Chapter 248. Mechanisms of Wound Repair, Wound Healing, and Wound Dressing

The terms “wound healing”, “wound repair”, or “tissue repair” are often used interchangeably, but actually “healing” and “repair” point to different sets of events and outcomes. First of all, before any distinction is made, one must recognize the rather obvious fact that “healing” and “repair” are not confined to the skin, but can encompass any organ system. Technically, wound healing is a term that should be used only in the context of true regeneration, when the original architecture and structure of an organ or anatomic part is completely restored to the way it was before injury. More primitive animals, such as small amphibians and reptiles, are still capable of this type of regeneration. However, as animals became larger and more complex during evolution, true regeneration was no longer possible. The human fetus is still largely capable of regeneration (especially in the early stages) but, in adults and with the possible exception of the liver (probably a compensatory enlargement and not regeneration), true regeneration does not take place. Rather, man and other higher vertebrates heal by a process of repair (wound repair or tissue repair), whereby the eventual outcome is not true anatomic restoration but a functional compromise. Still, because of well established terms and the published literature, even within our discussion here, we may at times use “healing” and “repair” interchangeably; we will be more specific when truly referring to the process of tissue regeneration.

There are probably evolutionary reasons for why repair occurs in higher vertebrates compared to true healing or regeneration. Teleologically and from an evolutionary standpoint, the process of repair for higher animals needed to be rapid, economical from an energy standpoint, and allow for the immediate survival of the organism. However, by necessity, repair leads to a rapid solution to injury and thus to scarring. Another important consideration is that most of the mechanisms of wound repair that have evolved are aimed at addressing acute tissue injury and not chronic conditions. Again from an evolutionary standpoint, humans were not meant to develop degenerative diseases or live long enough to develop arterial, venous and pressure ulcers, or neuropathic ulcers from diabetes. Therefore, humans are quite unprepared for these types of chronic wounds, and there are no specific mechanisms that have evolved to deal with them in a truly effective way.1–3 It should be noted that the extent of injury and its depth are important considerations. Thus, shallow wounds (i.e., shave biopsies) that retain skin appendages (hair follicles, sweat glands, etc.) in the wound bed have the capacity to heal from the “inside”, since keratinocytes associated with those skin appendages are still present. On the other hand, full-thickness wounds (punch biopsies are a good example) have to rely on keratinocyte migration and proliferation from the edges (Fig. 248-1). Not surprisingly, full-thickness wounds are associated with delayed healing and more scarring.2 Some animal models, particularly those in which certain genes are overexpressed or knocked-out, are providing greater understanding of the role of certain proteins and mediators.4–7 Impaired healing is found in mice with combined deficiency of molecules critical in inflammation (E- and P-selectins), and in mice without plasminogen, uPA and tPA (double knock-out), fibroblasts growth factor-2 (basic FGF or bFGF), or inducible nitric oxide (iNOS). On the other hand, impaired or delayed healing occurs in transgenic mice overexpressing some tissue metalloproteinases (MMP-1) and antisense to CD44, the receptor for hyaluronic acid. Some induced mutations lead to accelerated healing, as reported with Smad-3 or skn-1a knock-out mice.8 In the future, it may be possible to use these clues to stimulate the healing process in humans.

In general, there are four recognized phases that characterize the cutaneous repair process: (1) coagulation, (2) inflammatory phase, (3) proliferative and migratory phase (tissue formation), and (4) remodeling phase. The coagulation and inflammatory phases ore sometimes grouped together, so great is the overlap of mediators that are released. In a diagrammatic form, Fig. 248-2 illustrates these phases of healing in sequence, while Fig. 248-3 shows specific events that take place during the different phases. We will discuss these phases and identify the main components and events that characterize them. The cell types primarily involved in wound healing have been regarded to be the platelets, neutrophils and macrophages, fibroblasts, endothelial cells, epithelial cells. More recently, increasing importance is accumulating for the role of lymphocytes, either directly or indirectly.9 However, it should be noted that breaking down the overall process of wound repair into these seemingly defined phases is artificial, as they overlap considerably.2,3 Another consideration is that the occurrence and hypothesis of these phases of wound healing has been determined in laboratory animal experiments; it is assumed that the same processes are involved in human tissue repair.

Coagulation and Inflammatory Phases

Although we have shown these as separate phases, the degree of overlap is such that it is appropriate to discuss them together. These early phases begin immediately after an acute injury. Disruption of blood vessels leads to local release of blood cells and blood borne elements resulting in clot formation. While the blood clot within the vessel lumen provides hemostasis, the clot within the injury site acts as a provisional matrix for cell migration,10 further formation of new extracellular matrix (ECM),11 and a reservoir for cytokines and growth factors.12 The initial component of this phase is dominated by the platelet, which directs clotting of the fresh wound by the intrinsic and extrinsic pathways. Platelets also release a number of chemotactic factors that attract other platelets, leukocytes, and fibroblasts to the site of injury. Leukocytes are slowed down within the blood stream through the expression of selectins, which, coupled with integrins, bring inflammatory white cells into the wound.5 These cells have multiple roles, including debridement of necrotic material and bacteria, as well as the production of certain critical cytokines. A case in point is the production of connective tissue growth factor (CTGF) by inflammatory cells and expressed in wounds.13 After the first few days, the neutrophils are removed by macrophages.

