Epidermal Homeostasis: The Role of the Growth Hormone and Insulin-Like Growth Factor Systems

Abstract

GH and IGF-I and -II were first identified by their endocrine activity. Specifically, IGF-I was found to mediate the linear growth-promoting actions of GH. It is now evident that these two growth factor systems also exert widespread activity throughout the body and that their actions are not always interconnected. The literature highlights the importance of the GH and IGF systems in normal skin homeostasis, including dermal/epidermal cross-talk. GH activity, sometimes mediated via IGF-I, is primarily evident in the dermis, particularly affecting collagen synthesis. In contrast, IGF action is an important feature of the dermal and epidermal compartments, predominantly enhancing cell proliferation, survival, and migration. The locally expressed IGF binding proteins play significant and complex roles, primarily via modulation of IGF actions. Disturbances in GH and IGF signaling pathways are implicated in the pathophysiology of several skin perturbations, particularly those exhibiting epidermal hyperplasia (e.g., psoriasis, carcinomas). Additionally, many studies emphasize the potential use of both growth factors in the treatment of skin wounds; for example, burn patients. This overview concerns the role and mechanisms of action of the GH and IGF systems in skin and maintenance of epidermal integrity in both health and disease.

• I.

  Introduction

• II.

  The GH System

  • A.

    GH and GHR/GHBP

  • B.

    GH physiology

• III.

  IGF System

  • A.

    IGFs and IGF receptors

  • B.

    IGF physiology

  • C.

    IGFBPs

• IV.

  Skin Structure and Function

  • A.

    Layers, cell types, and appendages

  • B.

    Epidermal growth and differentiation

• V.

  GH and the IGF System in the Dermis

  • A.

    Expression and actions of components of the GH system

  • B.

    Expression and actions of components of the IGF system

• VI.

  GH and the IGF System in the Epidermis

  • A.

    Expression and action of components of the GH system

  • B.

    Expression and actions of components of the IGF system

• VII.

  Epidermal Dysplasia: Involvement of GH and the IGF System

  • A.

    Psoriasis

  • B.

    Papillomas and IGFBP-3

  • C.

    Melanocytic lesions: from benign to malignant

  • D.

    Transgenic mouse models: multistage carcinogenesis

• VIII.

  Wound Healing

  • A.

    GH: natural and therapeutic effects

  • B.

    IGFs: natural and therapeutic effects

  • C.

    IGFBPs: regulation and usefulness in wound healing

  • D.

    Gene therapy: a novel avenue for treatment of wounds

• IX.

  Summary and Discussion

I. Introduction

A ROLE FOR the GH and IGF systems in skin homeostasis is evidenced by an ever-increasing number of studies, from clinical observations to analyses in animal models and at the molecular level. Investigations have focused on how the GH and IGF systems could impact on normal skin growth and development and skin perturbations, including wound healing. Analysis of the GH and IGF systems in skin growth, development, maintenance, and repair has arisen from: 1) the clinical observation that an excess or absence of serum GH correlates with phenotypic changes in the skin (1, 2); 2) the knowledge that IGF-I can mediate some of the actions of GH, combined with the observation that components of the GH and IGF systems exhibit widespread tissue distribution and thus presumably regulate cellular function (3–6); 3) the accumulating evidence that constitutive IGF-I receptor (IGF-IR) signaling is a feature of many neoplasms (7, 8); 4) the motivation to improve the treatment of perturbed skin growth, including benign (e.g., psoriatic) and malignant (basal and squamous cell carcinomas) epidermal hyperplasias as well as conditions of wound healing (9–12); and 5) the desire to clarify the mechanisms of action of GH and IGFs at the cellular and tissue level (13).

The following review concerns the role of both the GH and IGF systems in skin. More specifically, a primary aim of this review is to summarize the advances in our understanding of how the GH and IGF systems impact on the growth, differentiation, and maintenance of two major cell types of the epidermis: keratinocytes and melanocytes. Although particular emphasis is given to human skin, relevant animal models are also discussed. The role of the GH and IGF systems in the growth and development of skin appendages, including hair, is not covered in this precis but is covered by several recent reviews (5, 14–16). The review is divided into three broad areas: 1) brief background concerning the GH and IGF systems; 2) the actions of GH and IGFs in the dermis; and 3) the role of GH and IGFs in epidermal homeostasis.

II. The GH System

The GH system comprises the ligand (GH), the GH receptor (GHR), and the GH binding protein (GHBP). GH was first recognized by its endocrine activity and the critical role it plays in postnatal linear growth in mammals (17). The “Somatomedin hypothesis” arose from the observation that growth was stimulated by pituitary-derived GH and was mediated by IGF-I (18). In particular, the hypothesis suggested that endocrine GH targeted liver GHRs, stimulated the production of IGF-I and then endocrine IGF-I targeted organs, including growth plate cartilage, and growth resulted. The Somatomedin hypothesis has been modified (19) because it is now apparent that 1) GH acts via GHRs expressed by a range of cells besides the liver (20–23); 2) GH mRNA and protein are expressed in extrapituitary sites and close to GHR expression (suggesting paracrine/autocrine action) (23–26); 3) GH does not solely rely on IGF-I to mediate its action (27–29); 4) IGF-I exhibits widespread tissue expression (including skin) and is not always reliant on GH for expression (22); and 5) not all circulating IGF-I is derived from the liver (19, 30, 31). Nonetheless, a recent study provides compelling evidence that normal bone growth and density require a threshold level of circulating IGF-I, presumably via pituitary-derived GH stimulation (32).

A. GH and GHR/GHBP

Pituitary GH secretion occurs in a pulsatile fashion and is modulated by multiple factors including GH-releasing hormone, somatostatin, ghrelin, glucocorticoids, sex hormones, age, body composition, nutritional status, and diabetes (33–37).

The human GHR gene codes for a transmembrane receptor and a truncated form (GHRtr) lacking most of the cytoplasmic domain (38–42). The GHBP, which corresponds to the extracellular portion of the GHR, regulates plasma GH abundance, and at the cellular level GHBP (and GHRtr) can act in a dominant-negative manner with GHR by blocking the binding of free GH (39, 40, 42–46).

GH signal transduction involves GHR dimerization, activation of JAK-2 (janus kinase), and JAK-2 phosphorylation of multiple intracellular GHR tyrosine residues (reviewed in Refs. 47–49). Several different signaling pathways may then be triggered, including signal transducers and activators of transcription, MAP (mitogen-activated protein), and PI-3 (phophatidyl-inositol-3′) kinase (50). The PI-3 kinase pathway requires IRS (insulin receptor substrate) and may therefore explain the ability of GH to exhibit insulin-like activity (51–53).

B. GH physiology

At the cellular level, the growth-promoting actions of GH are due to stimulation of proliferation and/or differentiation in a range of cells or tissues including skin (54–57). Direct GH-stimulated differentiation, via GHRs, has been observed in a rat preadipocyte cell line (27, 58) and growth plate chondrocytes (59, 60). The metabolic actions of GH are numerous, affect almost every tissue, and may also be mediated by local or circulating IGF-I (61). For example, GH can exhibit chronic antiinsulin-like actions both in vitro and in vivo and acute insulin-like effects, demonstrated in vitro in adipocytes and in an animal model (50, 62–64). GH excess leads to overgrowth of various organs including skin (65), whereas GH deficiency (or GHR loss) results in short stature and other phenotypic disturbances (66, 67). GH-deficient adults before and after GH treatment also reveal the importance of GH in a range of tissues or body compartments including muscle, bone, heart, and the vascular system (61).

III. IGF System

The IGF system is comprised of two ligands (IGF-I and -II), the IGF receptors (types I and II), six IGF binding proteins (IGFBPs), and a range of IGFBP proteases. Each component of the IGF system is subject to temporal and tissue-specific regulation in response to numerous factors including hormones, growth factors, developmental status, and physiological condition, including nutrition, injury, and disease (13, 68). Although the IGFs were originally identified as endocrine factors, the juxtaposition of the various proteins of the IGF system in tissues enables IGFs to act in an autocrine and/or paracrine manner (13, 69–73). The skin provides an excellent example in which exquisite regulation of components of the IGF system contribute to tissue homeostasis (refer to Sections V.B and VI.B). Numerous reviews describe in detail various aspects of the IGF system (e.g., Refs. 13 and 74–76).

A. IGFs and IGF receptors

IGFs were first identified in serum by their ability to mediate the actions of GH (77) and then by their nonsuppressible insulin-like activity (78). IGF-I and IGF-II are 60% homologous at the amino acid level and exhibit significant homology to insulin (79–83). Unlike IGF-I, IGF-II is generally not under the control of GH (84).

The biological activity of IGFs is transduced by the IGF-IR, a member of the family of transmembrane tyrosine kinase receptors (85–88). IGF-I exhibits a 2- to 15-fold greater affinity than IGF-II for the IGF-IR (74). The IGF ligands and insulin have weak affinities and actions for each other’s receptors (13, 87, 89–91). After ligand binding and IGF-IR autophosphorylation, adaptor molecules (including IRS-1 or IRS-2) utilize one of several pathways, including PI-3 kinase and/or MAPK to transmit a signal to the cell nucleus (reviewed in Refs. 76 and 92–96). Although the IGF-IR signaling pathways converge with the insulin receptor (IR) and other growth factor-initiated signaling pathways (including GHR), receptor-specific features ensure distinct responses (76, 92, 97, 98).

The IGF-IIR, which is identical to the mannose-6-phosphate receptor (M-6-PR), binds IGF-II with much greater affinity than IGF-I but does not play a direct role in IGF signaling (99). Although the major role of this receptor is to traffic lysosomal enzymes that contain M-6-P residues, it may regulate IGF-II availability (100–102).

B. IGF physiology

Transgenic mice in which specific components of the IGF system are ablated confirm the role of IGFs in somatic growth, specific tissue/organ development and indicate that each peptide exerts its growth-promoting actions at distinct developmental stages (102–108). IGF-IR null mice are severely growth retarded, die perinatally, and exhibit a perturbed epidermis (discussed in Section VI.B) (103). The growth-promoting actions of IGFs are attributed to their potent mitogenic action (74, 109–111). Prodifferentiation effects are evidenced in many cell types including those of hemopoetic origin, myoblasts, adipocytes, osteoblasts, and cells of the central nervous system (74). IGFs also exhibit strong antiapoptotic activity (112–117). Aberrant expression of components of the IGF system has been described in a range of malignant tumors and cell lines, thus implicating this system in the tumorigenic phenotype (112–118).

IGFs stimulate several cellular functions associated with a differentiated phenotype including hormone secretion, extracellular matrix (ECM) expression, chemotaxis, and cell recognition (74, 119–125). The major insulin-like effects of IGF-I include glucose uptake, glycolysis, and glycogen synthesis but not glucose oxidation (126). Acute hypoglycemic effects are reduced by IGFs being bound to IGFBPs in serum (127).

C. IGFBPs

The family of structurally-related IGFBPs (-1 to -6) bind IGFs with high (but varying) affinity; exhibit widespread serum, tissue, and extravascular fluid distribution (13, 75, 128, 129), and thus: 1) regulate ligand availability and half-life in the serum and interstitial compartment (130, 131); 2) modulate the IGF/IGF-IR interaction (132); 3) augment or inhibit IGF action (133, 134); and 4) exhibit IGF-independent activity (135). Hence, IGFBPs are multifunctional. Consequently, IGFBPs provide another level of complexity by modulating all aspects of endocrine-, paracrine-, and autocrine-specific cellular IGF activity (13, 75, 129).

