts with the focal contacts protein talin through its leucine-rich repeat domain.^92,93 Moreover, it appears to function as a negative regulator of wound healing in mouse skin (Fig. 2).^92,93 Specifically, overexpression of Flii in mice leads to impaired wound healing with larger wound beds, while mice with decreased expression of Flii show enhanced wound healing and migration.⁹² Though the mechanism for Flii function in wound healing is not fully understood, it may modulate the migration of skin cells via its effects on the pool of talin that is available for binding to and activating integrins.⁹⁴

Hemidesmosomes

Unlike fibroblasts and many epithelial cells, keratinocytes in intact skin and cornea assemble matrix adhesion attachment devices termed hemidesmosomes. The core of the hemidesmosome is composed of α₆β₄ integrin, which, via an interaction with collagen XVII (BPAG2, BP180), a type II membrane protein, and the plakin family members BPAG1e (BP230) and plectin, mediates cell surface interaction of the keratin cytoskeleton. α₆β₄ integrin and the extracellular collagen-like domain of collagen XVII bind to the matrix molecule laminin-332 in the basement membrane^8,9,95 (Fig. 2). Within the complex itself, β₄ integrin has binding sites for BPAG1e and plectin as well as collagen XVII. α₆ integrin also binds collagen XVII.⁹⁶ Moreover, there are a number of accessory proteins that have been reported to interact with the hemidesmosomes, including CD151.⁹⁷ It should be noted that there is far less molecular complexity to the hemidesmosome than the focal contact, at least with regard to the number of proteins associated with it (Fig. 2).

Hemidesmosomes are important for stable adherence of the epidermis to the dermis. Evidence for this comes from studies of numerous skin diseases where loss of hemidesmosomes results in blistering due to dysadhesion of the epidermis at or close to the basement membrane zone.⁸ The molecular defects in these diseases will be discussed next. Nevertheless, in wound healing, undamaged adherent keratinocytes at the margin of the wound release themselves from the constraints of the basement membrane zone, at least in part, by disassembling their hemidesmosomes (Fig. 1). Bona fide hemidesmosomes are only assembled after complete epithelialization of the wound bed. Such observations have led to the idea that hemidesmosomal proteins are not involved in the wound-healing process per se. Rather, hemidesmosomal proteins might be thought of as restricting motility over the ECM. However, recent studies from own lab argue otherwise and we have provided evidence for the role of several hemidesmosomal proteins in determining the motility of keratinocytes as detailed next.

Transmembrane components of the hemidesmosome

α₆β₄ integrin

α₆β₄ integrin binds laminin-332 in the ECM (Figs. 3 and 4).⁹⁸ The β₄ integrin subunit is unique among its β integrin relatives in that its cytoplasmic is more than 1,000 amino acids long.¹⁵ This allows it to interact with a number of high molecular cytoskeleton linker proteins, namely BPAG1e and plectin, which will be described next. BPAG1e and plectin bind the keratin cytoskeleton and, hence, α₆β₄ integrin possesses the ability to anchor keratin filaments to the cell surface.^9,95 It is well established from both antibody inhibition studies and analyses of knockout mice that ligation of α₆β₄ integrin by laminin-332 nucleates hemidesmosome assembly.^99,100 In addition, the long tail of β₄ integrin interacts with a number of signaling intermediates although whether α₆β₄ integrin is capable of signaling when it resides within the hemidesmosome is not clear.¹⁰¹ Rather, α₆β₄ integrin more likely signals to various pathways when not incorporated into the hemidesmosome proper.

On skin wounding, EGF is released into the wound and induces the Src kinase family member Fyn to phosphorylate β₄ integrin in those hemidesmosomes in the keratinocytes in the undamaged margins of the wound.¹⁰² This process may involve the interaction between syndecans and α₆β₄ integrin.¹⁰³ Indeed, it has been suggested that syndecans may function to bring the cytoplasmic domain of the β₄ integrin in close proximity to membranes to facilitate its phosphorylation.¹⁰³ Regardless, phosphorylation of the β₄ integrin cytoplasmic tail results in disassembly of hemidesmsomes, allowing keratinocytes to escape the constraints of the basement membrane.