Fig. 248-4 is a diagrammatic representation, using a tissue section, of the events that take place after injury. The fibrin plug resulting from the initial injury provides a temporary wound coverage and consists of platelets embedded within a complex meshwork of mainly polymerized fibrinogen (fibrin), fibronectin, vitronectin, and thrombospondin.3,14 Cross-linking of fibrin by Factor XIII also appears to be critical, and there is evidence that the function and probably migration of keratinocytes is impaired by uncross-linked fibrin.15 As they immediately aggregate, platelets release a wide range of growth factors, including platelet-derived growth factor (PDGF) and transforming growth factor-β1 (TGF-β1). In fact, platelets are the main storage site for the TGF-β1 isoform. The acute injury environment, with its hypoxia, proteases, and low pH, contributes to the activation of these growth factors.2 The principal mediators themselves produce others from their fragmentation or polymerization. A classic example is the chemotactic fibrinopeptides A and B, which are produced from the action of thrombin on fibrinogen.15,16 Another example is the formation of bradykinins, as well as C3a and C5a, which are activated by Hageman factor.17

As the inflammatory component of this early phase continues, within 24–48 hours after injury, monocytes replace neutrophils and become the predominant leukocyte. Monocytes are attracted to the injury site by some of the same chemoattractants responsible for recruitment of neutrophils, such as kallikren, fibrinopeptides, and fibrin degradation products.18 Other, more specific chemoattractants then take over in recruiting monocytes, and they include fragments of collagen, fibronectin, elastin, and TGF-b1. Monocytes undergo a phenotypic change to tissue macrophages and, unlike neutrophils, they are critical for the progression of wound healing.19,20 Macrophages phagocytose and kill bacteria, and scavenge tissue debris.21 They also release several growth factors, including PDGF, fibroblast growth factor (FGF), TGF-b, thereby stimulating migration and proliferation of fibroblasts, as well as production and modulation of extracellular matrix. The macrophage is generally regarded as the master cell in wound healing.22 However, this is an oversimplification, as it is important to consider that what may delay or impair healing is not so much the presence or absence of inflammation and certain cells but, rather, an inappropriate inflammatory response.17 For example, there is evidence that healing may occur in the absence of an inflammatory infiltrate.23,24 Conversely, experiments in mice constitutively expressing the chemotactic cytokine IP-10 show that an intense inflammatory infiltrate can impair neovascularization and the formation of appropriate granulation tissue.25 Therefore, the true role of inflammation in tissue repair remains somewhat controversial from an experimental point of view.26,27 From the clinical standpoint, one observes that in certain cutaneous wounds, such as pemphigus or pyoderma gangrenosum, downregulation of inflammation with the use of corticosteroids is effective. Possibly, modulation and “correction” of the inflammatory response by corticosteroids may be helpful in these selected clinical entities.28

Proliferation/Migration and Remodeling Phases

Fig. 248-4 illustrates a representation of the fibrin plug in place and certain events that occur in the next two phases of healing, the proliferative/migratory and remodeling phases. As inflammation has already played its major role early on after injury, the most important event in cutaneous wound healing now becomes reepithelialization. This critical event is not based on keratinocytes alone, because there is a high interdependence between keratinocyte movement through the provisional fibrin matrix, recruitment of fibroblasts and endothelial cells, and extracellular matrix formation (Figs. 248-2 and 248-3). Tissue matrix metalloproteinases (MMPs) and other enzymes (tPA and uPA) are critical to the movement of cells through provisional structural matrix components, as well as for freeing the keratinocytes at the edge of the wound from their hemidesmosomal and desmosomal attachments. Another critical set of molecules that help guide this migratory components are the integrins.29,30 The integrins, consisting of at least 24 αβ heterodimers (18 α and 8 β subunits), are transmembrane cell surface receptors that bind the extracellular matrix (ECM) to cytoskeletal structures.31 The integrin profile is very dynamic during the repair process. For example, dermal fibroblasts undergo a switch from α2 to α3 and α5 integrin subunits. As another example, endothelial cells cannot respond to angiogenic stimuli without the expression of αvβ5 integrin. Certain polypeptide growth factors are absolutely essential to angiogenesis, including basic fibroblast growth factor (bFGF) and vascular endothelial growth factor (VEGF).3,31 Table 248-1 shows some of the main cytokines and growth factors shown to play a role in the repair process. One must realize that such tables are oversimplifications of very complex biologic properties of these polypeptides. Only the main effects are noted, and it must be borne in mind that the actions of growth factors are often context specific and not the same in every biological situation.

Evidence points to low oxygen tension as an important early stimulus for fibroblast (and endothelial cell) activation. Fibroblast replication and longevity are enhanced in hypoxia,32 and low oxygen tension stimulates clonal expansion of dermal fibroblasts seeded as single cells.33 Moreover, the synthesis of a number of growth factors is enhanced in hypoxic cells. Macrophages secrete an angiogenic substance only when they are exposed to low oxygen tension. This reversible effect was observed at an oxygen tension of 15–20 mm Hg.34 TGF-β1 transcription and peptide synthesis are enhanced in cultures of human dermal fibroblasts exposed to a similar level of hypoxia.35 In addition, hypoxia upregulates the synthesis of endothelin-1,36 PDGF B chain,37 and VEGF38 in endothelial cells. It appears that, at least in some cases, the effect of hypoxic conditions is mediated by hypoxic inducible factor-1 (HIF-1), a DNA binding complex shown to contain at least two basic helix-loop-helix PAS-domain proteins.39