IGFBP-specific actions, both IGF-dependent and -independent, are attributed to several distinct structural features (e.g., heparin binding motifs) and posttranslational modifications (e.g., glycosylation, phosphorylation) that are not necessarily mutually exclusive (136). Every IGFBP may be subject to distinct proteolytic cleavage that alters IGF affinity and/or association with cell membranes or ECM proteins and ultimately modulates targeting of IGF with cell membrane-bound receptors (137). Generally, proteolysis of an IGFBP results in reduced ligand affinity, increased stimulation of IGF-IRs, and enhancement of IGF action (129, 137). Cleavage of IGFBPs also results in specific fragments that can exhibit IGF-independent cellular activity (discussed in Section III.C).

IGFBPs may mediate or augment apoptotic activity induced by many growth factors, hormones, or cellular stressors (e.g., TNFα; Ref. 138), TGFβ (139), UV irradiation, and serum starvation (140–143). IGFBP-3 is a transcriptional target of the proapoptotic, tumor suppressor protein p53 (140). Apoptosis involving IGFBP-3, however, is not always reliant on p53 activity and vice versa (144). The proapoptotic activity of IGFBPs is primarily attributed to sequestering antiapoptotic IGFs (141).

The IGF-independent activities of IGFBPs include cellular migration, stimulation or inhibition of proliferation, and proapoptotic activity (129, 135, 145). IGFBP-1 directly interacts with α5β1 integrin and stimulates human trophoblast migration (146). IGFBP-3 can stimulate apoptosis in IGF-IR null fibroblasts (144). Proteolytically derived IGFBP-3 or -5 fragments, in the absence of IGFs, can inhibit or stimulate cellular proliferation (147–151).

Although IGFBPs display IGF-independent activity, a definitive membrane-associated receptor that may signal such action remains elusive (144, 152). IGFBP-3 can, however, interact with the type V TGFβ receptor and prevent TGFβ interaction in mink lung epithelial cells (153, 154). Recent analyses in breast cancer cells further implicate the TGFβR family in IGFBP-3 cell signaling. Specifically, IGFBP-3 required TGF-βRII to stimulate phosphorylation of TGF-βRI and the signaling proteins Smad2 and Smad3, to ultimately modulate gene transcription and elicit an inhibitory growth response (155).

A nuclear role for IGFBPs has also been postulated (156). IGFBP-3 was localized to the nucleus of several cultured cell lines, including epidermal keratinocytes (157–159). Nuclear uptake of IGFBP-3 and -5 requires a nuclear localization sequence (156, 160). Subsequent yeast-two hybrid analyses revealed that IGFBP-3 associated with the nuclear receptor retinoid-X-receptor α (RXR-α; Ref. 161) and the nucleolar localized E7 oncoprotein derived from the human type 16 papillomavirus (162, 163) (further discussed in Sections VI.BandVII.B).

IV. Skin Structure and Function

A. Layers, cell types, and appendages

Skin is comprised of three broad layers: the epidermis, the dermis, and the hypodermis (subcutis) (Fig. 1). The epidermis is a stratified and dynamic structure, comprised primarily (95%) of many layers of differentiating keratinocytes. Nonkeratinocyte cell types found within the epidermis are melanocytes, Langerhans cells, and Merkel cells. Langerhans cells have an immunological function, and Merkle cells may play a role in sensory perception (164). Melanocytes, specialized neural crest-derived cells, are located between keratinocytes in the basal layer of the epidermis and provide protection against UV irradiation-induced sunburn, photocarcinogenesis, and photoaging. Melanin is produced in cytoplasmic organelles (melanosomes) and is transferred via dendritic projections of the melanocytes to keratinocytes (165). Melanin pigment protects the nuclei of dividing cells from the DNA-damaging effects of UV radiation, results in the tanning of skin, and is also responsible for “basal” skin color (see Ref. 166 for a review of melanocytes). Anatomically, the ratio of melanocytes to keratinocytes is approximately 1:36, and this arrangement is termed the epidermal/melanin unit (167). Multiple long dendritic projections enable single melanocytes to contact approximately 36 basal and suprabasal keratinocytes and thus allow for the widespread distribution of melanin pigment (168). The juxtaposition of melanocytes and keratinocytes is critical, rather like a symbiotic relationship.

Fig. 1

Diagrammatic representation of skin structure. Skin is divided into three major layers: the outer epidermis, inner dermis, and hypodermis (subcutis). The epidermis is comprised of several layers of morphologically distinct keratinocytes and sits on a basement membrane of ECM proteins. Melanocytes are located between keratinocytes of the basal layer of the epidermis. Dermal appendages include hair follicles, sebaceous glands, and sweat glands. Fibroblasts, nerve cells, and a vascular network are located in the dermis. The hypodermis is primarily comprised of adipocytes.

A basement membrane, rich in ECM proteins (including collagen type IV, epiligrin, laminin, fibronectin, nidogen, and heparin sulfate proteoglycans), separates the epidermis and the dermis and facilitates diffusion of nutrients and growth factors between the two layers (169). The basement membrane also ensures attachment of basal keratinocytes and is a critical regulator of keratinocyte differentiation (170).

The dermis provides a supporting matrix, which includes collagen and elastin, for extensive vascular and nerve networks. Fibroblasts are the predominant cells in the dermis along with endothelial cells and mast cells. Under conditions that may involve the immune system (such as wound healing and psoriasis), macrophages, lymphocytes, and leukocytes are also found in the dermis (164, 171). The hypodermis (subcutis) underlies the dermal layer and is comprised of adipocytes (fat cells) that are essential for energy storage and metabolism while contributing to insulation and protection against injury (164). Several appendages, comprised of both dermal and epidermal components, provide various functions including thermal control and protective covering. These include hair, sweat glands (apocrine and eccrine), and sebaceous glands (164).

B. Epidermal growth and differentiation

Keratinocyte differentiation involves a continuous and complex, but exquisitely choreographed, series of biochemical and morphological transformations that lead to terminally differentiated corneocytes that are eventually shed from the skin surface (172). Epidermal differentiation is associated with a cessation of proliferation, the induction of cellular migration, and ultimately cell death, and is thus considered a form of apoptosis. The process of keratinocyte differentiation takes 2–4 wk in humans (173) and is essential for maintaining the epidermal barrier. Each stage of keratinocyte differentiation is represented by a specific layer of the epidermis: basal, suprabasal/spinous, granular, transit, and cornified (Fig. 2). Because keratinocyte differentiation is a carefully ordered process, specific gene products (e.g., structural proteins, growth factors, and enzymes) are expressed in a temporal and cell-specific manner and therefore reflect a distinct stage.

Fig. 2

Differentiation of epidermal keratinocytes. Keratinocyte differentiation begins in the basal layer where selected daughters of SC become TA cells. After a finite number of cell divisions, TA cells withdraw from the cell cycle and commit to the process of terminal differentiation. PMD cells eventually detach from the basement membrane and move to the suprabasal layer where they continue to metamorphose as they migrate upward through the layers of the epidermis (basal, spinous, granular, cornified). Many local factors modulate keratinocyte differentiation, including an increasing calcium gradient. BM, Basement membrane (refer to Section IV.B). Circular green arrow on SC indicates that keratinocyte SC undergo self-renewal through cell division. Green arrows indicate progression of keratinocytes.

The process of keratinocyte differentiation begins in the basal layer and involves cross-talk between cells of the dermis and epidermis via growth factors [including the epidermal growth factors (EGFs) (174) and TGFs (175)], the vitamin-D receptor system (176), retinoid nuclear hormone receptor system (including RXRα and RARγ) (177) and ECM proteins (including integrins) (178, 179). Calcium (Ca²⁺) is another important regulator of keratinocyte differentiation. A steep intra- and intercellular Ca²⁺ gradient exists within the epidermis (180–182). More specifically, in murine epidermis basal and spinous keratinocytes are situated in an environment in which the extracellular Ca²⁺ is significantly lower than that of serum. Furthermore, low levels of intracellular Ca²⁺ are detected in these keratinocyte subtypes. In contrast, the granular and transit layer cells exhibit high extracellular and intracellular Ca²⁺ reservoirs. The cornified exhibits low intra- and extracellular Ca²⁺ levels. In corroboration with these studies, human and murine keratinocytes grown in low Ca²⁺ medium (0.05–0.1 mm) exhibit a proliferative response, but when the Ca²⁺ is increased to above 1 mm proliferation decreases and stratification or cornification of keratinocytes is induced (183–186).

Several groups have confirmed that three keratinocyte cell subtypes exist in the basal layer: stem cells (SC), transit-amplifying (TA) cells, and postmitotic differentiating (PMD) cells; a model of keratinocyte differentiation has been used for some years (Fig. 2) (170, 187–191). Epidermal SC provide a reservoir of keratinocytes that are capable of self-renewal with a high proliferative potential and are therefore critical for the continued growth and maintenance of the epidermal integument (189, 190–193). With the induction of keratinocyte differentiation, selected progeny of SC become TA cells. After a finite number of cell divisions, TA cells withdraw from the cell cycle and commit to the process of terminal differentiation. Committed cells (PMD cells) eventually detach from the basement membrane and move to the suprabasal layer where they continue to metamorphose as they migrate upward through the layers of the epidermis. Thus, proliferative keratinocytes are found only in the basal layer.

V. GH and the IGF System in the Dermis

Clinical observations (Section V.A.2) combined with the accessibility and ease of culturing dermal fibroblasts from a range of species has led to their widespread use as model cells for understanding the expression and mechanisms of action of both the GH and IGF systems. The next section focuses on and summarizes the studies that highlight the expression, regulation, and mechanisms of action of GH and IGFs by dermal fibroblasts (including human, murine, and bovine).

A. Expression and actions of components of the GH system

1. Expression

Immunohistochemical and in situ hybridization analyses indicate that GHR/GHBP protein and mRNA expression exhibit ontogenic regulation in the skin of a number of species. Initial immunohistochemical analysis, utilizing an antibody that recognizes both GHR and GHBP protein, in neonatal and adult skin from humans, rats, and rabbits revealed GHR/GHBP-positive staining in a range of dermal structures including fibroblasts, dermal papillae of hair follicles, Schwann cells of peripheral nerve fascicles, skeletal muscle cells, adipocytes, medial smooth muscle, and endothelial cells of arteries. Subsequent studies using a GHBP-specific antibody indicated that GHBP and GHR distribution correlated in rat skin (194). Lincoln et al. (195) published a more thorough analysis of GHR/GHBP immunoreactivity in normal and neoplastic tissue and cells from humans, confirming the presence of GHR/BP in the dermal cells.

A role for GH in dermal growth during fetal development is supported by the temporal and tissue-specific expression of GHR/BP protein and mRNAs in both humans and rats. Immunohistochemical analysis of human tissue indicates that GHR/BP protein is present in dermal fibroblasts as early as 8.5 wk gestation and remains present at 15–20 wk (21, 196). In situ hybridization analysis of fetal and adult rats indicates GHR/BP gene expression in cells of the dermis as early as embryonic d 16.5 (22, 197, 198). In adult male rats, GHR and GHBP distinct mRNA species were also localized to fibroblasts, adipocytes, and skeletal muscle cells (197).