Loss of expression of α₆β₄ integrin is the underlying cause of the blistering skin disease junctional epidermolysis bullosa (JEB) with pyloric atresia.⁸ This loss not only blocks hemidesmosome assembly but also results in dysadhesion of epidermal cells from the basement membrane. Indeed, this consequence has been used to suggest that the primary function of α₆β₄ integrin in the skin is to maintain epidermal-dermal adherence and that it is a stable anchor.²² Moreover, antibody blockade studies and analyses of motility of mouse keratinocytes deficient in β₄ integrin expression have also suggested that α₆β₄ integrin is a stable anchoring integrin.^99,100 However, there is now emerging evidence for the importance of α₆β₄ integrin in determining directed cell movement after wounding. First, α₆β₄ integrin localizes to the leading front of the cells migrating into wound sites in vitro and in tissue (Figs. 3 and 4B).¹⁰⁴ Moreover, Pullar and associates explored the directed migration of primary keratinocytes exposed to an electrical field, an early signal in the wound-healing process.³ They compared migration of wild-type keratinocytes and keratinocytes lacking β₄ integrin, isolated from a patient suffering from JEB with pyloric atresia, when exposed to an electrical field. Both the presence of laminin 332 and a functional β₄ integrin cytoplasmic tail were shown to be necessary for directed migration in this in vitro model. Moreover, such migration requires activation of Rac1, a finding that is consistent with several other studies.^3,105,106

Our own studies have also provided support for the idea that α₆β₄ integrin is required for directed migration. Wild-type keratinocytes move in a directed fashion along a linear track of laminin 332 matrix, whereas β₄ integrin-deficient keratinocytes move primarily in circles and lay down a laminin 332 matrix that is also in a circular pattern. Interestingly, the motility defect in the β₄ integrin-deficient keratinocytes can be corrected by either re-expressing wild-type β₄ integrin or by plating the β₄ integrin-deficient keratinocytes onto a matrix laid down by wild-type keratinocytes.¹⁰⁶ These results suggest that migration defect seen in the β₄ integrin-deficient keratinocytes is not due to a problem in the motility machinery, but rather is a consequence of an aberrant matrix deposition.

Rac1 immunoprecipates with α₆β₄ integrin in motile keratinocytes.¹⁰⁶ Moreover, Rac1 signals to activate cofilin, an actin severing protein that is involved in actin cytoskeletal remodeling in such cells.¹⁰⁶ Cofilin is activated by dephosphorylation, and our data indicate that members of the slingshot family of phosphatases are involved in this activation in migrating skin cells.¹⁰⁷ Slingshot activity, in turn, is regulated through its interaction with 14-3-3 binding proteins. When slingshot is phosphorylated, it interacts with 14-3-3 protein, leading to inactivation by sequestering the complex in the cytoplasm.^108,109 Intriguingly, α₆β₄ integrin binds 14-3-3 protein when the β₄ integrin cytoplasmic tail is phosphorylated.¹¹⁰ Thus, we speculate that this places 14-3-3 protein in a key position to regulate Rac1-mediated cell migration in keratinocytes. Specifically, inactive, phosphorylated slingshot binds 14-3-3, but on receipt of migratory signals, β₄ integrin becomes phosphorylated, allowing it to bind 14-3-3 to set up a scaffold for further signaling through Rac1.¹¹¹ In the absence of β₄ integrin, this signal pathway is compromised, leading to inactivation of both Rac1 and cofilin and perturbation in the actin remodeling machinery at the leading front of the cells. This is manifest in lamellipodial defects, in both their stability and formation, and a pronounced deficiency in directed migration¹¹¹