The importance of the hemidesmosome and desmosome is well known for many dermatological diseases where specific defects in their overall structural integrity are present (i.e., pemphigus, bullous pemphigoid, epidermolysis bullosa, etc.). Now, however, during the healing process and for keratinocyte migration to occur, there is a need to break down these complex structures anchoring the basal keratinocytes to the basement membrane and neighboring keratinocytes. This breaking down process is just as complex as the structure itself, and involves interactions between MMPs, integrins, growth factors, and structural proteins. In the normal resting state, laminin-5 is bound to α6β4 integrin, the latter linking the intracellular keratin filaments of keratinocytes to the basement membrane. In large part due to the interactions of integrins (including their phosphorylations status) with the ECM and receptor clustering on the surface of keratinocytes, there are important morphological changes (such as lamellipodia formation) that are required for keratinocyte locomotion.3,40–42 Molecular GTPases switches (Rho, Rac, Cdc42) are involved. Partially as a result of phosphorylation of the α6β4 integrin, α3β1 integrin promotes lamellipodia formation and keratinocyte locomotion. It has often been stated that migration of keratinocytes is key to resurfacing of the wound and that in fact such resurfacing is not dependent on proliferation.3 Overall, one must add, epithelial-mesenchymal transition (EMT), also important in cancer and epithelial adhesion and gain of mesenchymal features, empowers the epithalial cells to migrate in a way that is reminiscent of the embryonic stage. Recently, this process was reconstituted in vitro by exposure of epithelial cells to tumor necrosis factor-α (TNF-α), which led to the expression of vimentin, FSP1, and MMPs. The TNF-α mediated EMT may be secondary to bone morphogenetic proteins (BMP).43

In addition to the critical role of migration, keratinocytes do proliferate within the first hours after injury, and this is certainly more evident when the gap or size of the defect cannot be bridged temporarily with cell movement alone and is dependent on the cells beyond.2,3 By a week after injury, the repair process is well underway. Cytokines and growth factors (EGF, PDGF, TGF-β, FGFs, VEGF), matrix components, and MMPs continue to play an active role. In addition to the growth factors themselves, signaling proteins are critical. For example, certain kinases (MAPK) are activated in basal and suprabasal keratinocytes by further action of integrins or the release of interleukin (IL)-1α. Ion fluxes, including calcium levels and entry into the cells, are also critical to the overall process of keratinocyte migration and skin resurfacing.44 There is evidence that, in addition to lamellipodia extension, wound edge suprabasal keratinocytes may “leap-frog” over the basal cells near the wound. However, this system may not be very efficient, certainly not to the extent of the purse-string mechanism occurring in fetal healing and corneal healing, which is accompanied by other important differences in matrix deposition and growth factor profile.5,45 Wound contraction, rather than epithelialization, is common in major injuries, including burns. It is also the preferred mode of repair in many animal species, such as rodents. In experiments using knockout mice, MMP-13 has been implicated in keratinocyte migration, angiogenesis, and contraction, while the role of MMP-9 appears critical in keratinocyte migration.46

The remodeling of the ECM, as well as the movement of cells, is highly dependent on MMPs and serine proteases.47 A very important component of this dependence on MMPs is the MMPs-driven degradation of ECM and the resultant exposure of selective bioactive ECM segments that influence cell behavior, including migration and proliferation.47 The remodeling phase begins 5–7 days after injury.11 Upregulation of tPA and uPA are important for keratinocyte migration, which may depend on cross-talk and interactions between α3β1, keratinocytes, and collagen. These events lead to the induction of MMP-1 (collagenase-1 or interstitial collagenase), important in keratinocyte migration and epithelialization.11 MMP-9 plays a fundamental role in “cutting” Type IV and type VII collagen, which are essential components of the basement membrane and anchoring fibrils, and it also promotes inflammation and neutrophil migration. MMP-10 (stromelysin) breaks down other noncollagenous ECM components and facilitates migration.2,11,14,48 Other mediators, such as thymosin-β, have been found to upregulate MMPs during wound repair.49 Of critical importance during the remodeling process is the phenotypic switch in certain cell subpopulations from fibroblasts to myofibroblasts.2,50 Therefore, while the early process of healing relies heavily on matrix accumulation, which in turn facilitates cell migration, there is now a need to dampen the ECM formation and degrade it to a level that at least approximates the preinjury state. However, the remodeling phase is more than a breakdown of excess macromolecules formed during the proliferative phase of wound healing. Cells within the wound are returned to a stable phenotype, extracellular matrix material is altered (i.e., collagen type III to type I), and the granulation tissue that was so exuberant during the early phases of wound healing disappears.2,3

Over the last several years, considerable attention has turned to the role of stem cells from the follicular and interfollicular epidermis in wound healing. Using murine models, it has been reported that, at least after injury, cells from the hair-follicle bulge are recruited to the epidermis and migrate to the center of the wound.51 In subsequent studies, it has been shown that hair follicles form de novo after wounding and that epidermal cells in the wound assume a hair follicle stem cell phenotype. Some of these findings may be dependent on Wnt signaling.52