Cultured adult and fetal human dermal fibroblasts express GHR/BP mRNA and protein, providing further evidence for skin as a direct GH target (21, 57, 199, 200). GH mRNA has also been detected via RT-PCR in cultured human fibroblasts (201).

2. Actions of GH

A role for GH in skin growth and development is strongly supported by clinical states of GH excess. Acromegalic patients exhibit increased GH levels; thickened, coarse, and oily skin; skin tags; and acanthosis nigricans (1). Normalization of GH levels in acromegalics leads to a decrease in skin thickness and thus indicates that the effects of GH in the skin are reversible (202). In contrast, patients lacking GHR activity (Laron syndrome) and those with GH deficiency display thin skin and/or reduced elasticity due to a reduction in dermal thickness (2, 66, 203, 204).

Further clinical evidence for GH action on skin has been demonstrated by the advent of GH therapy. Patients with GH deficiency or low IGF-I plasma levels exhibit an increase in skin thickness (affecting dermis) and stiffness after GH treatment (205, 206). Treatment of GH-deficient men revealed that GH could augment the androgenic effects on pubic hair development in the skin (207). In further support, male GH-transgenic mice exhibit a greater increase in skin thickness and skin surface area when compared with their female counterparts (208). Castration of these mice prevents the increase in skin thickness and thus confirms a role for synergy between GH and androgens in skin growth. The observed effects of GH on skin thickness, and in some cases mechanical strength, are primarily due to an increase in collagen content in the dermis and are not due to epidermal expansion (208–210). More controversial is the nonlicensed use of GH as an antiaging treatment regime for rejuvenating skin (often advertised via the Internet). Although GH treatment may enhance dermal collagen content, and thus increase skin thickness, there is no scientific evidence to support the use of this hormone for esthetic improvement of skin.

GH can bind to cultured human fibroblasts (211), via GHRs, and elicit a proliferative response (55, 57, 200, 212). Akin to liver tissue, GH up-regulates IGF-I mRNA and IGFBP-3 mRNA and protein expression of cultured human fibroblasts (200, 213). Studies of GH action on avian fibroblasts revealed that GH alone can stimulate the production of collagen (214). In addition, IGF-I can synergize with GH and increase collagen production to levels greater than that achieved by either growth factor alone and provides further explanation for the increase in dermal thickness seen in conditions of GH excess (1).

B. Expression and actions of components of the IGF system

1. Expression of IGF system components

All components of the IGF system are produced by dermal fibroblasts. Cultured fetal and postnatal human dermal fibroblasts produce IGFs (I and II), particularly in response to numerous factors including GH (213, 215). In addition, IGF-I mRNA and protein are expressed by dermal fibroblasts in human skin sections (216). Both the IGF-IR and IGF-IIR are expressed by cultured human fetal and postnatal dermal fibroblasts (217–219). IGF-IR has been detected by immunohistochemical technology in dermal fibroblasts of infant and adult human skin (220).

In situ hybridization analyses by our group localized IGFBP-4 and -5 mRNAs to dermal fibroblasts in adult human skin (221). In contrast, cultured human fibroblasts express IGFBP-3, -4, -5, and -6 (222–224). The discrepancy between in vivo and in vitro fibroblast-derived IGFBP profiles could be due to the limitations in the sensitivity of in situ hybridization or may be a reflection of factors that are present or absent in the culture medium. Interstitial fluid, derived from artificially raised blisters, contains IGFBP-1, -2, -3, and -4 (225). IGFBP-3 is predominantly present as a proteolytically cleaved fragment of 29 kDa (225–227). The origin of IGFBPs in blister fluid is not known and may include serum or dermal and epidermal cells.

In vitro, the expression profile of fibroblast-derived IGFBPs is differentially regulated by a wide range of local or systemic factors, including IGFs (228, 229), GH (200, 213), TGFβ1 (224), estradiol (213), testosterone (230), and glucocorticoids (231).

Aging affects the biological functioning of fibroblasts both in vivo and in vitro, and this difference may be in part due to changes in IGF responsiveness. Although similar levels of IGF-IRs are expressed, only young, and not senescent, human fibroblasts proliferate in response to IGF-I stimulation in vitro (232). In addition, senescent fibroblasts fail to express IGF-I mRNA, thus indicating that potential autocrine IGF-I action is ablated (233). Late passage adult fibroblasts, those derived from Werner syndrome patients (who exhibit premature aging), and senescent fibroblasts express higher levels of IGFBP-3 when compared with fetal or younger fibroblasts (234–237). Furthermore, elevated IGFBP-3 expression by senescent fibroblasts in vitro significantly reduces IGF-I-stimulated DNA synthesis via sequestration of the ligand and thus highlights another potential mechanism for ablation of paracrine IGF-I action in vivo (237).

2. Actions of IGF

IGF actions on fibroblasts may include proliferation (238), survival (239), migration (240), and production of growth factors including TGFβ1 (241) that may act locally or exert paracrine effects in the epidermis. Fibroblast-derived ECM components are regulated by IGF-I. For example, IGF-I stimulates collagen and inhibits collagenase production (124, 125, 242). Numerous locally and systemically derived factors regulate IGF-I action on fibroblasts. For example, IGF-I can synergize with platelet-derived growth factor (74) or glucocorticoids (231) to enhance proliferation or the production of specific factors in vitro.

The IGF-II/M-6-PR may also play a role in fibroblast proliferation via TGFβ1 activation (243). Latent TGFβ1 is activated by binding to IGF-II/M-6-PR and undergoing a conformational change. Although produced by dermal fibroblasts, TGFβ1 may also be derived from the epidermis (244–246). Thus, epidermal-to-dermal interaction (247) may be regulated via fibroblast IGF-II/M-6-PR levels.

Fibroblasts have been used extensively as a model cell for elucidating the mechanisms of IGFBP action, often using proliferation or DNA synthesis as an end-point of cellular function (13). Such analyses indicate that depending upon the experimental parameters, IGFBPs can augment or inhibit IGF-I-stimulated proliferation. For example, IGFBP-3 inhibits IGF-stimulated proliferation via sequestration of the ligand from IGF-IRs (133). In contrast, pretreatment of fibroblasts with IGFBP-3 leads to cell membrane association and proteolysis of IGFBP-3, a reduction in affinity for IGF-I, and ultimately potentiation of IGF-I-stimulated proliferation via modulation of PI-3 kinase activation (133, 134, 248). IGFBP-3, via cell membrane association, can modulate the fibroblast IGF-I/IGF-IR interaction independent of its ability to sequester the ligand (149). Whether IGFBP-3 directly interacts with the IGF-IR to inhibit or displace ligand binding remains to be determined. It is also possible that IGFBP-3 may affect fibroblast IGF-IR signaling via interaction with a putative IGFBP cell membrane receptor (152, 249) (Section III.C). The ECM surrounding fibroblasts also affects IGFBP activity. When fibroblast-derived, phosphorylated IGFBP-5 associates with the ECM, it has a lowered affinity for IGF-I that leads to potentiation of IGF-I-stimulated DNA synthesis (250).

Proteases and protease inhibitors are involved in the complicated cross-talk between components of the IGF system in modulating fibroblast IGF activity (228, 251, 253). An IGF-II-dependent protease (pregnancy-associated plasma protein) that cleaves IGFBP-4, and thus enhances IGF-I-stimulated proliferation, has been identified and characterized in human fibroblast cultures (253–255). It is also possible that local IGFBP-3, via binding of IGF-II, may play a role in modulating the activity of fibroblast-derived, IGF-II-dependent IGFBP-4 protease (256). Local proteolytic fragments of IGFBPs may also alter fibroblast proliferation. Proteolytic fragments of IGFBP-3 may exert an antiproliferative effect by different mechanisms and, in specific instances, exhibit activity that is not evident with the intact IGFBP. Plasmin can cleave nonglycosylated, recombinant human IGFBP-3 into specific N-terminal fragments of 22–25 kDa and 16 kDa representing residues 1–160 and 1–95, respectively (257). The 22- to 25-kDa fragment exhibits greatly reduced affinities for IGF-I and -II but can still inhibit IGF-I-stimulated proliferation of chick embryo fibroblasts, albeit by only 50% of that seen with intact IGFBP-3. In contrast, the 16-kDa fragment, shown to lack affinity with IGFs by several techniques, was a potent inhibitor of IGF-I, and insulin stimulated cellular proliferation. Furthermore, the 16-kDa fragment was as potent as intact IGFBP-3 in preventing IGF-I stimulated proliferation. These IGFBP fragments thus provide a further level of cellular regulation.

Many of the IGF-independent actions of IGFBPs, particularly IGFBP-3, have been dissected in IGF-IR null fibroblasts. Fibroblast proliferation, in the absence of IGF activity, may be inhibited by both proteolyzed fragments of IGFBP-3 or intact IGFBP-3 (144, 149, 257, 258). In mouse embryo fibroblasts, a 16-kDa IGFBP-3 fragment, generated by plasmin digestion, can prevent signaling through FGF receptors in the absence of IGF-IR stimulation (149).

Exposure of cultured human fibroblasts to UV irradiation stimulates IGFBP-3 mRNA abundance, thus suggesting that IGFBP-3 may modulate cell survival and/or apoptosis in these cells (140). In addition, in mouse embryo fibroblasts the UV-stimulated increase of IGFBP-3 mRNA is dependent upon the tumor suppressor p53, presumably via transcriptional regulation (140).

In summary, fibroblasts clearly provide a potential reservoir of IGF ligands and IGFBPs that may exert autocrine or paracrine action in the dermis and the epidermis. The role of dermally located IGFBPs in the transport of IGFs from the dermal compartment, including those derived from serum, to the epidermis remains unknown. Nonetheless, analyses in cultured fibroblasts, in particular, indicate that IGFBPs may utilize a range of mechanisms (e.g., cell or ECM association) to target IGFs to IGF-IRs located on dermal or epidermal cells. Furthermore, the dermal fibroblasts can also add to the reservoir of IGFBPs that could also exert their IGF-independent actions in all compartments of skin.

VI. GH and the IGF System in the Epidermis

A. Expression and action of components of the GH system

1. Expression

a. Keratinocytes

Immunohistochemical analysis of neonatal and adult skin from humans, rats, and rabbits revealed GHR/BP-positive staining localized to all layers of the epidermis (basal, spinous, and granular) and epidermal layers of skin appendages including sweat glands, secretory ducts, and hair follicles (199, 259).

Like dermal cells, a role for GH in epidermal growth and differentiation during fetal development is supported by the temporal and tissue-specific expression of GHR/BP protein and mRNAs in both humans and rats. In humans, the germinal layer of the epidermis also exhibits weak immunostaining at 13–14 wk but is most intense at 19–20 wk gestation (21, 196, 260). Epithelial cells of hair follicles and sebaceous glands showed weak GHR staining at 13–14 wk and most intense staining at 19–20 wk gestation (260). GHR/BP mRNA expression is detected in epidermal cells of fetal (as early as embryonic d 16.5) and adult rats (22, 197, 198). In adult male rats, GHR- and GHBP-distinct mRNA species are coexpressed in the basal, suprabasal, and granular layers of the epidermis, sebaceous glands, and epithelial-derived cells of the follicular sheath (197).

