Skin Barrier Disruption - A Requirement for Allergen Sensitization?

Abstract

For at least half a century, noninvasive techniques have been available to quantify skin barrier function, and these have shown that a number of human skin conditions and disorders are associated with defects in skin permeability. In the last decade, several genes responsible for skin barrier defects observed in both monogenetic and complex, polygenic disorders have been elucidated and functionally characterized. This has led to an explosion of work in the last six years that has identified pathways connecting epidermal barrier disruption and antigen uptake as well as the quality and/or magnitude of the antigen-specific adaptive immune response. This review will introduce the notion that diseases arise from the dynamic crosstalk that occurs between the skin barrier and immune system using atopic dermatitis or eczema as the disease prototype. Nevertheless, the concepts put forth are highly relevant to a number of antigen-driven disorders for which skin barrier is at least transiently compromised such as psoriasis, allergic contact dermatitis and blistering disorders.

The skin provides a vital barrier structure that protects vertebrates from both routine and extreme environments including exposure to antigens, solvents, ultraviolet light, detergents, microorganisms, toxins, nanoparticles and a variety of physical insults (Elias, 2006). In terrestrial vertebrates, the epidermis, where most of the skin barrier function resides, is highly stratified and has an outermost layer that is cornified. Recent findings have shown that epidermal barrier dysfunction is pathologically involved in a variety of common, antigen-driven skin diseases, including psoriasis and atopic dermatitis (AD). In this review, we will briefly describe i) the barrier system of the human epidermis, ii) human disorders associated with skin barrier defects and allergen sensitization, iii) murine studies that have helped to clarify pathways connecting barrier and immunity and iv) epithelial-derived immune adjuvants released in response to barrier disruption. We will focus on immune recognition of allergens as a paradigm for all incomplete and complete environmental and microbial antigens. It is important to note that the skin microbiome, which consists of both commensal and pathogenic bacteria, affects the skin barrier and epithelial innate immune responses. Consequently, skin microbes are thought to play a critical role in the development of atopic dermatitis. These topics will not be discussed further as they are reviewed in this 75^th Anniversary series (Gallo, 2012; Kong and Segre, 2012).

Cutaneous allergen sensitization is a critical and early event in the pathogenesis of AD, but this may also be true for several other atopic disorders where defects in skin barrier genes associate with diseases that manifest in other organs including food allergy, asthma, allergic rhinitis and eosinophil esophagitis (Blanchard et al., 2010; Bremmer et al., 2008; Brown et al., 2011a; Brown et al., 2008; Weidinger et al., 2008). In subjects with AD, the initial exposure to allergens (sensitization phase) induces a systemic “allergic” T Helper (Th) 2 cell response that is magnified with each subsequent exposure (effector phase). Critical features of a Th2 immune response include the local production of Th2 cytokines (IL4, IL5 and IL13), bone marrow production, prolonged survival and activation of eosinophils and mast cells and production of allergen-specific IgE. For a long time, allergic diseases were considered primarily immunologic disorders. As a consequence, research and drug development focused on modifying the Th2 effector phase (Hanifin et al., 1985). Over the last decade, there has been a shift in our thinking, with the epithelium now recognized as a critical player in the development of allergic sensitization (Bulek et al., 2010; De Benedetto et al., 2009; Holgate, 2007). In particular, compelling data from human and mouse studies have implicated skin barrier impairment as an indispensable event in allergen sensitization (Cork et al., 2009; Fallon et al., 2009; Jin et al., 2009; O’Regan and Irvine, 2010; Oyoshi et al., 2009; Palmer et al., 2006; Smith et al., 2006; Spergel, 2010). The specific pathways connecting epidermal barrier disruption to allergen sensitization are beginning to be elucidated. The overriding hypothesis is that epidermal disruption would allow skin resident antigen presenting cells such as Langerhans cells (LC; epidermis) or dendritic cells (DC; dermis) to capture environmental antigens (Kubo et al., 2009) (Figure 1 & 2). Additionally, barrier-disrupted keratinocytes would release immune adjuvants that activate and mature LC/DCs as well as affect their ability to direct naïve Th cell polarization and thereby affect the character of the T helper response (Figure 3). Whether the adaptive immune response that ensues can feedback and affect barrier function has not been fully explored. In this review, we will summarize the evidence for and possible mechanism(s) involved in the induction of allergen sensitization in the context of skin barrier defects.

Figure 1. TJ penetration of LC dendrites.

(a and c) Layers of the stratum granulosum (SG) are designated SG1, SG2, and SG3, counting from the skin surface inwards. In murine epidermis TJs are found in the intercellular spaces between SG2 cells. Blue dotted lines represent TJs. (b and d) Activated LCs (e.g. MHC class II positive) elongate their dendrites to dock with and penetrate epidermal TJs. (a) ZO-1 and claudin-1 accumulate at penetration points (yellow arrows), where novel tricellular TJs are formed between LC dendrites and surrounding keratinocytes to prevent significant barrier disturbance. (b) Rotated views of an activated LC dendrite are shown. EZ-link sulfo-NHS-LC-biotin was applied topically on mouse skin. Trace amounts of this biotin tracer are observed in LC dendrites that have crossed the TJ barrier (arrowheads; upper b). (The images were originally published in JOURNAL OF EXPERIMENTAL MEDICINE (Doi: 10.1084/jem.20091527 (Kubo et al., 2009)).