In wound healing, the extracellular matrix (ECM) comprises four main components:47 (1) structural proteins (i.e., collagen, elastin); (2) multidomain adhesive glycoproteins (i.e., fibronectin, vitronectin, laminin); (3) matricellular proteins, including secreted protein acidic and rich in cysteine (SPARC), thrombospondin 1 and 2, tenascins, osteopontin; and (4) glycosaminoglycans (GAG), including hyaluronic acid and proteoglycans, that is, syndecans, perlecan (such as chondroitin sulfate and heparan sulfate). The glycosaminoglycan (GAG) hyaluronic acid (hyaluronan) is a particularly abundant component of the provisional matrix, and one whose deposition needs to be modified during the remodeling process. GAGs are often around other ECM proteins, including collagen and elastin.47 High levels of hyaluronic acid and fine reticular collagen are present in the early embryo, where they are thought to offer less resistance to cell migration.53 Indeed, embryonic wound repair is characterized by a hyaluronic acid–rich environment, thought to be at least in part responsible for the “scarless” healing of embryonic wounds. Fibroblasts from early granulation tissue produce large amounts of hyaluronic acid, and proliferating cells express CD44, which is the receptor for this GAG molecule.54,55 Besides the issue of offering less resistance to cell movement, hyaluronic acid may stimulate cell motility by altering cell-matrix adhesion. One example of this is that hyaluronic acid weakens the adhesion of heparan sulfate and fibronectin.56 Perhaps more importantly from a physical/spatial standpoint is that hyaluronic acid creates a highly hydrated structure leading to tissue swelling and interstitial spaces, and thus an environment more conducive to cell movement.2,57 The effects of hyaluronic acid are also regulated by growth factors and cytokines. Upregulation of the hyaluronic acid expression and its receptors (for example RHAAMM) by TGF-β1 stimulates fibroblast motility.58 Eventually, as remodeling occurs, hyaluronic acid is degraded by hyaluronidase and replaced by sulfated proteoglycans, which contribute a stronger structural role in late granulation tissue formation and in scars, while being less able to stimulate cellular movement. Two major proteoglycans, chondroitin-4-sulfate and dermatan sulfate, are produced by mature scar fibroblasts.

Three main classes of collagens are present normally in connective tissue: (1) fibrillar collagens (types I, III, and V), (2) basement membrane collagen (type IV), and (3) other interstitial collagens (types VI, VII, and VIII). These are examples of the different types of collagen present in skin. Importantly, however, the fibrillar collagens serve as the major structural collagens in all connective tissues.2,57,59 During the initial phases of wound repair, it appears that the wound tends to recapitulate the processes involved in embryogenesis.45 Thus, granulation tissue is initially comprised of large amounts of type III collagen, which is a minor component of adult dermis and indeed is present in larger amounts in fetal wound repair. During the phase of remodeling, type III collagen is gradually replaced by Type I collagen. Type I collagen replacement is associated with increased tensile strength of the scar. However, the final tensile strength of a scar is only about 70% of that of preinjured skin.3,57 The process of converting the collagen content of the dermis from type III to type I collagen is controlled by interactions involving synthesis of new collagen with breakdown of old collagen.14 Key to this process of conversion are matrix metalloproteinases (MMPs) and, specifically, the collagenases.

Matrix-degrading metalloproteinases (MMP) are proenzymes that need to be activated, and are considered to be the physiologic mediators of matrix degradation.48 The prototypic MMP is interstitial collagenase, but over 20 such enzymes (zinc-dependent endopeptidases) have been described.11 We have already mentioned these critical proteinases in the context of keratinocyte migration, but it is important to go over them in more detail. Five groups of these enzymes have been identified: (1) collagenases, (2) gelatinases, (3) stromelysins/matrilysins, (4) membrane-type MMPs, and (5) other types. Table 248-2 is a summary of certain MMPs that have a prominent effect in wound healing, and which supplements this discussion. The collagenases include interstitial collagenase (fibroblast collagenase, MMP-1), which acts on collagens I, II, III, VII, and X. Collagen type II is a particularly good substrate for MMP-1. Another important member of the collagenase class is neutrophil collagenase (MMP-1), which also degrades type II and III collagens but is particularly active against type I collagen. Gelatinases break down denatured collagen (gelatin). Among the most important gelatinases are gelatinase A (MMP-2), which breaks down gelatins, collagen IV, and elastin. Another key gelatinase is gelatinase B (MMP-9), which is produced by many cell types, including macrophases, neutrophils, and keratinocytes. Stromelysins have a relatively broad substrate specificity. Both stromelysin 1 (MMP-3) and 2 (MMP-10) act on proteoglycans, fibronectin, laminin, gelatins, and collagens III, IV, and IX. Another member of the stromelysin family, matrilysin (MMP-7) degrades mainly fibronectin, gelatins, and elastin. A member of the MMPs family is epilysin (MMP-28), which appears to be produced by proliferating keratinocytes distal to the wound edge. It may be needed to restructure the basement membrane.3,48,57 Deficiency of MMP-7 (matrilysin-1), which regulates inflammation, epithelialization, and inhibits apoptosis, is reported to lead to the most severe wound healing defect associated with MMPs.11

One of the most important clinical observations of the last several decades is the realization that wounds kept moist re-epithelialize faster.60–63 The evidence for this is best for acute wounds. However, even in chronic wounds, the moisture-retentive dressings that have been developed to create moist wound conditions do lead to a number of desirable outcomes, such as pain control, painless autolytic debridement, and stimulation of granulation tissue.63–66 There have been fears that keeping the wound moist will cause infection, but this fear is unfounded.67–69 Nevertheless, occlusion of the wound is still contraindicated in the presence of infection. There are a variety of wound dressings available to the clinician and which fit the particular clinical situation.65,66,70,63 The main types of dressings include transparent films, hydrocolloids, foams, gels, alginates, and the relatively new collagen products (Table 248-3). In determining the most appropriate dressing for a particular wound, the clinician must take into consideration the need for absorption of excessive exudate (foams and alginates), whether the wound is too dry and needs additional moisture (gels or hydrogel materials) and whether the wound and its epithelial edges can tolerate the often subtle but serious trauma that comes from removal of adhesive dressings such as films. Thin contact layers, consisting of different polymeric materials with perforations, allow wound fluid to escape and are useful in preventing tissue injury upon dressing changes by minimizing removal of the primary dressing in direct contact with the wound.