In contrast to cultured fibroblasts and immunohistochemical analyses of skin sections, cultured adult keratinocytes do not express GHR/BP mRNA (57). Such a discrepancy may be due to differences in culturing conditions and/or detection methodology.

b. Melanocytes

The role of the GH system in melanocyte growth and function is less well understood, probably due to the relatively low abundance of melanocytes in normal epidermis (1:36, melanocytes to keratinocytes). Nonetheless, GHR mRNA is expressed by cultured human melanocytes (57). In addition, low levels of GHR/BP have been detected by immunohistochemistry in the cytoplasm and nucleus of normal melanocytes located in human skin samples (195).

2. Actions of GH

a. Keratinocytes

As previously discussed (Section V.A.2), GH excess or treatment leads to an increase in skin thickness, primarily via an increase in dermal collagen (208–210). Thus, in the majority of examples of GH excess, either pathological or via exogenous administration, an increase in skin thickness was not associated with an expansion of the epidermal compartment via stimulation of keratinocyte proliferation and/or maturation. One recent study, however, revealed that a cohort of GH-deficient patients exhibited thin epidermis and showed that GH treatment did not fully reverse the epidermal deficiency (2). The mechanisms of GH action on the epidermis were not explored, and therefore the possibility of IGF-I-mediating GH action cannot be ruled out. In a human skin xenograft model, GH treatment of the host mouse led to increased epidermal proliferation and thickness (261). Intradermal administration of anti-IGF-I antibody ablated the proliferative response and indicated that IGF-I was acting as a mediator of GH action in the epidermis. This study is the sole example in the literature that confirms that IGF-I can act as a mediator of GH action in the epidermis.

Thus, although immunohistochemical analyses of human skin sections indicate that epidermal keratinocytes express GHR/BP protein (Section VI.A.1), a clear functional role is yet to be elucidated. Limited functional analyses indicate that the addition of GH to the growth medium does not induce primary human keratinocytes to proliferate, but this is expected given an observed absence (or loss) of GHR mRNA (57). It remains possible that GHR activation in the epidermis is a requirement of cellular activity not yet explored, including specific metabolic activity, or occurs in response to stressful conditions (e.g., infection). Alternatively, the immunoreactivity may represent GHRtr and/or GHBP, and thus GH activity would not be expected. It also remains possible that GH action on the epidermis is mediated via IGF-I (261). Hence, the direct effects of GH, via GHR signaling, on the epidermis remain unclear.

b. Melanocytes

In comparison to the IGF system, the role of the GH system in melanocyte growth and function is less well understood. The majority of published data provide circumstantial evidence for a role of GH action in melanocyte biology, particularly melanocytic lesions (refer to Section VII.C.1). Our recent studies, however, provide the first evidence that GH can stimulate the proliferation of primary human melanocytes only in the presence of basic fibroblast growth factor (bFGF) or IGF-I (262). Thus, it seems that the GHR mRNA expressed by cultured human melanocytes produces functional GHRs (57).

B. Expression and actions of components of the IGF system

1. Expression

a. Keratinocytes

IGF-IR expression in the epidermis generally correlates with proliferating keratinocytes. IGF-IRs are localized to the basal layer of the epidermis of normal human skin and undifferentiated epithelial cells of skin appendages (263). However, one study provides conflicting data because it describes the localization of IGF-IR mRNA and protein to all layers of the epidermis and thus to quiescent, proliferating, and differentiating cells (216). The qualitative differences between these studies may be due to differences in IGF-IR epitope recognition by each antibody or the sensitivity of each in situ hybridization analysis. Our in vitro analyses, however, indicate that IGF-IR mRNA and protein are not reduced by the calcium-induced differentiation of the human keratinocyte cell line, HaCaT (264), and is in agreement with studies in murine-cultured keratinocytes (265). These two in vitro analyses support the observation of IGF-IR in many layers of the epidermis (216).

The cellular derivation of IGFs in skin is also controversial. Some studies indicate that cultured primary human keratinocytes do not produce IGFs (57, 266), and the necessity to include IGF-I (or supraphysiological levels of insulin) in primary keratinocyte culture medium correlates with a lack of autocrine stimulation. Furthermore, irradiated fibroblasts, via production of IGF-I and IGF-II, maintain keratinocyte growth and proliferation (215, 266, 267). Thus, in vitro dermally derived IGFs act in a paracrine manner to stimulate keratinocyte IGF-IRs. These observations, combined with the knowledge that IGFs are expressed by dermal fibroblasts (266), led to the general conclusion that IGF-IR ligands originated primarily from the dermis. This conclusion agrees with the general observation that IGFs are usually produced by cells of mesenchymal origin (22). In contrast, however, the HaCaT cell line can produce IGF-II mRNA and protein (268), and primary human keratinocytes in a cultured skin substitute produce IGF-I (269).

Moreover, there are several examples showing that IGFs are expressed by epidermal keratinocytes. Rudman et al. (216) demonstrated the expression of IGF-I mRNA and protein in the epidermis of human skin sections. In particular, IGF-I was localized to the epidermal granulosum layer and differentiating, epidermally derived cells of the hair follicle. Wound healing studies indicate that IGF-I and IGF-II mRNAs are expressed by keratinocytes during skin repair (263, 270, 271). More recently, IGF-I and IGF-II immunoreactivity was localized throughout the epidermis of normal human skin (272). IGF-II has also been localized to the epidermis of the human fetus (12 wk gestation) (273). In contrast, IGF-I mRNA is absent in the epidermis of fetal rats and thus suggests species variation (22). In addition, Rho et al. (274) suggest that IGF-I mRNA expression is low to nondetectable in murine epidermis. The discrepancy between cultured keratinocytes and those in vivo may be a reflection of keratinocyte differentiation status, the growth medium, or a restriction in rodents. It therefore seems reasonable to conclude that keratinocytes in vivo produce IGFs under specific conditions, including fetal growth and wound healing. Hence, it is possible that IGF-I, and even IGF-II, may operate via an autocrine or paracrine manner within the epidermis. Epidermal paracrine activity may also be possible because melanocytes produce IGF-I (57).

Knowledge of the expression, regulation, and potential role of IGFBPs in modulating IGF action in the epidermis has been primarily provided by studies in our laboratory. We found specific IGFBP expression profiles in human epidermis and cultured human keratinocytes (221, 275, 276). In adult human skin, only IGFBP-3 mRNA and protein are produced by selected basal keratinocytes (221, 277). In contrast, IGFBP-1, -3, -4, and -6 mRNAs were found in human fetal epidermis (278), suggesting different control of the IGF system in the fetus when compared with adult human epidermis. Akin to our in vivo analyses, IGFBP-3 is the major IGFBP produced by cultured primary human keratinocytes (under basal growth conditions) and an immortalized human keratinocyte cell line (HaCaT) (275, 276). IGFBP-2, -4, and -6 were also detected in primary human keratinocyte-conditioned medium in the presence of EGF and bovine pituitary extract (275), thus indicating that local skin factors may regulate the expression of keratinocyte-derived IGFBPs.

Our regulation studies revealed that IGFBP-3 expression (mRNA and protein) is reduced by EGF, TGFβ, and calcium-induced (i.e., elevation of CaCl₂ from 0.09 to 1.2 mm) differentiation of normal and/or a human keratinocyte cell line, HaCaT (264, 279, 280). The differentiation status of HaCaT keratinocytes has been associated with the presence of IGFBP-6 expression. In particular, Kato et al. (245) and Marinaro et al. (268) describe the presence of IGFBP-6 in HaCaT conditioned medium when cells were grown in the absence of serum and high calcium (DMEM, 1.8 mm CaCl₂), conditions that would induce differentiation. Our study of HaCaT keratinocyte differentiation failed to demonstrate the induction of secreted IGFBP-6 by an increase in the calcium level in the medium (keratinocyte growth medium, 1.2 mm; our unpublished data and Ref. 264). It remains possible that the lack of IGFBP-6 expression in that study is related to differences between culture media including calcium concentration. It should also be noted that although elevated calcium levels are frequently used to induce differentiation of monolayers, keratinocytes exhibit a more complete differentiation program in the presence of mesenchymal support (e.g., organotypic cultures or when xenografted onto athymic mice) (281). Our findings are however consistent with in vivo studies showing a lack of IGFBP-1, -2, -4, -5, and -6 in adult human epidermis, except in specialized sebaceous glands and sweat glands (221). The suppression of keratinocyte-derived IGFBP-3 expression by factors that affect the proliferation and differentiation status of keratinocytes may reflect the mechanisms that regulate IGFBP-3 distribution in the epidermis (refer to Section IX and see Fig. 5 for a thorough discussion).

Skin interstitial fluid contains proteolyzed IGFBP-3, which suggests that epidermal IGFBP-3 immunoreactivity may also include cleaved forms (226, 227, 277, 282). Thus, it remains possible that keratinocyte-derived IGFBP-3 is primarily present as the cleaved form as a result of local proteases. In support of this postulate, plasminogen-treated HaCaT keratinocytes produce a protease which is similar to that found in skin interstitial fluid and which cleaves IGFBP-3 (227).

A possible nuclear role of IGFBP-3 in keratinocyte biology is indicated by our confocal microscopic analyses that revealed IGFBP-3 immunoreactivity in the nucleus of dividing HaCaT keratinocytes (159). Although IGFBP-3 has also been localized to the nucleus of other cell types (Section III.C), the significance remains elusive. Nonetheless, it is interesting to note that IGFBP-3 can associate with RXRα in HeLa cell nuclear extracts (161). The same study showed that the antiproliferative and proapoptotic activity of IGFBP-3 also required a functional RXRα system (161). It remains possible that epidermal IGFBP-3, via an interaction with RXRα, may modulate RXRα activity in the epidermis, including keratinocyte proliferation and differentiation (177).

b. Melanocytes

Primary human melanocytes express IGF-I mRNA, and our recent studies indicate that they express IGFBP-4 in minimal amounts (262). IGF-IR must also be expressed by melanocytes as they respond to IGF-I (262, 283).

2. Actions of the IGF system

Unequivocal evidence for the critical role of the IGF system in epidermal homeostasis, at least during fetal development, is provided by IGF-IR knockout mice, which present with thin, transparent skin lacking a spinous layer and a reduction in hair follicle number and size (103). Unfortunately, postnatal mortality of these mice abrogates the opportunity to further dissect the functional role of IGF-IR signaling in normal and disturbed postnatal epidermal development. Thus, in the absence of IGF-IR gene-targeted deletion, the mechanisms of IGF action in keratinocyte biology must be inferred from conventional in vivo and in vitro investigations.

a. Keratinocyte proliferation

Similar to many cultured cells, supraphysiological levels of insulin (5 μg/ml) stimulate keratinocyte growth (284, 285). The ability to substitute insulin with lower levels of IGF-I indicated that insulin was operating via IGF-IRs. Other studies reveal that activation of the IR plays a distinct role in keratinocyte proliferation. Firstly, ablation of keratinocyte IGF-IR signaling, by inclusion of an anti-IGF-IR antibody (αIR3), confirms the mechanism by which IGF-I, IGF-II, and insulin stimulate proliferation (285). The ability of αIR3 to only partially block insulin-stimulated keratinocyte growth supported a role for IRs in epidermal homeostasis. Recent analysis of IR-null mice confirmed the role of the IR in keratinocyte proliferation and revealed that both insulin and IGF-I stimulate intracellular IRS-1 phosphorylation (286). Subsequent studies revealed, however, that combined IGF-I and insulin stimulate murine-cultured keratinocyte proliferation in an additive fashion and utilize distinct intracellular pathways (287). More specifically, insulin but not IGF-I, signaling was mediated by protein kinase C δ via activation of α2 and α3 Na+/K⁺ pump subunit isoforms. In addition, PI-3 kinase activity was not required for insulin-induced proliferation. The intracellular effectors downstream of IRS-1 that are used by the IGF-IR in keratinocyte proliferation, however, remain to be elucidated.