©Kubo et al., 2009. Originally published in JOURNAL OF EXPERIMENTAL MEDICINE. doi: 10.1084/jem.20091527

Figure 2.

The epidermis has two formidable barrier structures that are analogous to a castle’s moat (SC; blue water) and portcullis (TJ; white gate). Epicutaneous sensitization requires that the antigen be engulfed by an antigen presenting cell such as an epidermal dendritic cell or Langerhans cell (LC). (a) Under resting conditions the immune system does not respond to environmental factors such as allergens. (b) When the SC is breached (e.g. drawbridge is down) allergens may cross the moat, but will still be blocked by an intact TJ barrier (portcullis). LC dendrites are found below TJ (e.g. behind the portcullis). (c) When the portcullis is opened (TJ loosened), LC dendrites extend through these weakened TJ and take up allergens and initiate an adaptive immune response. Additionally, it is hypothesized that small allergens may penetrate leaky TJ and be taken up by LC/DC whose dendrites are below the TJ. It is not clear whether a transient break in both epidermal barriers (SC and TJ) are required for LC dendrites to penetrate TJs. This dual barrier system uniquely found in the skin may explain why we do not respond to a myriad of antigens that reach our skin surface daily and therefore why our skin is usually uninflammed.

Figure 3. Epidermal barrier function and immune responses are tightly linked.

Primary barrier defects lead to the release of a number of epidermal-derived mediators including ones that are considered pro-Th2 and pro-Th17. Some of these cytokines are released in an autocrine fashion through proteinase-activated receptors (PAR)-2 activation or by epithelial–derived “danger signals” that act on innate immune receptors expressed on keratinocytes. The presumption is that barrier disruption comes in different flavors each resulting in a specific adjuvant profile that would either favor a Th2, Th17 or another adaptive immune response to an antigen. Several Th2 and Th17 products establish an autocrine feedback loop and further aggravate barrier disruption (e.g. secondary barrier defects). To date, this has been best characterized for Th2 cytokines (Huppert et al., 2010; Sehra et al., 2010).

I. Overview of the epidermal barrier

Stratum corneum – the proverbial “moat” that is the preliminary defense for the “castle”

The horny layer of the epidermis was initially considered to be a loose collection of amorphous keratin filaments separated by wide empty spaces (Rothman, 1954). The “basket-weave” appearance of this layer as it is observed in formalin-fixed specimens has long misled histopathologists. Kligman and Christophers demonstrated that this appearance was an artifact of sample preparation. They were the first to recognize its cellular nature coining the term “corneocytes” (Christophers and Kligman, 1964b; Kligman, 2006; Kligman, 2011). Corneocytes were noted to appear as disk-shaped cells with a polygonal outline (Goldschmidt and Kligman, 1967), and the sodium hydroxide immersion technique demonstrated that the stratum corneum (SC) was about 15 cells thick in adult human skin samples from most body sites (Christophers and Kligman, 1964a). This same technique also demonstrated that corneocytes are encased by well-defined “cornified envelopes”, which have been extensively studied since then (Candi et al., 2005; Eckert et al., 1993; Madison, 2003; Proksch et al., 2008). At first, the inter-corneocyte space seemed empty as viewed by electron microscopy. This turned out to be an artifact of fixation and sample processing. Ruthenium tetroxide fixation and staining demonstrated that the intercellular spaces were filled with stacks of lamellar structures, which are now recognized as multiple stripes of hydrophobic and hydrophilic structures formed by various intercellular lipids (Elias, 2005). These intercellular spaces also include various proteases that control desquamation and antimicrobial peptides that act as a microbial barrier and control the growth of both commensal and pathogenic bacteria (Braff et al., 2005).

Studies demonstrating enhanced skin permeability after SC removal highlight the importance of the SC as a barrier structure (Blank and Gould, 1959; Malkinson, 1958; Monash, 1957). Although, it is not clear whether several or all of these approaches to disrupt the SC may also disrupt the tight junctions (TJ) found immediately below the SC. Dr. Elias and his colleagues have published a large number of studies characterizing the SC components responsible for barrier function and how they are altered in normal and pathologic conditions (reviewed in (Elias, 2005; Elias, 2006). It was initially proposed by Scheuplein in 1971 that exogenous substances transited though the SC by direct transmembrane diffusion (Scheuplein and Blank, 1971). Later, it became clear that the intercorneocyte space provided the pathway through which exogenous, more lipophilic substances pass to reach the sub-SC layers of the epidermis. The surface proteins of the “cornified envelope” were shown to be tightly bound to hydroxyl-acylsphingosine molecules, which suggests that each corneocyte envelope has its own lipid envelope providing a barrier against the passage of water and water-soluble substances (Swartzendruber et al., 1987). Various physical and chemical penetration enhancers including exposure to water alone were shown to have the ability to disrupt the intercellular multi-laminar membranes, creating openings within the intercellular spaces called “lacunae” through which hydrophilic and hydrophobic substances could penetrate (Menon and Elias, 1997; Warner et al., 1999). In summary, the SC is composed of two components, protein-rich corneocytes and intercellular lipid domains, and it has been likened to a “brick and mortar” structure (Elias, 1981).