The exact mechanisms by which moist wound conditions facilitate keratinocyte migration are unknown.2,66,70 A number of explanations, some supported by experimental work, have been proposed and appear plausible. Intuitively, a dry wound bed and the presence of a scab or crust may render keratinocyte migration more difficult. Some have proposed that the electrical gradient is altered in a beneficial way when wounds are occluded and kept from getting dry. Certainly, there is considerable amount of work indicating that electrical currents can indeed affect the healing process.71–73 Retention of growth factors and cytokines in the wound bed has been proposed as yet another explanation for how moisture retentive dressings work. The wound fluid taken from acute wounds has been shown in vitro to stimulate the proliferation of a number of cell types, including fibroblasts and endothelial cells.74 Conversely, there is work indicating that the opposite is true for wound fluid harvested from chronic nonhealing wounds.75 Whether this difference in the wound fluid characteristics helps explain the greater effect of moist wound healing on keratinocyte migration in acute wounds remains unknown. Indeed, another view has been to remove the wound fluid from chronic wounds, which can be accomplished with certain devices consisting of a film dressing applied to the wound and connected to a source of negative pressure.76

The biology of acute wound healing has direct clinical application to dermatologic surgery. For example, reconstructions following Mohs micrographic surgery routinely involve the four main forms of acute wound healing that include secondary intention healing, side to side closure, cutaneous flaps, and skin grafts. Secondary intention healing is a variation of acute wound healing in which the wound is left to heal on its own. The biology of this process is characterized by all of the usual steps of wound healing but differs in the predominance of the inflammatory reaction and granulation tissue, as well as that of wound contraction.77 Side to side and flap closures are examples of primary intention healing and follow all of the phases of wound healing described previously.

The healing of a skin graft, however, is quite different from what was described earlier with acute injury. One distinguishing feature of a skin graft is the complete dependence of the graft on the recipient wound bed for revascularization,78–80 a feature that requires several unique physiological events. On the basis of histological studies, graft healing has classically been divided into three identifiable phases: (1) imbibition, (2) inosculation, and (3) neovascularization.81

Imbibition Phase

As in conventional acute wound healing, the immediate event is the formation of a fibrin clot at the graft interface,82 an event that is also accompanied by infiltration of leukocytes from the recipient to the graft.83 What distinguishes a skin graft from a conventional wound is the diffusion of plasma from the recipient site into the overlying skin graft, resulting in a weight increase of the graft of up to 40% within the first day.84 This net diffusion of the plasma into the graft was termed imbibition, and its significance had initially been thought to be limited to keeping the graft moist and the graft vessels patent.84,85 Other investigators had hypothesized that imbibition also played a role in providing nutrition to the graft,86,87 but support for this nutritional role was only demonstrated years later, when bromodeoxyuridine was shown to incorporate into skin graft cells after systemic injection in a mouse model.88

The imbibition phase ends as the venous and lymphatic drainage is reestablished, with a concomitant reduction in the weight of the graft.84 In a rabbit graft model, lymphatic drainage of contrast material has been shown to begin slowly on day 2 and to improve until day 12, with histologically visible anastomoses of the graft and recipient lymphatic vessels by day 4.89 A fundamental biological issue of skin graft healing has been to determine the relative contributions of the graft and the recipient tissues to vasculature of the graft, a more precise understanding of which has been achieved by recent molecular biological approaches.

Revascularization Phases of Inosculation and Neovascularization

Inosculation was originally proposed as the anastomoses of graft blood vessels with the recipient-derived vessels, and neovascularization as the ingrowth of new blood vessels from the recipient wound bed into the graft. The early histological studies had suggested that the graft maintained its own vasculature after grafting,90,91 in support of the idea of inosculation. However, others felt that histological studies could not distinguish graft (or donor-derived) from recipient-derived vessels. Lambert concluded that at least during the first several days, the vessels of a skin graft were donor-derived92 based on combined histology and radiographic methods. Using a human tissue–engineered skin equivalent with a network of endothelialized vessels,93 Auger and his group demonstrated a significantly increased speed of vascularization by inosculation compared to the rate of vascularization of the nonendothelialized skin equivalent,94 highlighting the higher efficiency of inosculation over that of neovascularization.