IGF-II can also stimulate human keratinocyte proliferation, but IGF-I is more potent (288) and may reflect the higher affinity of IGF-I for IGF-IRs (74). We have demonstrated that IGF-I also stimulates HaCaT keratinocyte proliferation (264, 276), and a reduction of IGF-IR mRNA and IGF-IR abundance via antisense oligonucleotide methodology significantly repressed cell growth (12). Finally, IGF-stimulated murine primary keratinocyte proliferation is impaired in the presence of high glucose levels (20 mmol/liter) and thus correlates with impaired wound healing of diabetic subjects (289).

IGF-I (or insulin) combined with EGF in low calcium culture medium stimulates keratinocyte growth to a much greater level than that achieved by either growth factor alone (290–292). Such synergistic activity appears to involve cross-talk between the two systems. IGF-I stimulates an autocrine loop via activation of EGF receptors (EGFRs) and up-regulation of EGFR ligands [TGFα, amphiregulin, and a heparin-binding EGF-like protein (293–295)]. The addition of an EGFR antibody abrogates IGF-I stimulated keratinocyte proliferation (296). However, because EGF alone cannot stimulate keratinocyte cell number to that seen with both growth factors, it seems clear that other factors or signaling systems are involved. For example, IGF-I also synergizes with bFGF to enhance keratinocyte proliferation in vitro (290).

Endogenous IGFBP-3 inhibits IGF-I-stimulated proliferation of the human keratinocyte cell line, HaCaT, and therefore suggests that this IGFBP may regulate IGF-stimulated keratinocyte growth in vivo (276). The colocalization of IGFBP-3 and IGF-IR (216, 263, 297) in the basal layer of the epidermis supports such a role (220, 277). The addition of exogenous IGFBP-3 to HaCaT keratinocytes can either stimulate or inhibit IGF-I-induced proliferation, but the response depends upon whether IGFBP-3 is coincubated or preincubated with the cells (227). The differing effects of IGFBP-3 on IGF-I-stimulated keratinocyte proliferation mimic those described in cultured human dermal fibroblasts (133, 134) and are yet to be clarified in vivo.

b. Keratinocyte differentiation

The incomplete or thin epidermis of the IGF-IR knockout mouse suggests that IGF-IR signaling may also be important for keratinocyte differentiation/stratification (103). In addition, IGF-I can induce stratification/differentiation of an SV40 transformed keratinocyte cell line (298). Rudman et al. (216) suggest that IGF-IR immunoreactivity in the suprabasal layers of the epidermis indicates that IGFs also regulate keratinocyte differentiation. In contrast, however, calcium-induced differentiation of cultured murine keratinocytes leads to the ablation of IGF-IR signaling but no change in the cellular distribution of IGF-IR (265). In further support, we found that calcium-induced differentiation of HaCaT keratinocytes did not alter the IGF-I-stimulated proliferative response, although IGF-IR abundance remained the same and IGFBP-3 abundance was reduced (264), thus suggesting that postreceptor signaling mechanisms are likely to play a more important role. It is also important to note that although keratinocytes express both IGF-IRs and IRs, they exhibit different mechanisms of action during calcium-induced differentiation and perhaps indicate that specific signaling pathways are involved in epidermal differentiation (265). Specifically, differentiation of keratinocytes is associated with a reduction and loss of phosphorylation of IR and IGF-IRs, respectively.

Because IGFBP-3 is expressed by specific basal keratinocytes and is reduced upon induction of differentiation, it is also possible that IGFBP-3 modulates epidermal differentiation via both IGF-dependent and/or IGF-independent mechanisms (129, 221, 264, 277). In particular, the distribution of IGFBP-3 in the basal layer indicates that it may play a role in modulating the early stages of keratinocyte differentiation, specifically the evolution from keratinocyte SC to TA cell and then PMD cells. This postulate is discussed further in Section IX and Fig. 5.

c. Keratinocyte motility

Keratinocyte motility is essential for normal differentiation and wound-healing conditions. Monolayer and coculturing models indicate that human keratinocytes migrate in response to IGF-I (119–123). Thus, the expression of IGF-I and IGF-II by migrating keratinocytes during wound healing may reflect the actions of each ligand on keratinocyte migration in vitro (270, 271).

d. Keratinocyte apoptosis

Ultraviolet B (UVB) absorption by keratinocytes is a major cause of genomic DNA damage and thus a major contributor to the development of epidermal skin cancers. Epidermal keratinocytes appear to use two mechanisms to abrogate the effects of UVB radiation: 1) DNA repair during cell cycle arrest with subsequent resumption of normal cell division, or 2) apoptosis (299). Although Hansson et al. (300) described the transient induction of IGF-I in mouse skin after UV stimulation, two recent studies support a role for the IGF system in keratinocyte cell survival and protection from apoptosis after UVB irradiation. In the first study, exogenous IGF-I, via IGF-IR activation, protects cultured human keratinocytes from UVB-induced apoptosis by promoting cell survival in preference to proliferation (301). Furthermore, IGF-IR activation, via PI-3 and MAPK signaling pathways, combined with UVB irradiation induce keratinocytes to become postmitotic. In the second study, transgenic mice that overexpress IGF-I in the basal layer of the epidermis exhibited suppression of UVB-induced apoptosis (302).

e. Keratinocytes: glucose transport

In cultured murine keratinocytes, IGF-IR activation increases glucose transport via glucose transporters 2 and 3, and this is in contrast to insulin-stimulated glucose uptake via glucose transporters 1 and 5 (303). Thus, the IGF system regulates keratinocyte glucose metabolism in a different manner to the insulin system.

f. Melanocytes

Evidence for IGF-I as a stimulator of melanocyte growth was indicated by the necessity to include supraphysiological levels of insulin (5 μg/ml) in culture medium (304). The ability to replace insulin with lower levels of IGF-I confirmed that the IGF-IR was the primary signaling process for these factors (283).

Low levels of melanocyte-derived IGFBP-4 enhance IGF action in vitro (262). The role of melanocyte-derived IGFBPs in modulating IGF action in vivo remains uncertain because adjacent basal keratinocytes produce abundant amounts of IGFBP-3 (221, 277). Specifically, the juxtaposition of melanocytes with basal keratinocytes suggests that IGFBP-3 may also significantly modulate IGF-I interaction with melanocyte IGF-IRs.

VII. Epidermal Dysplasia: Involvement of GH and the IGF System

A. Psoriasis

Psoriasis is a nonmalignant, autoimmune skin condition that is characterized by epidermal hyperproliferation and abnormal differentiation of keratinocytes overlying a dermal inflammatory reaction (171). The etiology of psoriasis is not clearly understood but appears to be multifactorial; both genetic and environmental factors seem to contribute to manifestation of the skin lesions (305). Psoriatic skin lesions appear as red, scaly patches that can cover all parts of the body.

1. Psoriasis and GH

The role of GH in psoriasis is controversial. Weber et al. (306) were the first to report elevated circulating GH levels in psoriatic patients. Subsequently, two patients with pituitary hyperplasia and increased GH levels exhibited psoriatic lesions (307), and GH treatment of a psoriatic patient correlated with lesional relapses (308). In contrast, elevated serum levels of neither GH nor IGF-I could be substantiated in other cohorts of psoriatic patients (309–312), and the literature does not report an increased incidence of psoriasis in acromegalic patients. Moreover, the use of somatostatin in the treatment of psoriatic patients did not lead to the regression of psoriatic lesions and therefore provides more conflicting data (313, 314). Bjorntorp et al. (312) concluded that the GH/IGF-I axis was not involved in psoriasis because they found that urinary levels of GH, IGF-I, and IGFBP-3 in psoriatic patients were not different from those of nonpsoriatic patients. Furthermore, GHR abundance in psoriatic lesions was not different from normal skin, nor did GHBP levels in serum differ. Although lymphocytes (315) and fibroblasts (201) potentially express GH, they found no evidence of GH mRNA in the psoriatic lesion. Thus, on balance, it seems that the majority of evidence does not support a role for GH in psoriasis.

2. Psoriasis and IGFs

Several lines of evidence implicate local perturbations at all levels of the IGF system in the hyperproliferative psoriatic epidermis. Blister fluid from psoriatic lesions contains higher levels of IGF-II, but not IGF-I (226). Keratinocytes from psoriatic patients, even from uninvolved regions, demonstrate a greater proliferative response to IGF-I when compared with keratinocytes from an unaffected individual (316). IGF-I and EGF exhibit greater proliferative synergy in psoriatic keratinocytes (317). In psoriasis, IGF-IR expression in the epidermis is more widespread (basal and suprabasal layers) and correlates with the expanded proliferative compartment (263). Furthermore, IGF-IRs of psoriatic lesions exhibit up-regulated tyrosine kinase activity (297). Our recent studies, however, have provided conclusive evidence for the importance of IGF-IR signaling in the etiology of the psoriatic lesion. More specifically, down-regulation of the IGF-IR mRNA, via antisense oligonucleotide technology, led to normalization of the hyperplastic epidermis of psoriatic skin grafted onto athymic mice (Fig. 3) (12).

Fig. 3

IGF-IR antisense oligonucleotide treatment normalizes the hyperplastic epidermis of human psoriatic skin xenografts. Psoriatic skin biopsies were grafted onto athymic and subjected to intradermal injections of IGF-IR-specific antisense oligonucleotides (AON), a random-sequence ODN in PBS or with PBS alone, every 2 d for 20 d. Grafts were then analyzed histologically. A, Donor A graft treated with AON no. 50 showing epidermal thinning compared with pregraft and control (PBS)-treated graft. Donor B graft treated with AON no. 27 showing epidermal thinning compared with pregraft and control (R451)-treated graft. E, Epidermis. Scale bar, 400 μm. B, Mean epidermal cross-sectional area over the full width of grafts was determined by digital image analysis. Results are represented as mean ± sem. Shaded bars, Control treatments (R451, random ODN sequence); black bars, treatments with IGF-IR AON. Asterisks indicate a significant difference from the vehicle-treated graft (P < 0.01); ++ indicates a significant difference from R451-treated grafts (P < 0.01). C, In situ hybridization analysis of IGF-IR mRNA in psoriatic skin before and after grafted and ODN. IGF-IR mRNA is seen as black grains that are almost absent from the lesion treated with IGF-IR AON (no. 27). Arrowheads indicate the basal layer of the epidermis. Scale bar, 50 μm. [Adapted with permission from Wraight et al.: Nat Biotechnol 18:521–526, 2000 (12). © Nature Publishing Group.]