Tight junction barrier - the proverbial “portcullis” that is the final defense for the “castle”

In the 1940s, Dr. Rothman first suggested that another barrier structure in the skin controlled diffusion of water and solutes through the paracellular pathway (Rothman and Flesch, 1944). In his words, “The impermeability of the skin to water and electrolytes is caused neither by the presence of a greasy-waxy cover of the skin nor by the presence of the horny layer. The seat of the absorption-barrier is to be placed in the transitional layers between cornified and noncornified epithelium, i.e., in the stratum granulosum (SG)…” (Rothman and Flesch, 1944). This absorption-barrier is called TJ and is the only barrier structure in mono-layered or simple epithelia demarcating the body surface of urochordates but has also been recognized more recently in the stratified epithelia of vertebrate skin. Thus, TJ are an ancient barrier system of the body surface, which have been conserved from urochordates and fish that lack SC to mammals. Although Hashimoto had described the appearance of TJ structures in the granular layer of human epidermis in 1971 (Hashimoto, 1971), this barrier has long been overlooked or ignored in mammalian skin. The importance of this barrier structure in mammalian epidermis was revived by identification of claudins as integral transmembrane proteins found in all TJ. The profound epidermal permeability abnormalities observed in claudin-1 knockout mice and claudin-6 transgenic mice solidified the importance of this second skin barrier structure (Furuse et al., 2002; Turksen and Troy, 2002). TJs seal the intercellular spaces between stratum granulosum keratinocytes (Furuse et al., 2002; Kubo et al., 2009). TJs are not just physical barriers; they exhibit ion and size selectivity and their barrier function varies significantly in ‘tightness’ depending on cell type and physiological requirements, enabling dynamic regulation of substances that traffic between extra-TJ and inside-TJ compartments including dendritic processes from antigen presenting cells (Kubo et al., 2009; Tsukita and Furuse, 2002)(Figure 1).

II. Skin barrier defects promote allergen sensitization - evidence based on human diseases

Disorders characterized by enhanced desquamation – Netherton’s Syndrome and Peeling Skin Syndrome Type B

Association of skin barrier defects with allergic manifestations has been demonstrated in two genodermatosis that are disorders of desquamation, namely Netherton’s syndrome (NS; OMIM 256500) and Peeling skin syndrome, type B (PSS-B; OMIM 270300) (Table 1). Both diseases are characterized by superficial intra-epidermal detachment and atopic diathesis. In the SC, the corneocytes are tightly bound together via corneodesmosomes, the end product of epidermal desmosomes modified by the incorporation of corneodesmosin (CDSN). In the desquamation process, kallikrein (KLK)-related peptidases such as KLK5 and KLK7 degrade CDSN and induce cleavage of corneodesmosomes, resulting in detachment of corneocytes (Ovaere et al., 2009). The protease activity of KLKs is inhibited by lymphoepithelial Kazal-type 5 serine protease inhibitor (LEKTI), encoded by the serine protease inhibitor, Kazal type 5 (SPINK5) gene. Loss-of-function mutations in SPINK5 (Chr 5q32) leading to accelerated desquamation, is the defect in Netherton’s syndrome, a severe autosomal recessive ichthyotic condition characterized by a chronic atopic dermatitis-like condition, allergen sensitization and other atopic disorders such as asthma and allergic rhinitis (Frenk and Mevorah, 1972). Interestingly, coding polymorphisms in the SPINK5 gene have been found in association with AD and disease severity (Cork et al., 2009; Walley et al., 2001), and more recently with food allergy in children with AD (Kusunoki et al., 2005). Although Hubiche et al. (Hubiche et al., 2007) did not find an association in a French AD population, they did observe an association between the E420K polymorphism and serum IgE levels, suggesting that this barrier defect predisposes subjects to a Th2 response to environmental allergens.

Table I. Epidermal barrier defects and allergen sensitization in human diseases.