Collagen XVII

A second transmembrane component of the hemidesmosomes is a 180 kD type II trimeric protein with a cytoplasmic amino terminus and a long extracellular tail composed of collagen repeats, termed collagen XVII (Fig. 2).^112,113 This protein was first identified using autoantibodies in the serum of patients afflicted with the disease bullous pemphigoid (BP).¹¹⁴ Hence, it is sometimes termed BP180. Mutations in collagen XVII result in a nonlethal variant of a JEB-related blistering disease, characterized by decreased epidermal adhesion and skin fragility.⁸ Collagen XVII has been shown to contain multiple binding sites for hemidesmosomal proteins, including BPAG1e and β₄ integrin in the cytoplasm, and interacts with laminin 332 via its collagen tail and α6 integrin via its NC16A domain.^95,96 Collagen XVII is also able to link β₄ integrin and BPAG1e when the β₄ integrin binding site for BPAG1e has been deleted.¹¹⁵

The role of collagen XVII in migration is controversial. Keratinocytes derived from patients lacking the protein show enhanced migration over controls.¹¹⁶ However, studies using HaCat cells reveal that knockdown of collagen XVII expression using siRNA results in a decrease in migration in single-cell motility assays.¹¹⁷ To further complicate matters, keratinocytes isolated from a GABEB patient display an increase in motility in scratch wounds and cell scattering assays.¹¹⁶ To add to the complexity, collagen XVII is also proteolytically processed, releasing the extracellular region into the basement membrane region.¹¹⁸ One study has shown that this shed domain functions as an inhibitor of migration in a scratch wound assay, while other groups have demonstrated that the collagen XVII ectodomain is involved in both transmigration of squamous cell carcinomas and invasion of extravillous trophoblasts in the placenta.^119,120

Our recent work using lentiviral-mediated shRNA knockdown of collagen XVII in cultured keratinocytes has shown that the depleted cells have impaired lamellipodia dynamics, loss of directed migration, and reduced Rac1 activity. Although surface expression of β₄ and α₃ integrins is unchanged in the knockdown cells, loss of expression of collagen XVII results in a decreased expression of BPAG1e.¹¹⁵ This indicates a role for collagen XVII in assembly of a α₆β₄ integrin/BPAG1e signaling complex in motile skin cells.

CD151

CD151 is a member of the tetraspanin family of transmembrane proteins and is found in association with hemidesmosomes (Fig. 2).⁹⁷ All the tetraspanins share a similar structure of series of four passes through the cell membrane with one large extracellular loop that can interact with α integrin subunits and one small extracellular loop. The amino and carboxy termini are in the cytoplasm, and the carboxy tail is involved in signaling.¹²¹ These proteins are involved in cell adhesion, migration, and signaling.¹²² Tetraspanins may function as scaffolds or organizers of larger complexes of proteins, linking them to signaling enzymes and to each other. This complex has been termed a tetraspanin “web”¹²³ or tetraspanin-enriched microdomain.¹²⁴

In mouse skin, CD151 expression is up-regulated at the migrating front of the epithelium at the wound edge.¹²⁵ Moreover, using blocking antibodies, tetraspanin CD151, along with its relative CD9 and CD81, is implicated in keratinocyte wound healing in scratch wound assays.¹²⁶ Skin explant cultures from punch biopsies of neonatal skin from CD151 knockout mice also demonstrate defects in keratinocyte motility.¹²⁷ In addition, in vivo studies from the CD151 knockout mice reveal wound-healing defects.¹²⁵ The mechanisms involved in the integrin-tetraspanin interaction during keratinocyte migration are not well understood. One possibility is that CD151 plays a crucial role in integrin trafficking.¹²⁸ During migration, some integrins are endocytosed from the rear end of the cell and recycled to the leading front.¹²⁹ Interestingly, the cytoplasmic tail of CD151 contains an endocytosis/sorting motif. When residues of this motif are mutated, endocytosis of the CD151-associated integrins is compromised, and cell motility is defective as measured by scratch wound assays.¹²⁸ Moreover, knockdown of CD151 expression using RNAi technology results in defects in cell migration on laminin-332, as well as on collagen and fibronectin.¹³⁰ These authors propose a role for CD151 in regulating internalization and recycling of α₃β₁ integrin during cell migration and perhaps influencing the role of α₃β₁ integrin in maintaining polarization of migrating cells during migration by Rac1 activation.¹³⁰