More precise dissection of the relative roles of graft and recipient in the creation of the final vasculature had to await molecular biologic methods to distinguish graft cells from recipient cells. Specific in-situ hybridization and immunolabeling were used to distinguish graft from recipient in studies of human split thickness skin grafted onto nude mice. It was found that the recipient endothelial cells grew into the preexisting donor capillary tubes, decisively supporting the process of immediate inosculation.95 Most recently, transgenic mice were used to distinguish donor-derived from recipient-derived vascular growth to reveal an even more precise picture. Capla and colleagues showed regression of the graft vessels starting at day 3, with simultaneous vessel ingrowth from the recipient primarily into the graft's preexisting vascular framework, resulting in inosculation by day 7.96 Presumably, neovascularization without anastomosing with the graft's vascular framework is not the predominant event. Using a mouse bone marrow transplant model, the same investigative group further showed that bone marrow-derived endothelial progenitor cells contributed 20% of the recipient-derived endothelial growth within the graft.96

Other Processes in Skin Grafts

As with the blood vessels, there have been analogous issues about the relative contributions of graft and recipient tissue for the final reinnervation of the graft.97 In a porcine graft model, there was a suggestion based purely on histological studies that recipient nerves enter the graft and follow the course of the graft tissue's preexisting neurolemmal sheaths.98 In another study, human skin was grafted onto nude mice in which the graft structures could be distinguished from recipient tissue using species-specific antibodies. In that report, neurites from the recipient were shown to grow into the graft but did not appear to connect with preexisting donor graft nerve trunks.95 Using histochemical labeling of cholinesterases, other studies had suggested that the reinnervation of the graft also involved chemotactic attraction of growing neurites to adnexal structures.99

Much of the remainder of skin graft healing involves events in common with the normal wound healing process. Infiltration by fibroblasts in the graft occurs 3 to 5 days after grafting, followed by a progressive increase in both graft and recipient fibroblasts within the graft.95 Full-thickness skin grafts are thought to reduce wound contraction compared to the significant contraction found in secondary intention healing, with one study suggesting that a graft in the exact size of the original defect minimizes this contraction.100 Hyperpigmentation can be an occasional undesirable side effect of grafting, and histological changes of melanocytes have been observed after grafting.101 It has been proposed that skin grafts, in a fashion similar to that of cultured autografts, may also stimulate wound healing of the recipient site, especially after prewounding of the donor site from which the grafts were harvested.102,103 In further support of this hypothesis, expression of keratinocyte Ki-67 and β-1 integrin after grafting has been noted, in addition to the production of stimulatory growth factors and cytokines. This may imply an additional phase of graft healing.104

Finally, the mechanisms of contraction of the actual graft are unclear; however, a working model proposes that keratinocytes and myofibroblasts separately contract the graft in two phases. During the initial phase, which occurs during the first two days, epidermal cells contract the graft in a purse string fashion.105 The initial epidermally mediated contraction is followed, starting at 3–5 days, by a longer contractive phase mediated by myofibroblasts (which have differentiated from fibroblasts).106 The contraction phases have been modeled in vitro using collagen gels106 and tissue engineered skin.107 Among the findings from in-vitro studies is that TGF-β does not drive graft contraction as it does hypertrophic scarring.108

The previous discussion has focused on cutaneous wound repair after acute injury. It must be recognized, however, that the applicability of those events and processes is limited when it comes to chronic wounds, such as in diabetes, venous, and arterial insufficiency, and in a variety of situations complicated by inflammatory processes and faulty host response. The linear and one-way relationship between the different phases of wound repair is lost in chronic wounds (Fig. 248-5). There is no specific orderly sequence, and parts of the chronic wound may be in different phases at different times. Acute wounds, such as those created by surgery or by trauma, have a predictable time-frame for healing and generally heal quite readily. However, chronic wounds display what has been called a “failure to heal” or impaired healing. A number of observations have been made with regard to chronic wounds, and hypotheses have been proposed for impaired healing. It must be noted that it has proven difficult to study chronic wounds from a pathophysiological standpoint. A major drawback has been the unavailability of animal models that truly display impaired healing.

Some chronic wounds are the complex result of ischemia, pressure, and infection. A case in point for this pathophysiological triad is the diabetic ulcer.3 There is still considerable controversy whether hyperglycemia itself plays a pathophysiological role in the development of diabetic ulcers, although neutrophil function is impaired and thus the propensity to infection is enhanced in the diabetic state.3 Importantly, the notion of “small vessel disease” in diabetes has been more properly evaluated and shown not to be an obstructive phenomenon, but rather a more physiological one.109 This being the case, revascularization of the diabetic foot is now viewed as the right approach in the presence of large vessel disease and good run-off circulation. Perhaps the best example of truly impaired healing, not related to undue pressure and poor arterial supply, is venous ulceration. The underlying abnormality in the development of venous ulcers is venous hypertension, which refers to the inability of venous pressure in the foot to decrease in response to exercise.110 There have been several hypotheses proposed for how venous hypertension then results in venous ulceration or the development of lipodermatosclerosis. Some of these hypotheses have emphasized the accumulation of fibrin around dermal blood vessels and defects in fibrinolytic activity. Another hypothesis has centered on damage to the microvascular endothelial cells, which then results in leakage of fibrinogen and dermal pericapillary fibrin cuffs. The leakage of macromolecules from the blood vessels into the dermis in venous disease is substantial, prompting other investigators to propose that such molecules can trap growth factors and other important components and render them unavailable to the healing process or the maintenance of tissue integrity.111–115 There is evidence that growth factors in chronic wounds are bound to and trapped by macromolecules leaking into the dermis, such as albumin, fibrinogen, α-2-macroglobulin. The latter molecule is a scavenger for growth factors, including PDGF.114,115