The potential role of IGFBPs in psoriatic skin lesions has also been investigated. Xu et al. (226) reported that blister fluid from psoriatic lesions exhibited an increase in intact IGFBP-3 abundance attributable to a reduction in IGFBP-3 proteolytic activity due to the presence of a specific protease inhibitor. The derivation of IGFBP-3 in interstitial fluid has not been determined but could be serum, dermal, and/or epidermal. Our analyses reveal disturbances in the distribution and abundance of IGFBP-3 mRNA and protein in the psoriatic epidermis (Fig. 4) (277). Although strictly localized to basal keratinocytes, as in normal epidermis, high levels of IGFBP-3 mRNA and protein are tightly associated with relatively quiescent (Ki-67-negative) cells. It should be noted, however, that our analyses of psoriatic and normal skin used an anti-IGFBP-3 antibody that does not distinguish between intact and proteolyzed forms. Therefore, the identity of IGFBP-3 immunoreactivity in the epidermis remains to be determined.

Fig. 4

IGFBP-3 mRNA and protein are localized to the basal epidermal layer in normal and psoriatic human skin. IGFBP-3 mRNA, detected by in situ hybridization, is seen as black grains over selected keratinocytes of the basal layer of the epidermis in normal (panel A; scale bar, 100 μm) and psoriatic (panel C; scale bar, 50 μm) skin. IGFBP-3 protein, detected by immunohistochemistry, is seen as brown staining over selected basal keratinocytes in normal (panel B; scale bar, 50 μm) and psoriatic (panels D and E; scale bars, 250 and 50 μm, respectively) skin. As indicated, IGFBP-3 exhibits disturbed and elevated expression in the suprapapillary epidermis of the psoriatic lesion when compared with normal skin. [Adapted from Wraight et al.: J Invest Dermatol 108:452–456, 1997 (277). © Blackwell Publishing.]

The above-mentioned studies provide potential insights into the mechanisms of IGF action in the epidermis. In psoriatic skin, an increase of intact IGFBP-3 abundance could lead to the observed elevation in the local IGF pool, via protection of the ligand from degradation (129). Thus, an increase of IGF ligand in the extracellular milieu may lead to excess IGF-IR stimulation and ultimately epidermal hyperproliferation. If IGFBP-3 in normal skin is usually present in the proteolyzed form, then, taking into consideration the above-mentioned psoriatic skin data, IGFBP-3 fragments may exert antiproliferative effects on keratinocytes via mechanisms similar to that described in cultured chick fibroblasts, i.e., via inhibition of IGF-stimulated proliferation (149) (Section V.B.2). Furthermore, it is possible that IGFBP-3 fragments could inhibit excess keratinocyte proliferation via IGF-independent mechanisms (148–150). Hence, the observed absence of IGFBP-3 in selected psoriatic basal keratinocytes correlates with their increased proliferative (Ki-67-positive) activity (277). Alternatively, if epidermal IGFBP-3 is usually intact, it may modulate basal keratinocyte proliferation by simply sequestering IGFs and preventing IGF/IGF-IR interaction (131, 132) or via IGF-independent mechanisms observed in other cells types (144, 149, 257, 258) (Section III.C). These data are further discussed in a model describing the putative actions of the IGF system in modulating epidermal homeostasis (Section IX and Fig. 5).

Fig. 5

Model of IGF action in the epidermis. A, Normal skin: IGF-I, produced by dermal fibroblasts (elongated orange/brown cells) and epidermal melanocytes (dendritic-like, red cell situated between keratinocytes), acts in a paracrine fashion on basal keratinocytes. As indicated, IGFBP-3, produced by SC and TA, may modulate keratinocyte proliferation in the basal layer via a range of mechanisms. Specifically IGFBP-3 may sequester IGF-I from basal keratinocytes [1], target IGF-I to basal cells without [2] or with [3] cell surface association, and potentially via IGF-I independent mechanisms when intact [4] or proteolyzed [5] (further described in the text, Section IX). Finally, IGFBP-3 expression by PMD and upper epidermal keratinocytes may be suppressed by local expression of TGFβ, TGFα, or increasing calcium concentration (right side of panel A). Green arrows indicate progression of basal keratinocytes. Circular green arrow on SC indicates that keratinocytes SC undergo self-renewal through cell division. Scissors, IGFBP protease. B, Psoriatic skin: IGF-I, produced by dermal fibroblasts (elongated orange/brown cells), inflammatory cells (rounder red/brown cells), and melanocytes (not drawn), acts in a paracrine fashion on basal and suprabasal keratinocytes. In this model, IGFBP-3 is only produced by keratinocyte SC and is absent from most TA and all PMD cells. The absence of IGFBP-3 produced by most TA cells results in their excess proliferation in response to IGF-I in synergy with TGFα (EGFR ligand). In addition, excess proliferation of TA cells may also be as a result of a loss of IGFBP-3 (intact or proteolyzed fragments) acting on these cells in an IGF-I-independent manner (as described in Fig. 5A and in the text, Section IX). Thus, in the regions above the TA cells that have lost IGFBP-3 expression, the psoriatic epidermis exhibits hyperplasia (proliferating cells in the suprabasal layer) and abnormal differentiation (including the presence of nucleated cells in the outer layers). Relatively normal epidermal differentiation is seen in the right column of cells (refer to Section IX for a more thorough description).

B. Papillomas and IGFBP-3

Human papillomaviruses (HPV) are responsible for benign skin lesions that exhibit epidermal hyperproliferation in humans (318). Furthermore, the HPV gene E7 is partially responsible for the immortalization of epidermal keratinocytes (319). Observation of E7/IGFBP-3 interaction led to the discovery that keratinocyte-derived IGFBP-3 is degraded by E7-expressing keratinocytes (162). Thus, a loss of intact IGFBP-3, via E7-induced degradation, may be an important feature of the immortalization and may ultimately involve the increased exposure of keratinocytes to the potent mitogen IGF-I.

C. Melanocytic lesions: from benign to malignant

The transition of melanocytes to melanomas is not clearly understood and does not always follow a distinct program of phenotypic changes (320). Abnormal growth of melanocytes, however, includes benign melanocytic lesions and primary melanomas that are contained within the epidermis, primary melanomas that have invaded into the dermis, and metastatic melanomas that are distributed in the dermis and elsewhere in the body. Both the GH and IGF systems have been implicated in the etiology of melanocytic lesions.

1. GH action

Bourguignon and colleagues (321, 322) have published several studies concluding that GH treatment of numerous patients (children), including those with Turner’s syndrome and hypopituitarism, increased the number and growth rate of melanocytic naevi, together with phenotypic changes. Several studies by another research group (323, 324) refute the findings of Bourguignon and colleagues. In particular, GH treatment of children, including those with Turner’s syndrome, did not result in an increase in the number or size of melanocytic naevi. In support of these findings, it is argued that acromegalic patients do not exhibit an increased frequency of melanocytic lesions. A recent report, however, described the presence of benign and malignant melanomas in the eyes of two acromegalic patients (325).

Evidence for indicating that GH may act directly in melanocytic lesions including melanomas has been provided by the immunohistochemical localization of GHR/GHBP (195). Low levels of GHR/GHBP were detected in the cytoplasm and nucleus of benign melanocytic lesions such as compound naevi. Moreover, intense GHR/GHBP levels were localized to malignant melanomas, with greater abundance in the cytoplasm than in the nuclei. Three melanoma cell lines exhibited scant levels of GHR/GHBP immunoreactivity in the nucleus with high levels in the cytoplasm and golgi.

Thus, all of the evidence to link the GH system with the growth and development of melanocytic skin lesions remains circumstantial, and therefore the functional role of GH remains to be clarified.

2. IGF action

The recognition that the IGF system could play a role in the pathophysiology of tumor cell growth led to the characterization of this system in melanocytic lesions, especially melanomas. In situ hybridization and immunohistochemical analyses provide circumstantial evidence that changes in the expression of components of the IGF system may be involved in the evolution and/or maintenance of melanomas. Fleming et al. (326), using melanocytic lesions grouped according to the Clark scale, showed that IGF-I mRNA levels exhibited progression-associated expression. In particular, the highest levels of IGF-I mRNA were seen in dysplastic naevi and primary melanomas. However, advanced and metastatic melanomas had the lowest levels of IGF-I mRNA and were comparable to common naevi (the most basic of melanocytic lesions).

An analysis of IGF-IR protein levels in congenital pigmented naevi indicated widespread IGF-IR expression that was not necessarily elevated in proliferating cells (Ki-67-positive) (327). However, recently a study showed increasing IGF-IR protein levels associated with progression from benign naevi to metastatic melanomas and correlation with an increased rate of cellular proliferation (118).

Numerous studies on cultured melanocytic cells, from primary cells through to melanoma cell lines, further underscore the importance of the IGF system in their growth. Initial studies indicated that exogenous IGF-I or insulin was the primary growth stimulator of several metastatic melanoma cell lines, and inclusion of antibodies to the IGF-IR blocked this action (328). Consistent with these analyses, although IGF-I mRNA has been detected in cultured melanocytes (57), primary melanocytes and naevus cells require IGF-I (or insulin) and other growth factors for maximum growth in vitro (304). Two studies showed that primary melanocyte cell cultures grow efficiently in the presence of IGF-I (or insulin), bFGF, 12-O-tetradecanoylphorbol-13-acetate (TPA), and α-melanocyte stimulating hormone, whereas naevus cells were less dependent on bFGF and TPA (329, 330). In contrast, IGF-I or insulin were the only exogenous growth factors required by primary melanomas, and metastatic melanomas grew in the absence of any exogenous growth factors. This suggests that an enhanced autocrine growth factor system enabled metastatic melanoma growth.

Blockade of IGF-IR signaling using antisense cDNA plasmid constructs or oligonucleotides, antibodies, or inhibitors of glycosylation results in growth inhibition of many metastatic melanoma cell lines and in xenotransplants in athymic mice (11, 331–334). IGF-IR abundance, and thus activity, may also reflect the status of the tumor suppressor p53 in some melanoma cells because p53 modulates IGF-IR gene transcription (335). In particular, ablation of mutant p53, via antisense oligonucleotide treatment, led to a drastic down-regulation of IGF-IR abundance, IGF-IR tyrosine phosphorylation, DNA and cell replication, and an increase in apoptosis in three melanoma cell lines (336). This study indicates that mutant p53, although relatively rare in melanomas (337), may stimulate IGF-IR gene transcription, and this is in contrast to the repression of IGF-IR gene transcription by wild-type p53 (335). Furthermore, this study confirms the role of the IGF-IR in melanoma growth and survival.

A study by Resnicoff et al. (333) suggests that in one human melanoma cell line (FO-1) IGF-I signaling may be more important for the transformed phenotype rather than mitogenesis. In particular, FO-1 cells stably transfected with antisense IGF-IR cDNA plasmid constructs or treated with antisense IGF-IR oligonucleotides were only growth inhibited in xenotransplants and soft agar assays and not when grown in monolayer cultures.

The IGF system may also play an important role in the metastatic nature of advanced melanomas. Two early reports attest to the ability of IGF-I and -II to act as chemotactic agents for a melanoma cell line (A2058). Inhibition of motility was achieved by blockade of the IGF-IR with an antibody (338, 339).

It should be noted that the IGF system might not be the only mechanism of growth by metastatic melanomas. Specifically, Furlanetto et al. (11) described the lack of growth inhibition by anti-IGF-IR antibody in two metastatic cell lines both in vitro and in athymic mice. The lack of inhibition is perhaps reflective of the genotypic and phenotypic differences between each melanoma. This observation also correlates with two studies in which constitutive IGF-I or -II mRNA was not detected in a panel of metastatic melanomas (340, 341).