Human disease	Barrier-related defects	Local Immune phenotype	ATOPY/ Th2 immune response	Reference (PMID)
Desquamation abnormalities
Netherton’s syndrome	↓ SPINK5; ↑ activity of proteases (ELA2, KLK5, 7 and 14); abnormal FLG processing & lipid lamellae	↑ TSLP	Association with atopic disorders and ↑ IgE	10835624 19414552 20179351
Peeling Skin Type B	↓ CDSN	ND	Association with atopic disorders, ↑ IgE and ↑ EOS	21191406 20691404
Disturbed Lipid Metabolism
Ichthyosis Prematurity Syndrome	↓ FATP4; membrane inclusions in SC	ND	Association with AD- like condition, ↑ EOS	19119129
Abnormalities in SC structural proteins
Ichthyosis Vulgaris	↓ or absent FLG	ND	Association with atopic disorders. ↑IgE	16550169 18159904
Atopic Dermatitis	Dysregulated EDC (including ↓ FLG); ↓ proteases inhibitors and ↑ proteases; lipid abnormalities; TJ abnormalities, ↓ CLDN1 and 23, ↑ Cx26	↑ pro-Th2 cytokines (e.g. TSLP, IL25, IL33); Th2, Th22 and Th17 (only in acute lesions); Th2 and Th1 in chronic lesions.	Association with atopic disorders, ↑ IgE and ↑ EOS	16550169 18396323 19494826 21163515 21388665 21419481
Psoriasis	Dysregulated EDC	Th1 and Th17	Rarely observe modest ↑ IgE	18432274 21388665 21419481
Full thickness epidermal disruption
Bullous Pemphigoid	Subepidermal blisters	Th2 cytokines, tissue eosinophils (early lesions)	↑ IgE and ↑ EOS	6345605 9452886 12783647
Pemphigus Vulgaris	Suprabasilar acantholysis	Th2 cytokines, tissue eosinophils (early lesions)	↑ IgE	4217592 20015693 11161984
Epidermolysis bullosa pruriginosa	Anchoring fibril abnormalities, mutation in COL7A1	ND	↑ IgE	8204470

Abbreviations: Abs, antibodies; CDSN, corneodesmosin;CLDN, claudin; COL7A1, collagen, type VII, alpha-1; Cx26, connexin 26; EDC, epidermal differentiation complex; ELA2, elastase 2; EOS, eosinophilia; FATP4, fatty acid transport protein 4; FLG, filaggrin; KLK, kallikrein; ND, not determined; SPINK5, serine peptidase inhibitor Kazal type 5; SC, stratum corneum; TJ, Tight junction; TSLP, thymic stromal lymphopoietin.

Recently, two distinct mutations in the CDSN gene (Chr 6p21) have been shown to be associated with PSS-type B (Israeli et al., 2011), an autosomal-recessive ichthyosiform erythroderma characterized by peeling of the skin and allergen sensitization (Israeli et al., 2011; Oji et al., 2010). PSS-B patients have been reported to have high serum IgE levels, eosinophilia and food allergies (Oji et al., 2010). In summary, defects in two different proteins that result in enhanced SC desquamation are strongly associated with a Th2 adaptive immune response to environmental allergens as noted by elevated IgE levels and peripheral eosinophilia and clinical manifestations of an atopic disease(s). It is important to note that the CDSN gene is located in PSORS1, one of the major psoriasis (PS) susceptibility loci, and polymorphisms in this gene have been associated with PS in some populations (Capon et al., 2004; Helms et al., 2005; Jenisch et al., 1999). Since psoriasis is a disease of delayed desquamation, it is assumed that the CDSN mutations observed in PS would likely inhibit their degradation by kallikreins and therefore would not likely explain PS subjects barrier defect.

Disorder of disturbed lipid metabolism - Ichthyosis Prematurity Syndrome

Ichthyosis prematurity syndrome (IPS; Chr 9q34) is a rare syndrome characterized by the clinical triad of premature birth, thick caseous desquamating epidermis, and neonatal asphyxia (Niemi et al., 1993). The diagnosis is often made by characteristic ultrastructural findings demonstrating conspicuous membrane inclusions in the SC, and most cases are associated with mutations in a gene encoding for the fatty acid transport protein 4 (FATP4) (Klar et al., 2009; Sobol et al., 2011). The FATP4 protein plays a central role in the transport and activation of fatty acids in the endoplasmic reticulum of keratinocytes and is thought to be important in the maintenance of normal epidermal barrier function (Milger et al., 2006). Patients with this condition can have peripheral eosinophilia, an atopic dermatitis-like condition, an extensive keratosis pilaris-like condition and a family history of atopic disorders (Bygum et al., 2008). Although the clinical features of the few subjects reported in the literature to date are incomplete, there is still a suggestion that a SC lipid defect, that causes a barrier phenotype in genetically altered mice (Moulson et al., 2007), is associated with markers of systemic Th2 polarity (e.g. eosinophilia) and atopy (Table 1).

Disorders characterized by reduced expression of a key SC structural protein - Ichthyosis Vulgaris and Atopic Dermatitis