Cytoplasmic components of the hemidesmosome

BPAG1e

In the skin, BPAG1e (BP230) was originally identified as a 230 kD component of the plaque of the hemidesmosome using autoantibodies present in the serum of patients afflicted with BP.¹¹⁴ It is a plakin family member sharing homology with desmoplakins and plectin.¹³¹ Several alternatively spliced proteins are transcribed from the BPAG1 gene and have distinctive distribution across epithelial, neuronal, and muscle tissues.¹³² BPAG1e, the epithelial form, is located within the intracellular hemidesomsomal plaque on the basal surface (Fig. 2).^9,114 Mice lacking BPAG1e have poorly formed hemidesmosomes, lacking keratin cytoskeleton connection, and a fragile skin prone to trauma-induced wounds and delayed wound healing.¹³³

A role for BPAG1e in wound healing is suggested by the finding that wounds in the skin of the knockout mouse heal more slowly than those in wild-type animals and is supported by the link between an increase in BPAG1e expression and invasive squamous cell carcinomas.^133,134 Moreover, on wounding skin, both β₄ integrin and BPAG1e colocalize at the leading edge of the migrating keratinocytes (Fig. 5B).¹³⁵ More direct evidence for a role for BPAG1e in keratinocyte migration derives from studies of keratinocytes that are depleted of BPAG1e.¹³⁵ They have defects in front–rear polarity, exhibit a significant reduction in lamellipodial stability, and display aberrant motility with loss of directionality.¹³⁵ Moreover, the amount of Rac1 immunoprecipitated with α₆β₄ integrin is reduced in BPAG1e-deficient keratinocytes compared with that in control cells.¹³⁵ Taken together, these results strongly suggest a role for BPAG1e in the formation of a signaling complex with Rac1 that mediates α₆β₄ integrin-regulated directed migration.

Plectin

Plectin is a high-molecular-mass (over 600 kDa) protein that also indirectly binds to α₆ integrin within hemidesmosomes (Fig. 2).⁹⁵ This plakin family member can be found in both focal contacts and the plaques of hemidesmosomes via a direct interaction with β₄ integrin.¹³¹ Plectin cooperates in some way with BPAG1e in connecting keratin filaments to the basal plasma membrane at the site of the hemidesmosome. Its N-terminal globular head region contains binding sites for the cytoplasmic tail of β₄ integrin and actin, while the C-terminal tail associates with keratin.¹³⁶ Structural studies indicate that plectin does not interact with β₄ integrin and actin at the same time.¹³⁶

Plectin plays a general role in the reinforcement of mechanically stressed cells such as those in the skin.¹³⁷ Moreover, patients with mutations within the PLEC gene often have either no plectin or a truncated form and display a trauma-induced blistering phenotype within the epidermal basal layer.¹³⁸ Fewer hemidesmosomes are formed, compromising anchorage and resulting in blister formation. Plectin-null keratinocytes migrate with higher velocities than their wild-type counterparts indicating that plectin plays a role in stabilization matrix attachments to their underlying matrix.¹³⁹

Matrix, focal contacts, and hemidesmosomes

During wound healing, migrating keratinocytes move over or under a complex of matrix proteins composed of fibronectin and fibrin (the clot) and over the collagen present in the dermis.¹ In addition, the migrating skin cells lose their connection to the laminins in the intact basement membrane and as they move, they deposit a fresh matrix rich in laminin-332.⁶ We will discuss fibronectin and laminin-332, focusing on our recent data indicating a complex interplay between the functions of fibronectin and laminin-332 in keratinocyte migration.^30,32,140

Fibronectin is a large dimeric glycoprotein containing two similar polypeptide subunits of about 250 kDa linked by disulfide bonds (reviewed in Singh et al.¹⁴¹). It is a component of blood plasma and is also secreted by fibroblasts and keratinocytes. The fibronectin molecule contains three distinct modules termed type I, II, and III fibronectin repeats. These modules are arranged along the fibronectin monomer and contain various matrix and integrin-interacting domains, with the most famous one being the arginine-glycine-aspartic acid (RGD) motif, which interacts with the focal contact-associated α₅β₁ integrin to promote cell attachment and migration of keratinocytes.^15,35