Two studies suggest that alterations in IGFBP expression profile may contribute to the etiology of melanomas. Stracke et al. (339) detected a cell surface-associated IGFBP of 38 kDa (potentially IGFBP-3) in the melanoma cell line A2958. Olney et al. (341) described the production and regulation of IGFBP mRNAs in another melanoma cell line (HT-144). mRNAs for IGFBP-2, -3, and -5 were present, but only IGFBP-2 and -3 were increased with IGF-I stimulation. These IGFBP profiles differ from cultured human melanocytes that display only low levels of IGFBP-4 (262).

D. Transgenic mouse models: multistage carcinogenesis

Several lines of evidence indicate that keratinocyte dysplasias (benign and malignant) involve perturbations of the IGF system.

Three transgenic mouse models in which IGF expression was specifically targeted to keratinocytes confirm a role for the IGF system in epidermal hyperplasia (302, 342, 343). Targeted overexpression of IGF-I or IGF-II to the suprabasal (via a human keratin 1 or bovine keratin 10 promoter, respectively) or basal (via a human keratin 5 promoter) layer induced a hyperplastic and hyperkeratotic epidermis. Each transgenic animal model exhibited dermal and hypodermal thickening, thus indicating epithelial to mesenchymal cross-talk. In contrast to these studies, several transgenic mouse models overexpressing IGF-I (under the control of ubiquitous promoters) have been described, a significant change in skin phenotype was not reported, but the possibility of a lack of elevated IGF-I in the epidermis was not clarified (105). Another study by Eming et al. (344) described the production of primary human keratinocytes that overexpressed IGF-I. When grafted onto athymic mice, these transgenic cells produced stratified epidermis that was morphologically comparable to control grafts. Although markers of proliferation or regenerative epithelium (Ki-67 and K16 immunostaining) were enhanced in the epidermis formed from modified keratinocytes, the authors noted that terminal differentiation was unaffected. More specifically, Eming et al. (344) concluded that overexpression of IGF-I could not induce the gross phenotypic changes seen in the hyperplastic epidermis of psoriatic skin lesions. It is, perhaps, not surprising that altered expression of one ligand can induce the severe morphological changes seen in such a complicated skin lesion. Nonetheless, it is also possible that IGF-IR abundance, and thus signaling, may be a rate-limiting step in controlling the hyperplastic epidermis in psoriatic skin. This possibility is supported by the observation that the proliferative (and transformation) capacity of cells is modulated by IGF-IR abundance (number/cell) (345). In further support of this postulate (Section VII.A), psoriatic epidermis exhibits elevated levels of IGF-IR (263), and down-regulation of IGF-IR mRNA can result in regression of the grossly altered and expanded epidermal compartment (12). Alternatively, local IGFBPs may have modulated the activity of transgene-derived IGF-I (221).

Multistage carcinogenesis mouse models have been used to dissect the correlation between IGF-IR signaling and epidermal tumor development. In mice, topical application of TPA did not alter IGF-I and IGF-IR mRNA levels. In contrast, elevated levels of IGF-I mRNA were found in epidermal tumors (squamous cell carcinomas), and an up-regulation of IGF-IR mRNA occurred in some papillomas and squamous cell carcinomas. These initial studies therefore suggested that alteration in the IGF system may be coincident but not necessarily causative of epidermal tumors (274).

Transgenic mouse models provide further confirmation for the essential role of IGF-IR signaling in epidermal tumor promotion. Mice overexpressing IGF-I in the basal (HK5/IGF-I) or suprabasal (HK1/IGF-I) layer exhibit enhanced tumor formation after an initiation-promotion protocol (7,12-dimethylbenz(a)anthracene, then TPA, chrysarobin, okadaic acid, or benzoyl peroxide treatment) (302, 343, 346). Moreover, treatment with TPA only led to an enhanced incidence of papillomas and tumors in these mouse models. Thus, IGF-I overexpression may substitute for an initiating event in tumor formation. Overexpression of IGF-I in the basal layer of the epidermis (HK5/IGF-I) was also associated with increased spontaneous tumor formation (302).

VIII. Wound Healing

The healing of skin is a multifaceted and complex process involving epidermal and dermal cellular cross-talk requiring cytokines, growth factors, and ECM modulation (347, 348). The process of wound healing is, for clarity, often considered to involve three major stages: 1) the inflammatory reaction, 2) granulation tissue formation, and 3) scar tissue remodeling. During inflammation, blood clots form, and inflammatory cells move into the wound area. The process of reepithelialization also begins during this phase as keratinocytes begin to migrate across the wound surface. The second stage involves cellular proliferation and tissue formation to restore the functional skin barrier. Keratinocytes within the wound and those at the edge begin to proliferate and thus contribute to the migrating population and ultimately reepithelialization (348). Dermal matrix restoration (fibroplasia) and angiogenesis also occur during this phase. Finally, scar tissue matrix remodeling is accompanied by vascular regression and a reduction in dermal cell density. Wound healing has been assessed in a broad range of conditions, from simple cuts through large-scale burns and recalcitrant wounds of diabetic patients. Clinical studies along with animal models, including human skin grafts, cultured cells, and organotypic-/coculture systems, have been used to dissect the finely tuned process of wound healing. The GH and IGF systems, due to their anabolic actions, have been analyzed for their role and potential therapeutic value in wound healing. Such studies provide further evidence that the GH and IGF systems are an essential feature of epidermal/dermal cellular cross-talk in normal and perturbed skin states.

A. GH: natural and therapeutic effects

In the majority of studies, systemic therapy has been used to investigate the effects of GH on skin wound healing. Collectively, these studies indicate that GH treatment can significantly improve wound-healing rates. Elevation of serum GH in burn patients can lead to enhanced wound healing as evidenced by improved reepithelialization, increased granulation tissue, increased coverage by basal lamina, and a decrease in healing time (9, 349). GH treatment also accelerates the healing of skin graft donor sites in severely burned children and adults (9, 350). Animal models reveal that GH therapy also improves healing time and increases collagen content, granulation tissue, and thus the tensile strength of skin (209, 351–354). Interestingly, systemic administration of GH does not enhance tensile strength of wounds in old rats (355) and could be due to alterations in expression of components of the IGF system and/or responsive to IGFs by aged fibroblasts (232–237).

GH treatment corrects the reduction in serum IGF-I levels induced by serious thermal injury and stimulates hepatic IGF-I mRNA transcription and IGF-IRs at the wound site (9, 356–359). Combined systemic GH/IGF-I treatment synergistically improves reepithelialization of the wound when compared with either hormone alone in a skin burn rat model (360). In another wounded rat model, systemic GH treatment combined with locally applied IGF-I significantly improved reepithelialization rates when compared with either hormone alone or no treatment (361).

B. IGFs: natural and therapeutic effects

The IGF system appears to be essential for wound healing, even in the absence of GH treatment. Local IGF-I and IGF-II mRNAs and protein abundance are modulated during the wound-healing process. Generally, maximum IGF ligand expression in wound fluid or tissue is seen during the early hours and/or days post wounding and particularly correlates with cellular proliferation and migration (270, 295, 362). IGF-I levels in human wound fluid are at maximum levels 24 h post injury, then return to baseline at the completion of healing (362). In a pig model, IGF-I abundance was elevated in wound fluid for the first 6 d post wounding and was then undetectable, correlating with complete reepithelialization (295). The derivation of IGFs in wound fluid may include migrating keratinocytes, epithelial cells of adjacent hair follicles, granulation tissue fibroblasts, inflammatory cells, and serum (270, 271). Retarded wound healing may be partially due to alterations in the IGF system. For example, wounds of diabetic rats exhibit a delay in IGF-I and IGF-II expression (271). A human study by Blakytny et al. (272) revealed that IGF-I is reduced or absent in the epidermis and dermal fibroblasts of diabetic skin and recalcitrant wounds. In humans, chronic skin wounds exhibit up-regulation of IGF-IR expression in the epidermis, which correlates with an expanded proliferative compartment (263). The role of the IGF system in wound healing correlates with IGF activity in cultured keratinocytes. As previously discussed (Section VI.B.2), under culture conditions IGF-I is essential for maximum keratinocyte proliferation and also stimulates keratinocyte migration (123, 285), two features that are critical to complete reepithelialization in wounds.

Hypertrophic and keloid scars, which display an abnormal dermal fibroproliferative reaction and excess collagen deposition, also exhibit alterations in the IGF system (240, 363). Deep dermal fibroblasts of hypertrophic scars express higher levels of IGFBP-3 that, unlike normal fibroblasts, are not down-regulated by TGFβ (363). Keloid fibroblasts overexpress IGF-IRs, and their invasive capacity is increased by IGF-I stimulation (240).

Although the application of IGF-I (or GH) via systemic means can lead to improved wound healing in several instances, the problem of adverse side effects, including hypoglycemia, electrolyte imbalance, and edema, cannot be ignored (361). Furthermore, transfer to the extravascular space may compromise the effectiveness of systemic IGF-I treatment. Thus, investigations have also addressed the efficacy of locally administered IGF-I. Application of IGF-I to the wound site results in an increase in reepithelialization (364), angiogenesis (365), dermal collagen deposition (366), and improved wound-healing rates. Local IGF-I treatment can also ameliorate impaired wound healing associated with diabetes (367) or corticosteroid treatment (368).

C. IGFBPs: regulation and usefulness in wound healing

The efficacy of systemic and locally expressed IGFs may be modulated by changes in local and systemic IGFBP expression. Thus, the involvement of IGFBPs and their proteases in regulating IGF action in skin wound healing has been investigated both in serum and at the tissue level.

Severely burned patients exhibit a reduction in serum IGFBP-3 levels, which correlates with decreased serum IGF-I, for at least 20 d post burn (369). Furthermore, while IGFBP-1 levels remain relatively constant, serum IGFBP-2 and -4 are raised. The changes in serum IGFBP abundance may affect the transfer of IGF-I or IGF-II to extravascular sites in the wound tissue and thus impair or enhance the repair mechanism (370).

IGFBP-3 and -1, in conjunction with IGF-I, are initially up-regulated and then return to baseline levels in human wound fluid (362). In a Hunt-Schilling rat wound chamber model, IGFBP profiles of wound fluid were similar to that seen in serum except that IGFBP-3 levels were reduced by the presence of protease activity (371). A similar IGFBP profile and protease activity have been described for human interstitial fluid (225, 371). Again, the cellular derivation of IGFBPs in wound fluid is not clear but may be a reflection of both serum and/or tissue derivation (134, 221).

The efficacy of systemic GH and IGF-I treatment in wound healing may also be determined by regulation of IGFBP production at the tissue level. Interestingly, a recent study showed that systemic administration of GH and IGF-I up-regulates IGFBP-3 mRNA and protein in whole skin samples (372). Although the precise cellular source was not clarified, it is apparent that serum GH and IGF-I levels can regulate local expression of IGFBP-3 mRNA.

The knowledge that IGFBPs can enhance IGF action has also provided the impetus for assessing the therapeutic advantages of delivering IGF-I in combination with selected IGFBPs at wound sites. Topical IGF-I is a more potent stimulator of wound healing, via reepithelialization and granulation tissue deposition, when coadministered with either dephosphorylated IGFBP-1 or IGFBP-3 (373, 374). Furthermore, complexed IGF-I/IGFBP-1 also enhances collagen contraction (375). The augmentation of IGF-I bioactivity by IGFBP-1 involves the interaction of IGFBP-1, via an Arg-Gly-Asp (RGD) sequence, with the α5β1 integrin receptor (376). The stimulatory effect of dephosphorylated IGFBP-1 also correlates with a reduced affinity for IGF-I (377). Because IGFBP-3 does not contain an RGD sequence, the mechanism of enhanced IGF-I action is not clear. It is possible that IGFBP-3 may extend IGF-I activity by protection from local proteases or by targeting of IGF-I to cell membrane IGF-IRs.