Filaggrin (FLG) is expressed in SG layers as a > 400-kDa precursor protein called profilaggrin. Profilaggrin is dephosphorylated and cleaved by proteases into FLG monomers, with keratin-binding activities which are thought to contribute to the cell compaction observed in the lower SC (Dale et al., 1978). FLG monomers are further degraded into “natural moisturizing factors” that maintain hydration of the upper SC (Rawlings and Matts, 2005; Sandilands et al., 2009; Scott and Harding, 1986) and also have anti-staphylococcal properties in vitro (Miajlovic et al., 2010). Mutations in the FLG gene were identified initially as the cause of ichthyosis vulgaris (IV; OMIM 146700; Chr 1q21) and subsequently as a major predisposing factor for AD (Table 1) (Palmer et al., 2006; Smith et al., 2006). To date, the strong association of FLG mutations with AD is one of the most robust genotype–phenotype linkages observed in human complex genetic disorders (van den Oord and Sheikh, 2009). Several case–control studies have also demonstrated strong association between FLG mutations and early AD onset, disease severity, eczema herpeticum, AD-related asthma and greater allergen sensitization (Gao et al., 2009; Schuttelaar et al., 2009; van den Oord and Sheikh, 2009; Weidinger et al., 2006). Interestingly, several studies have shown FLG mutations confer a substantial risk for other atopic disorders recognized as Th2 polarized diseases including allergic rhinitis, IgE-mediated peanut allergy, and eosinophilic esophagitis. These associations were observed even after controlling for co-existent AD (Blanchard et al., 2010; Brown et al., 2011b; Weidinger et al., 2008). This is a bit difficult to understand as FLG immunostaining is restricted to skin, oral mucosa and the cornified epithelium of the nasal vestibule with no detectable staining observed in epithelium from bronchial biopsies as well as gastrointestinal epithelium (De Benedetto et al., 2008; Weidinger et al., 2008; Ying et al., 2006). Therefore, FLG mutations are unlikely to affect barrier function and allergen sensitization in the organs where these diseases manifest. One interpretation of these findings is that FLG mutations may drive disease at distant mucosal sites by enabling allergen sensitization through a defective skin (and/or oral) barrier while subsequent allergen responsiveness occurs in these other epithelial beds (e.g. upper airway or gastrointestinal tract). This model seems feasible as we know that not all subjects with FLG mutations develop AD. For example, the carrier frequency is as high as 12% in healthy Northern European controls, suggesting that these mutations by themselves may not be sufficient to induce AD and/or a Th2 adaptive immune response. Additionally, the model presupposes that subjects with allergic rhinitis and food allergy likely have a barrier defect in the relevant epithelial surface that may be either genetic or acquired to explain their organ-specific disease manifestations. In the case of allergic rhinitis, the barrier defect may occur in the context of a viral upper respiratory infection which often proceeds an exacerbation of the disease, and in the case of food allergy one could argue that the infants and young children who develop this allergic disease have a delay in the normal intestinal barrier maturation that occurs shortly after birth.

Between 37 to 70% of IV patients present with atopic conditions, most commonly AD, but also allergic rhinitis and asthma (Bremmer et al., 2008; Brown et al., 2008; Wells, 1966). Interestingly, some IV subjects have allergic rhinitis and high IgE levels without concomitant AD providing further support for the notion that FLG mutations simply confer a risk for allergen sensitization through the skin (Oji et al., 2009). Somewhat surprisingly, a small German IV cohort study found no difference in the prevalence of atopic disorders based on FLG genotype (e.g. one or double mutation) (Oji et al., 2009). They also evaluated the number of epidermal dendritic cells identified by CD1a staining, as a marker of early immunologic activity, based on previous work demonstrating that subjects with more severe AD had increased epidermal DCs (Novak et al., 2004). They observed increased epidermal CD1a+ cells only in the atopic IV subgroups, which was independent of FLG genotype suggesting that cutaneous reactivity in these subjects was not determined by the presence or absence of FLG expression. Collectively, clinical assessments of AD and IV subjects suggest that FLG mutations predispose subjects to allergen sensitization but that these mutations are not sufficient as other genetic as well as environmental influences are likely promoting the Th2 immune response observed in susceptible individuals. One speculation is that other epidermal barrier defects might also contribute to and/or modulate the epicutaneous sensitization and ultimately the allergic phenotype. Table 1 highlights several human skin diseases that are thought to have a barrier defect and summarizes what we know about their local and systemic adaptive immune responses.

It is well known that the magnitude of the barrier defect in AD, as measured by transepidermal water loss (TEWL) at nonlesional sites, correlates with disease severity and serum IgE (Gupta et al., 2008; Hon et al., 2008; Lee et al., 2006). Using this physiological measure of epidermal barrier function, Boralevi et al. demonstrated that AD infants with two or more positive allergen patch tests had higher TEWL than infants with one or no positive tests (Boralevi et al., 2008). Further studies are needed to clarify the relationship between physiologic measures of barrier health and integrity (e.g. TEWL, surface pH, SC hydration, SC cohesion, lipid composition of SC, barrier recovery) and allergen sensitization in human subjects. Surprisingly, the biochemical basis for elevated TEWL is still not known. The literature strongly suggests that nonlesional TEWL abnormalities are not explained by FLG mutations (Hubiche et al., 2007; Jungersted et al., 2010; Nemoto-Hasebe et al., 2009; O’Regan et al., 2010). Although, Flohr et al.(Flohr et al., 2010) showed that FLG mutations were associated with higher TEWL in clinically normal appearing forearms in a small cohort of 3 month-old infants, which was not dependent on AD status. Collectively, these studies strongly suggest that FLG mutations may play a role in the TEWL abnormalities observed in AD subjects, but they are certainly not the entire story.