We have earlier discussed at some length the critical importance of MMPs in cutaneous wound repair, particularly in the context of keratinocyte migration. However, in chronic wounds, MMPs may contribute to the overall failure to heal. For example, in venous ulcers and other types of chronic wounds, the wound fluid contains excessive amounts of metalloproteinases, which break down the extracellular matrix and, most likely, cytokines and growth factors.116,117

It should also be recognized that tissues around chronic wounds are not normal, and indeed they are altered by the primary pathogenic mechanisms or in the face of inability to heal readily. Clinically, the best example of this is the intense fibrosis surrounding venous ulcers, which is termed lipodermatosclerosis.118 Once lipodermatosclerosis becomes established, it can lead to ulcers, and becomes the site of ulcer recurrence. Indeed, venous ulcers surrounded by lipodermatosclerosis are much more difficult to heal. The reason for these observations may lie in an increasing body of evidence suggesting that the cellular makeup of wounds that do not heal is altered. Thus far, most likely because of the ease with which they are grown and studied in culture, the best evidence is with wound fibroblasts. We now know that ulcer fibroblasts are senescent,119 and unresponsive to certain selected cytokines and growth factors.120–122 For example, it has been shown that venous ulcer fibroblasts are unresponsive to the action of transforming growth factor-β1 (TGF-β1)120,121 and platelet-derived growth factor (PDGF)122 The unresponsiveness of venous ulcer fibroblasts to TGF-β1 may be due to decreased expression of type II TGF-β receptors. This receptor abnormality also leads to decreased phosphorylation of key TGF-β signaling proteins, including Smad2, Smad3, and MAPK.121,123 The synthetic program of cells in diabetic ulcers may also be altered, so that such chronic wounds are said to be “stuck” in a certain phase of the repair process.124 A close relationship has been reported between some of these abnormalities and the inability to heal.119

Basic Standards of Care

Arterial ulcers need vascular reconstruction; anything less is a temporary measure and ultimately leads to amputation. The latter may be the only option in inoperable patients. Arterial ulcers that are necrotic and stable with an overlying eschar should probably not be disturbed unless a vascular reconstructive plan is in place.3 Leg elevation and limb compression are the fundamental therapy for venous ulcers. Contact dermatitis is extremely common in patients with venous disease and, therefore, avoidance of topical agents is required in many patients.125 Pentoxifylline has been tested in several large randomized trials for its ability to accelerate the healing of venous ulcers. The results have varied, and it may be that a high dose of pentoxifylline, at 800 mg three times a day, is more effective than the usual dose of 400 mg three times a day.126 Whether the use of pentoxifylline should be considered standard therapy for venous ulcers is unclear at the moment. The anabolic steroid stanozolol has been effective in diminishing the induration of lipodermatosclerosis, which is the fibrotic component of venous ulcers, and in the acute and painful phase of lipodermatosclerosis, when compression bandages and stockings are too painful to use.127 However, this agent is no longer available; danazol may be a useful substitute (V. Falanga, unpublished). For now, the only known medical approach for decreasing recurrence of venous ulcers seems to be graded elastic stockings, with a pressure at the ankle in the range of 40 mm Hg. Stockings should be considered a lifelong therapy to help prevent ulcer recurrence and the other manifestations of venous disease.

The triad of neuropathy, ischemia, and trauma is the main pathophysiological underpinning for the development and recurrence of foot ulcerations in patients with diabetes.3 Measures addressing these components will accelerate healing. Glucose control and optimal management of other systemic complications of diabetes, including renal failure, are an important aspect of therapy for diabetic foot ulcers. From a local wound care standpoint, off-loading is critical and is the main standard of care for neuropathic ulcers. Patients with diabetic ulcers due to vascular insufficiency or with ulcers that are mixed in etiology, that is, having both a neuropathy and surgical compromise, need to be referred immediately for vascular studies and surgical consultation.3

Aggressive surgical debridement of the necrotic tissue and the callus surrounding the ulceration has recently become part of standard care in the management of diabetic ulcers.128 The callus probably provides additional pressure and perpetuates the ulceration. As with neuropathic diabetic ulcers, off-loading is the standard of care for pressure (decubitus) ulcers. Turning and shifting the position of the patient every 2 hours is said to be effective but actually is not proven to be so.

Surgical Approaches to Chronic Wounds

In the management of venous ulcers, surgical therapy may be considered in the absence of deep vein obstruction and when significant superficial or perforator incompetence is documented. Recently, considerable promise has come from less invasive ligation of perforators using subfascial endoscopic perforator surgery (SEPS). This approach, sometimes combined with or followed by more definitive procedures, such as the removal of the long saphenous vein (Linton procedure), may be effective in healing the ulcers.129 However, it is yet unclear whether this type of therapy prevents ulcer recurrence.

The realization a few years ago that the concept of an occlusive microangiopathy in diabetic ulcers is incorrect brought about a fundamental change in how revascularization of the diabetic foot is viewed; it has now become common practice to revascularize the diabetic limb.109,130 Therefore, standard vascular surgical techniques are applicable to diabetic patients as long as there are adequate distal vessels. Vascular reconstruction may need to be combined with limited digital, ray, or forefoot amputation. Ostectomy procedures can alleviate pressure from bony prominences in selected cases.

Ostectomy can also be useful in the management of pressure ulcers, to broaden bony prominences. In the appropriately selected patient, random or musculocutaneous flaps are indicated.