D. Gene therapy: a novel avenue for treatment of wounds

Gene therapy has provided another avenue with which to explore the therapeutic value of IGF-I or GH in wound healing. Moreover, the targeted expression of cDNA constructs in skin cells ameliorates the potentially deleterious side effects of systemic cytokine administration (refer to Section VIII.B). In rats, intradermal injected or liposome encapsulated application of IGF-I cDNA plasmid constructs to thermal burns increased local IGF-I mRNA and protein abundance, stimulated keratinocyte proliferation, and accelerated reepitheialization (10, 378–380). In addition, the improvement of burn wounds by the IGF-I cDNA application correlated with a reduction of IL-1β and TNFα mRNA expression and therefore a possible reduction in prolonged local inflammation (381).

Keratinocytes transfected with GH cDNA constructs and transplanted onto full-thickness wounds of pigs secreted GH into the wound fluid for 10 d but did not result in a greater healing rate when compared with transplanted, nontransfected keratinocytes (382). The lack of a GH effect on wound healing is in contrast to that seen in burns patients (as discussed in Section VIII.A). These data therefore suggest that the concentration of GH in the transgenic wound fluid was not enough to enhance healing. These data also suggest that GH derived from the transgene could not directly stimulate keratinocyte proliferation and support the general observation that GH does not stimulate proliferation of keratinocytes even if GHRs are expressed by keratinocytes (as discussed in Section VI.A.2.a). Furthermore, the study suggests that GH did not utilize IGF-I as a mediator of its actions as previously discussed (261). Finally, the loss of local GH expression may have been due to the loss of transfected keratinocytes through the process of differentiation and perhaps led to the lack of a significant improvement in wound-healing rates. This study potentially highlights the importance of isolating and targeting keratinocyte SC if a transgene is required permanently in the epidermis (383).

IX. Summary and Discussion

The widespread distribution of components of both the GH and IGF systems is perhaps a reflection of the importance of the ligands throughout the whole organism, and this review clearly indicates that the skin is no exception. Collectively, the literature points to the significance of the GH and IGF systems in normal skin growth and suggests that disturbances in these pathways are involved in the pathophysiology of distinct skin perturbation, including psoriasis and a range of epidermal carcinomas. Indeed, the majority of studies implicate the direct actions of the IGF system in modulating epidermal homeostasis. More specifically, IGF-I acts on epidermal keratinocytes and melanocytes to modulate several functions including proliferation, differentiation, migration, and survival. In contrast, it seems that GH primarily ensures normal skin growth and development by direct stimulation of GHRs located on dermal cells, particularly fibroblasts, and then appears to use IGF-I as a mediator of its actions in the dermis and epidermis. It also seems likely that IGF-I can modulate epidermal homeostasis without needing to be a mediator of GH action. Although GHRs on keratinocytes have been detected by immunohistochemical methodology, the direct actions of GH in these cells remains unknown. In vitro analysis, however, suggests that GH may directly regulate melanocyte proliferation, but this role remains to be clarified in skin.

The current literature, primarily emanating from our laboratory, implicates IGFBP-3 in regulating epidermal homeostasis, possibly via regulation of the initial stages of keratinocyte terminal differentiation (221, 226, 227, 264, 276, 277, 279, 280). Figure 5A is a very simplified model of IGF action in the epidermis of normal skin, including the potential mechanisms of IGFBP-3 activity. Paracrine IGFs (particularly IGF-I, derived from dermal fibroblasts; orange cells) (266), serum, epidermal melanocytes (red cell) (57), and perhaps suprabasal keratinocytes (216), cross the ECM milieu to stimulate IGF-IRs expressed by basal keratinocytes (263, 297). IGF-IR stimulation ensures keratinocyte proliferation (285), particularly in TA cells, and cell survival (103, 384). In this hypothetical model, IGFBP-3 is expressed by SC and TA keratinocytes and is absent in PMD cells. Such distribution correlates with our observations that IGFBP-3 is strictly localized to selected basal keratinocytes (221, 277), is not present in the upper epidermal layers (suprabasal, granular, corneocytes), and is lost with the induction of calcium-induced differentiation of a keratinocyte cell line (264). Because IGFBP-3 colocalizes with IGF-IRs (297), it would control the IGF-stimulated proliferative response of basal keratinocytes, predominantly the TA cells, and ensure that the cells are not overstimulated by local IGFs. Specifically, IGFBP-3 would regulate IGF-IR stimulation by inhibiting or augmenting IGF/IGF-IR interaction, with the emphasis on inhibition. Appropriate IGFBP-3 modulated IGF-IR stimulation could be achieved by two mechanisms. Firstly, IGF/IGFBP-3 complexes may associate with cell surfaces or ECM, leading to IGFBP-3 proteolytic cleavage, a concomitant reduction in IGF affinity and ultimately IGF-IR activation (133, 134, 248). Secondly, proteases could also produce fragments of IGFBP-3 that inhibit IGF-I-stimulated proliferation of keratinocytes, akin to the unknown mechanism(s) described in fibroblasts (148). Finally, IGFBP-3 may also inhibit keratinocyte proliferation irrespective of IGF-IR activity, possibly via its own unidentified cell surface receptor (144, 149, 257, 258).

Several studies, particularly those of psoriatic skin (Section VII.A.2), support the above-mentioned model and thus a primarily inhibitory role for IGFBP-3 in the epidermis. IGFBP-3 inhibits IGF-I-stimulated proliferation of keratinocytes (276) and is present in quiescent and absent from proliferative basal keratinocytes in the hyperplastic epidermis of psoriatic skin lesions (277). Additionally, blister fluid from psoriatic skin exhibits elevated levels of intact IGFBP-3, suggesting that in normal skin proteolytic fragments of IGFBP-3 may modulate epidermal proliferation via IGF-dependent and IGF-independent actions (previously discussed in Section VII.A.2) (132, 385). Our observation that keratinocyte-derived IGFBP-3 is profoundly down-regulated by high levels of EGF (5–50 ng/ml) (279, 280) also correlates with perturbations in the EGF system in psoriatic skin. Specifically, TGFα (an alternative ligand for the EGFR) and EGFRs are overexpressed in the psoriatic epidermis (386–389) and could thus aid suppression of normal IGFBP-3 expression in the basal layer (particularly TA cells). The inhibition of IGFBP-3 could then further contribute to the observed greater IGF-I/EGF proliferative synergy in psoriatic keratinocytes (293–295). Figure 5B highlights these putative mechanisms of IGF-I/EGF action in psoriasis and particularly emphasizes how the loss of IGFBP-3 in TA cells of the basal layer would contribute to the grossly disturbed and hyperplastic epidermis.

In the current model of normal skin (Fig. 5A), IGFBP-3 modulates the early stages of keratinocyte terminal differentiation, i.e., the transition from keratinocyte SC to TA cells. An absence of IGFBP-3 in the upper layers of the epidermis also supports the postulate that IGFBP-3 primarily acts as an inhibitor of keratinocyte proliferation. Specifically, under normal circumstances, proliferative keratinocytes are present only in the basal layer of the epidermis (170, 187–191). The loss and continued suppression of IGFBP-3 expression in the upper epidermal layers may be partially supported by TGFβ1 (280), expressed by suprabasal keratinocytes (246), EGFR stimulation (317, 386, 390), and the increasing calcium concentration (another essential requirement for differentiation) (180, 264, 391, 392).

The juxtaposition of melanocytes (red cell, Fig. 5A) with basal keratinocytes may aid the suppression of melanocyte proliferation and thus maintenance of the epidermal melanocyte unit (i.e., the ratio of 1:36, melanocytes to keratinocytes; Ref. 167). In particular, because epidermal keratinocytes require the IGF system for growth and maturation, it is possible that melanocytes would also be exposed to IGF-I in the ECM milieu. Because IGF-I is a primary regulator of melanocyte proliferation (304), the ability to blunt IGF-I action seems imperative. It therefore seems possible that keratinocyte-derived IGFBP-3 and melanocyte-derived IGFBP-4 (262) would play an important role in modulating melanocyte IGF-I action, with local IGF-I being derived from dermal fibroblasts or melanocytes themselves.

Future analyses should continue to dissect the functional role of GH and IGFs in ensuring normal skin growth and development, particularly the epidermis. More specifically, in vivo functional analyses are required to further elucidate the mechanism by which IGFBP-3 modulates epidermal homeostasis. It is interesting to note, however, that in psoriatic lesions (277) and a range of skin perturbations exhibiting a hyperplastic epidermis (i.e., dermatofibroma, seborrheic keratosis, squamous keratosis; S. R. Edmondson, G. A. Werther, and C. J. Wright, unpublished observations), IGFBP-3 mRNA is absent in the basal layer. This observation alone strongly implicates IGFBP-3 in the protection of keratinocytes from overexposure to paracrine IGF-I.

Many perturbations of the epidermis, both benign (e.g., psoriasis) and malignant (e.g., melanomas), exhibit altered IGF signaling. Indeed, our recent success at blocking epidermal IGF-IR activity, via antisense oligonucleotide therapy, and normalizing epidermal thickness in psoriatic skin provides an excitingly strong basis for the development of novel drugs to alleviate the debilitating effects of epidermal overgrowth (12). Thus, the future may see a new suite of drugs that specifically modulate the activity of the IGF system and ultimately lead to normalization of hyperplastic skin lesions.

Finally, the advent of gene therapy has thrown the epidermis into the spotlight (393). The accessibility, ability to culture keratinocytes and melanocytes, ability to insert genes of interest, and regrafting procedures mean that keratinocytes and melanocytes are attractive commodities. Current literature strongly indicates that GH and IGF cDNAs will provide the basis for gene therapy regimes designed to accelerate the repair of severely damaged skin or recalcitrant wounds.

Acknowledgments

This work was supported by the National Health and Medical Research Council of Australia, Murdoch Children’s Research Institute, and the Royal Children’s Hospital Research Institute.

Abbreviations:

 
• bFGF,

  Basic fibroblast growth factor;

 
• ECM,

  extracellular matrix;

 
• EGF,

  epidermal growth factor;

 
• EGFR,

  EGF receptor;

 
• GHBP,

  GH binding protein;

 
• GHR,

  GH receptor;

 
• GHRtr,

  truncated GHR;

 
• HaCaT,

  human keratinocyte cell line;

 
• IGFBP,

  IGF binding protein;

 
• IGF-IR,

  IGF-I receptor;

 
• IR,

  insulin receptor;

 
• IRS,

  IR substrate;

 
• M-6-PR,

  mannose-6-phosphate receptor;

 
• MAP,

  mitogen-activated protein;

 
• PI-3,

  phophatidyl-inositol-3′;

 
• PMD,

  postmitotic differentiating;

 
• RXR-α,

  retinoid-X-receptor α;

 
• SC,

  stem cell(s);

 
• TA,

  transit amplifying;

 
• TPA,

  12-O-tetradecanoylphorbol-13-acetate.