FLG genotypes do seem to play a role in a number of physiologic measures of barrier function including skin hydration (Kezic et al., 2008; Scott and Harding, 1986), surface pH, SC cohesion and paracellular permeability of SC with the water soluble tracer, lanthanum (Fluhr et al., 2010; Gruber et al., 2011; Krien and Kermici, 2000). AD subjects have a number of barrier abnormalities in addition to increased TEWL, including increased surface pH, reduced SC hydration, and enhanced barrier recovery. As we discover the biochemical basis for these abnormalities, we will be able to address how each of them interact in a bidirectional way with the immune system (Figure 3).

We have recently demonstrated that AD subjects have a defect in TJ as noted by a remarkable impairment of bioelectric properties of their ex vivo epidermis, and this may be due to reduced expression of the TJ components, claudin-1 and 23 (De Benedetto et al., 2011). Claudin-1 levels were inversely correlated with Th2 biomarkers (IgE and eosinophilia) suggesting that reductions in this TJ protein may affect the systemic immune phenotype or vice versa (De Benedetto et al., 2011). This introduces the notion that AD subjects may have defects in both barrier structures, SC and TJ, and the relative contribution of each may contribute to the heterogeneity characteristic of this disease (e.g. disease onset, natural history, magnitude of allergen sensitization, comorbid atopic conditions). Recent studies of skin biopsies from FLG heterozygous and homozygous IV subjects have demonstrated a FLG gene dose-dependent inhibition in TJ protein expression suggesting that there is a dynamic interaction between these two epidermal barrier structures, SC and TJs (Gruber et al., 2011). Whether these changes affect TJ function is still not known. Developing a better understanding of how these two key epidermal barrier structures interact will likely yield information that will have broader implications than just inflammatory skin disorders. In summary, these studies suggest that damage to both epidermal barriers (SC and TJ) might promote the elongation of LC dendrites through TJ barrier and consequently their uptake of allergens/antigens from the skin surface (Figure 2 & 3).

It is interesting to note that several autoimmune disorders and genodermatosis that result in full epidermal detachment are characterized by Th2 inflammation. For example, in the autoimmune blistering disorders, bullous pemphigoid, herpes gestationis and pemphigus vulgaris there is a clear Th2 response which is observed early in the disease, but it is unclear whether it comes before the full thickness epithelial disruption or as a consequence of it (Arbesman et al., 1974; Borrego et al., 1999; Bushkell and Jordon, 1983; De Pita et al., 1997; Fabbri et al., 2003; Feliciani et al., 1999; Nagel et al., 2010) (Table 1). Similarly, in the genodermatosis, epidermolysis bullosa pruriginosa (OMIM 60412) (Mellerio et al., 1999; Yamasaki et al., 1997), a skin fragility disorder caused by mutations in the COL7A1 gene, and characterized by anchoring fibril abnormalities and sublamina densa blistering, the patients have been reported to have high serum IgE and eosinophilia (McGrath et al., 1994) (Table 1). The complete repertoire of IgE antigen-specificity is unclear even in bullous pemphigoid, herpes gestationis and pemphigus vulgaris where BP180, BP230 or desmoglein 3 accounts for only part of the serum IgE measured in these patients. The assumption is that these patients have IgE reactivity to other antigens, and the specificity of these responses would be interesting to characterize further. One is left wondering if the profound barrier disruption characteristic of these disorders promotes the production of epithelial-derived Th2 adjuvants (e.g. TSLP, IL33, IL25, etc).

III. Skin barrier defects promote allergen sensitization – What have murine models taught us?

There are a remarkable number of mouse models for which a barrier defect phenotype has been observed (Table 2). We have chosen a few examples to highlight the complexity inherent to epidermal barrier integrity. We have highlighted genetically altered mice that denote the import of both SC and TJ structural proteins as well as components of other intercellular junctions such as adherens and gap junctions, transcription factors, nuclear receptors, proteases/antiproteases and proteins relevant for lipid metabolism (Table 2). Studies such as these will continue to inform us about the complex network required to form both skin barrier structures (SC and TJ) under both homeostatic and inflammatory conditions.

Table 2. Overview of transgenic mouse models characterized by epidermal barrier defects and atopy.