Laminin-332 is a major component of the epidermal basement membrane. The structure and functions of laminin-332 have been extensively described in a recent review and so, we will only briefly discuss laminin-332 structure here.⁹⁸ It is a heterotrimer that is composed of α₃, β₃, and γ₂ subunits. In intact skin, laminin-332 protein is the ligand for the integrin that is associated with the hemidesmosome and provides firm anchorage to the underlying matrix (see above).⁹ The importance of this component of the basement membrane of skin in maintaining the interaction of the epidermis and the dermis is highlighted by the effects of mutations in the subunits of laminin-332, the genetic defect that is the underlying cause of blistering in a variant of JEB.⁸ Loss of laminin-332 expression in the skin of these patients leads to dysadhesion of the epidermis at the level of the basement membrane.⁸

Laminin-332 is an atypical laminin with short β and γ subunit side arms compared with the comparable side arms in laminin-111.⁹⁸ Moreover, both the α₃ subunit and the γ₂ laminin subunits are subject to proteolytic processing after secretion.^142,143 In the case of the α₃ subunit, this processing occurs in the C-terminus such that the mature molecule is missing domains 4 and 5 of its G domain.¹⁴² There is also evidence of processing occurring in the N-terminus of the α₃ subunit, while the γ₂ subunit undergoes an N-terminal cleavage.⁹⁸ During tissue remodeling, additional cleavage of the γ₂ subunit appears important in allowing cells to break their connection with the basement membrane. In this regard, syndecan-1 interacts with the short arm of the γ₂ chain.¹⁴⁴ This interaction, in turn, suppresses EGF-stimulated phosphorylation of the β₄ integrin subunit, inhibits hemidesmosome disassembly, and promotes migration into the wound site.¹⁴⁴

In addition to its role in stabling anchoring the keratinocytes in the basal layer of the epidermis to the dermis, laminin-332 is a crucial mediator of the migration of keratinocytes. Laminin-332 is not only localized to the wound edge of the epidermis, but it also supports the migration of cells in vitro.⁶ Moreover, live cell imaging of keratinocytes expressing a tagged subunit of the laminin-332 protein demonstrates that migrating cells deposit laminin-332 in linear arrays or tracks along which they move.^106,145

Laminin-332 containing an unprocessed γ₂ subunit has also been reported to drive the hypermotility of keratinocytes.¹⁴⁶ Moreover, there is evidence that matrices rich in laminin-332 containing uncleaved α₃ subunits support migration and that they do so in an α₃β₁ integrin-dependent manner.⁶ In addition, several heparin-binding domains have been identified in the G4 and 5 domains of the unprocessed α₃ laminin, suggesting that heparan sulfate proteoglycans could interact with these regions.^47,147,148 Moreover, a synthetic peptide encompassing the syndecan-binding domain of α₃ laminin LG4 domain has been shown to promote keratinocyte migration by its interaction with syndecan-4, induction of conformational changes, and activation of β₁ integrin. Interestingly, inhibition of p38MAPK partially blocked the ability of keratinocytes to migrate in the presence of the soluble peptide, implying syndecan-4 ligation to laminin-332 containing an unprocessed α₃ activates a migratory signaling cascade involving p38MAPK.⁴⁷ Further support for a syndecan-dependent motility inducing role for the G4 and 5 domains of the α₃ laminin subunit derives from the finding that mutations perturbing syndecan-4 binding inhibit keratinocyte motility.¹⁴⁷

Interestingly, the ability of laminin-332 to sustain migration appears fine-tuned by fibronectin. Specifically, addition of fibronectin to laminin-332 deposited matrix leads to a reduction in speed of the motile cultured keratinocytes, while enhancing their adhesion to substrate.³² We speculate that fibronectin acts to prevent detachment of keratinocytes by actively migrating over the wound bed in a laminin-332 dependent fashion.³²