Growth factors are polypeptides with a variety of potent effects on cell proliferation and synthetic capacity (Table 248-1). To date, the only growth factor that has proven clinically effective is topically applied platelet-derived growth factor (PDGF). The PDGF-BB isomer, becaplermin, is approved by the FDA for the treatment of diabetic neuropathic foot ulcers. Several clinical trials have tested the effectiveness of becaplermin gel (30 μg/g) in diabetic foot ulcers.128,131 The results showed an incidence of wound closure of up to 48% compared to 33% for the control group. Becaplermin must only be used when combined with optimal wound bed preparation and elimination of other factors that impair healing.

In the last decade, a number of tissue engineering products have become available for clinical use.132,133 Two main types of bioengineered skin have been tested and proven to be effective in venous and diabetic neuropathic foot ulcers. One construct, approved for use in diabetic neuropathic ulcers, is comprised of neonatal foreskin fibroblasts in an absorbable suture material.134 The other type of construct, approved for use in both venous135,136 and diabetic neuropathic foot ulcers,137 is bilayered and consists of both fibroblasts and keratinocytes from neonatal foreskin. The exact mechanism of action of bioengineered skin is unknown, but there is no prolonged engraftment of cells.138 In clinical trials, both constructs were applied repeatedly to the wound, in order to stimulate healing. However, in clinical practice, it appears that a more limited number of applications is required. As with the use of becaplermin, optimal preparation of the wound bed is required for these constructs to be effective. Moreover, there is evidence that bioengineered skin is only helpful in ulcers that are of long duration and have not responded to conventional therapy.136 However, the field of tissue engineering is extensive, making use of viable or nonviable cells and matrix constructs, which can be dermal, epidermal, and dermal–epidermal (bilayered).

Even more recently, stem cell therapy of difficult to heal wounds is becoming a reality and is considered a great promise in the field of regenerative medicine, including cutaneous wound healing.139 We have discussed the possible role of stem cells generated from hair follicles.51,52 However, in addition to the still controversial use of human embryonic stem cells, the hope is that the recently achieved reprograming of adult differentiated cells (i.e., skin fibroblasts) into induced pluripotential stem cells (iPS) would lead to the possibility of either accelerating the repair process or true regeneration. The iPS have been generated from adult differentiated cells for the most part through the use of key transcription factors “cocktails”. For the most part, though the field is rapidly evolving, two different combinations of transcription factors were originally used: (1) (NANOG, OCT3, SOX2, LIN28)140 and (2) (OCT3/4, SOX2, KLF4, AND c-MYC). Many if not all of the characteristics of embryonic stem cells are present with iPS cells, but we keep learning more about the consistency or lack thereof of iPS and how to control differentiation.141 An exciting new development is the use of very small embryonic-like stem cells (VSELs), which are present in adults and are pluripotential.

While rapid developments are occurring with the use of embryonic stem cells and iPS cells, extensive progress is also being made with the use of adult multipotent stem cells, that is, from bone marrow. We can only address representative studies here. Most studies of adult stem cell therapy of human wounds have used cultured bone marrow–derived mesenchymal stem cells (MSCs).142 In pilot studies, clearly defined MSCs have been used successfully in the treatment of nonhealing chronic wounds and in acute wounds resulting from Mohs micrographic surgery defects not amenable to reconstruction.143 In that study, the MSCs were applied to the wounds using a novel fibrin spray delivery method. The improvement of wounds with adult stem cells may be due to either integration of the stem cells or by their paracrine effects.144

Predicting Wound Closure

Several recent studies have now placed us in the position to predict from simple observation and in the first 3–4 weeks of therapy whether a wound will heal in a timely fashion. The methods used to predict wound closure range from simple measurements of wound size (width and length) and change in wound area, to computerized planimetric analysis and assessment of migration of the wound edge.145 In a study of 56,488 wounds, it was shown that a percent change in area of approximately 30% at 4 weeks could predict wound closure with a sensitivity of 0.67, a specificity of 0.69, and had a positive and negative predictive value of 0.80 and 0.52, respectively.146 In even more practical terms, the appearance of the wound edge is important, so that steep edges imply no progress of the wound, while in healing wounds the edges become less steep and begin to migrate toward the center. The ability to predict closure is very important. By four weeks, the clinician should be able to make a determination of whether the present therapy should be continued or whether a change is required, including a complete reassessment of the clinical situation. The prognostic value of the healing rate by 4 weeks of therapy has been confirmed by additional investigators.147

Characterizing the events occurring after tissue injury and their overlaps represents a rather daunting task. Yet, over the last several decades, the orderly progression from injury to inflammation and coagulation, to the development of a provisional matrix, to the formation of granulation tissue and to tissue remodeling, has been elucidated to a high degree. Although somewhat artificial, the different phases of wound repair described here are the platform on which more knowledge can be built. Much still needs to be learned, but the framework is there for understanding tissue repair and for developing ways to accelerate it. Indeed, due to breakthrough in science and technology, progress has been exponential in the last few years. Today, we are in the position of using smart dressings and stem cells. With the use of cultured adult stem cells or even iPS cells or VSELs that are pluripotent, we can aim for skin regeneration and not simply for tissue repair. There are still challenges, in terms of improving traditional dressings, smart biological dressings, bioengineered skin, stem cell therapy; we will need continued and increased understanding of the science involved. It is also possible that lessons learned from failure to heal, as in chronic wounds, will provide valuable lessons for the general principles of surgical and acute wound healing.