Mouse model	Barrier-related defects	BARRIER effects	ATOPY/ Th2 immune response	Reference (PMID)
Cornified envelope/SC
SPINK5R820X/R820X	Absence of LEKTI; ↑ FLG monomers	Detachment of SC; dehydration	ND; ↑ TSLP; Early death	15590704
LOR−/−	Absence of L-granule	≈ percutaneous dye penetration (outside-in barrier) after E17.5; ≈ TEWL.	ND	11038185
Intercellular junctions
CLDN1−/−	↓ CLDN1	↑ TEWL; ↑ biotin permeability (inside-out barrier)	ND	11889141
(k14- cre)EcadherinFl	Abnormal TJ; ↓ CLDN1	↑ TEWL; ≈ percutaneous dye penetration (outside-in barrier); ↑ biotin permeability (inside-out barrier)	ND	15775979
Inv-CLDN6-tg	↑ CLDN6; ↓ CLDN1 and abnormal differentiation	ND, Lethal	ND	17914196
Inv-DSG3-tg	Thin SC with a compact lamellar pattern; Loss of corneocyte adhesion	↑ TEWL;	ND	11309406
Inv-Cx26-tg	↑ gap junctions in SB	↑ TEWL; ↑ percutaneous dye penetration (outside-in barrier)	ND	16628254
Transcription Factors
(k14-cre)IKK1EKO	↓ CLDN23; ↓ OCLD; ↑ DSG3; Lipids abnormalities in SC	↑ TEWL; ↑ percutaneous dye penetration (outside-in barrier)	ND	17351639
p63 null	↓ CLDN1; SC abnormality	↑ TEWL	ND	18648642
RARα mutant	Absence of multilamellar structure	↑ TEWL	ND	7867929
Spontaneous dermatitis
NC/Nga (ARC)	Unknown; spontaneous dermatitis at week 8	↑ TEWL	After EC	11167677
Flaky tail (ft/ma)	Lack FLG; unknown matted phenotype	≈ basal TEWL	After EC	11121144
Flaky tail (ft/ft)	Lack FLG	≈ basal TEWL	After EC	19349982
Miscellaneous
Epidermal IL4-tg	Spontaneous dermatitis	ND	↑ IgE; asthma	11676841
APOC1+/+	Lipid abnormalities	↑ TEWL	↑ IgE	18049452
ADAM10−/−	Reduction of SS	↑ TEWL	↑ TSLP	21205794

ARC, air-regulated conditions; ADAM, A Disintegrin And Metalloprotease; APOC1, apolipoprotein C1; CLDN, claudin; Cx26, connexin26; DSG, desmoglin; EC: epicutaneous allergen sensitization; FLG, filaggrin; IKK1, I-kappa-B kinase 1; Inv, involucrin; ND: not determined; LEKTI, lymphoepithelial Kazal-type-related inhibitor; LOR, loricrin; OCLD, occludin; RAR, retinoic acid receptor; SB, stratum basale; SC, stratum corneum; SS, spinous layers; TEWL, trans epidermal water loss; tg, transgenic; TJ, Tight junction; TSLP, thymic stromal lymphopoietin; ≈, no difference.

How epidermal barrier defects impact the adaptive immune response to an antigen has been studied by a number of groups (Herrick et al., 2000; Kondo et al., 1998; Oyoshi et al., 2009; Spergel et al., 1998; Strid et al., 2006; Strid et al., 2004; Wang et al., 1996). These studies have demonstrated that epicutaneous sensitization with a protein allergen elicits a local and systemic Th2-predominant response, as noted by increases of IL4, IL5 and antigen-specific IgE and IgG1, but low or absent induction of IFNγ and IgG2a, in the skin and draining lymph nodes in both C57BL6 and BALB/c mouse strains which are at baseline thought to be inherently Th1 and Th2 prone, respectively (Herrick et al., 2003; Kondo et al., 1998; Spergel et al., 1998). Importantly, in all studies, the Th2-biased immune response was achieved only after a number of barrier disrupting methods were employed including shaving, repeated tape-stripping and prolonged (≥ 4 days) and often repeated patch application(s) with an occlusive dressing that would result in maceration and thereby lead to barrier disruption. At least one group has shown that the magnitude of the Th2 response, as measured by IL-4 expression in tissue and draining lymph nodes (Kondo et al., 1998), was enhanced by the degree of skin barrier disruption. Interestingly, IFNγ levels (e.g. Th1 response) did not change with variation in barrier disruption (Kondo et al., 1998). No studies have evaluated IL17 levels as a marker of a Th17 response under similar experimental conditions of graded barrier disruption.

Tissue neutrophilia and the local expression of IL17, both markers of a Th17 adaptive immune response, have been observed in IL4, IL13 and dual knockout mice or in mice that are deficient in the transcription factor STAT6, which is critical for the differentiation of Th2 effector cells (He et al., 2007; Herrick et al., 2003). This work combined with in vitro studies suggesting that Th2 cytokines may reduce expression and biological functions of Th17 cytokines (Eyerich et al., 2009; Nograles et al., 2010) suggests that these two Th phenotypes counteract each other. This is in keeping with findings in humans where there is a remarkable paucity of tissue neutrophils in AD skin biopsies, despite modest expression of prototypic Th17 cytokines, suggesting that Th2 inflammation dampens the functional Th17 response (Guttman-Yassky et al., 2011a, b; Nograles et al., 2009; Suarez-Farinas et al., 2011). This is in sharp contrast with psoriasis, a prototypic Th17/Th1 skin disease, characterized by tissue neutrophilia and relative paucity of Th2 cytokines, for which the antigens and conditions under which sensitization occurs remain a mystery (Guttman-Yassky et al., 2008; Koga et al., 2008).