Crosstalk between focal contacts, hemidesmosomes, and their proteins

Although we have described focal contacts and hemidesmosomes in separate sections, their functions in adhesion and migration are closely interlinked.¹² Moreover, although they share only one component, namely plectin, it is becoming increasingly clear that there is crosstalk between the proteins in focal contacts and hemidesmosomes. We have already mentioned data suggesting that the focal contact proteins kindlin and syndecan regulate hemidesmosome protein localization and possibly hemidesmosome protein functions. In addition, the focal contact protein Flii may be involved in this phenomenon, as mice overexpressing Flii have poorly formed hemidesmosomes and more fragile skin compared with wild-type mice.⁹³ Another example of crosstalk is the regulation of translation of the α₃ integrin subunit by α₆β₄ integrin-driven activation of the Akt-mTOR pathway in an initiation factor 4E binding protein 1-dependent manner.³⁰ Moreover, it was also recently shown that α₃ integrin subsequently regulates the transcription of α₂ integrin via an unknown mechanism.³⁰

Future Directions

Further work on the exact manner via which α₆β₄ integrin and the BP antigens regulate signaling that determines motile behavior is clearly needed. Moreover, their role in the context of the complex skin tissue environment also needs elucidation. Knockout mice have proved useful in dissecting out the functions of a number of proteins in wound healing, but mice are problematic for testing the activities of proteins that are involved in epithelialization. This is because mice wounds heal by contracture rather than epithelialization, the process via which human skin wounds repair. Thus, technical improvement in animal models such as the use of splints that prevent contracture and the use of three-dimensional culture systems may circumvent this problem in the future.¹⁴⁹

The hemidesmosome is associated with keratin, but its components, most notably α₆β₄ integrin and plectin, also show association with the actin cytoskeleton.²² The mechanisms underlying the regulation of such associations that appear central in their role in migration still require elucidation. Moreover, intriguingly, a role for the keratin cytoskeleton in wound healing is also indicated. Specifically, the wound-induced keratin 6 negatively regulates Src kinase and the migratory behavior of keratinocytes.¹⁵⁰ However, more work in this area is still needed to clarify whether keratins play a positive role in motility.

Finally, a better understanding of focal contact/hemidesmosome protein crosstalk should provide novel clues to the molecular mechanisms that regulate both adhesion and migration in skin cells. Such studies may reveal new strategies for the treatment of not only blistering skin diseases but also the treatment of chronic wounds and other defects in the repair of skin after tissue damage.

Take-Home Messages.

• • Focal contacts and hemidesmosomes and/or their components play critical roles in wound repair.

• • Focal contact proteins regulate the traction forces and signals that are necessary for forward movements of cells as they migrate over a wound.

• • Although the hemidesmosome is disassembled to allow keratinocytes to move onto the wound bed and/or over the provisional wound matrix, hemidesmosome proteins play important regulatory and signaling roles in determining aspects of the motile behavior of skin cells.

• • Crosstalk between focal contact and hemidesmosome proteins appears central to the molecular mechanisms and fine-tuning of the repair response.

Abbreviations and Acronyms

BP

bullous pemphigoid

BPAG

bullous pemphigoid antigen

ECM

extracellular matrix

EGF

epidermal growth factor

FAK

focal adhesion kinase

FERM

4.1 protein, ezrin, radixin, moesin

Flii

Protein flightless-1 homolog

ILK

integrin-linked kinase

JEB

junctional epidermolysis bullosa

JNK

c-Jun amino-terminal protein kinase

MAPK

mitogen-activated protein kinase

Acknowledgments and Funding Sources

The research in the Jones lab is supported by grants from the NIH (RO1 AR054184 and RO1 HL092963). They are very grateful to Dr. John Olerud, Dr. Robert Underwood, and their coauthors for providing original micrographs from their publication for the generation of Fig. 5.¹⁰⁴

Author Disclosure and Ghostwriting

The authors disclose no commercial associations or known conflicts of interest in connection with this work. No ghostwriters were used to write this article.

About the Authors

The authors are members of the Department of Cell and Molecular Biology at the Feinberg School of Medicine at Northwestern University in Chicago. Their research centers on the functions of matrix receptors and matrix in both the adhesion and motility of keratinocytes. Jonathan C.R. Jones, PhD, is Professor of Cell and Molecular Biology and Dermatology.