These murine models of epicutaneous allergen challenge induce skin lesions that are AD-like (Spergel et al., 1998). The sensitized mice have increased scratching behavior, thickened epidermis, inflammatory infiltration of CD4+ T cells and eosinophils, expression of Th2 cytokines, with modest increases in IFNγ, Th2 chemokines such as eotaxin and the Th2 adjuvant, TSLP as well as allergen-specific IgE (Herrick et al., 2003; Kondo et al., 1998; Spergel et al., 1998). Additionally, these mice developed airway hyper-responsiveness to methacholine and eosinophilia after a single dose of inhaled OVA. This demonstrates that antigen education epicutaneously on barrier-disrupted skin is sufficient to elicit systemic Th2 allergic inflammation in a distant organ such as the lower airways and esophagus following relevant challenge (e.g. inhaled vs oral, respectively)(Akei et al., 2006; Spergel et al., 1998; Strid et al., 2005). This may help explain the so-called “allergic march” where AD is the earliest atopic disorder to present followed, not infrequently, by the development of other atopic diseases such as allergic rhinitis, asthma, and food allergy. Additionally, it highlights the importance of cutaneous allergen sensitization even if the ultimate allergen elicitation occurs in a distant organ.

To demonstrate the importance of the route of immunization (e.g. sensitization phase), Strid et al. evaluated whether the immunologic response differed if the allergen (e.g. peanut protein) was applied on disrupted epidermis (24 hours after tape stripping) or injected subcutaneously with complete Freund’s adjuvant (CFA) (Strid et al., 2004). They found that the epicutaneous route generated a Th2 immune response in draining lymph nodes and spleen (higher IL4 and antigen-specific IgE and lower IFNγ, IL10 and IgG2a) in contrast with the subcutaneous route, which induced a Th1 immune response (high IFNγ and IgG2a and low IL4 and antigen-specific IgE). This is not too surprising since CFA is a Th1 and possibly Th17 adjuvant commonly used in many vaccines. A more interesting comparison would be to evaluate the adaptive immune response to subcutaneous administration of the antigen with a Th2 adjuvant. In subsequent work, this group demonstrated that immunization with an antigen applied on barrier-disrupted epidermis was able to switch an established Th1 response to a Th2 response (Strid et al., 2006). Interestingly, the prevention of the Th1 immune response was specific to the antigen applied on the skin, as exposure to a different antigen was still able to induce a Th1 response in the same animal. Collectively, this work suggests that epidermal barrier disruption can be a “real-life” Th2 adjuvant.

A reductionist approach to barrier disruption is necessary to delineate the importance of specific barrier defects and their relevance to the immunologic abnormalities observed in AD. Fallon and colleagues have done just that by identifying the FLG mutation present in the spontaneous recessive mouse mutant referred to as flaky tale (ft) mouse (Fallon et al., 2009). To eliminate the unknown effect of the matted mutation, Fallon et al. backcrossed the flaky tail mouse onto the C57BL/6 strain (Fallon et al., 2009; Presland et al., 2000). They (Fallon et al., 2009) found that C57BL/6 ft/ft homozygotes generated allergen-specific IgE (as well as IgG2a and IgG1) and a mixed Th1 (IFNγ), Th2 (IL5, IL4 and IL13), and Th17 (IL17) response after epicutaneous allergen challenge as compared to heterozygotes (ft/wt) or wild type (wt/wt) mice. In this study, the epicutaneous sensitization was elicited with relatively little barrier disruption (shaving and allergen application for 5 days × 3 cycles) and with this protocol the C57BL/6 ft/wt and wt/wt had no inflammation compared to saline challenged sites. In other words, this sensitization protocol was not sufficient to induce a Th2 response in the C57BL/6 mouse strain, which Herrick et al., have shown is possible with sufficient barrier disruption (Herrick et al., 2000). Oyoshi et al using the naturally occurring flaky tail mouse (with the matted mutation) and a similar epicutaneous allergen sensitization protocol found an enhanced Th17 (IL17 and IL23) with similar Th2 (IL-4 and IL-13) and Th1 (IFNγ) response compared to wild type strain controls (BALB/c and C56BL/6) (Oyoshi et al., 2009). Collectively, this work highlights the permissive nature of the FLG knockout mouse for epicutaneous sensitization and demonstrates the promiscuous nature of the adaptive immune response, which does not favor a specific T helper profile. In the Fallon study, although baseline TEWL was not different among the three groups (ft/ft, ft/wt and wt/wt) there was a substantial increase in TEWL observed only in the ft/ft after allergen sensitization that was commensurate with the cellular infiltrate, suggesting that it was the result of an inside-out mechanism. Lastly, although human studies have shown that FLG mutations predispose AD subjects to the development of asthma (van den Oord and Sheikh, 2009), epicutaneous sensitization in the ft/ft mouse model did not evoke an effector response in the airway as has been seen in other barrier-disrupted, epicutaneous allergen challenge models ((Herrick et al.; Spergel et al.). It is worth mentioning that the allergen sensitization and AD-like phenotype was only observed in ft/ft and not in heterozygous (ft/wt) animals, while the majority of FLG-associated AD cases have a single FLG mutated allele (e.g. are heterozygotes). The authors suggested that the lack of lung sensitization in this model could reflect the use of the C57BL/6 strain, which is less susceptible to allergen-induced Th2 inflammation as compared to BALB/c. Alternatively, the lack of lung reactivity may also suggest that additional skin barrier defects need to be present at the site of allergen sensitization, and these may prove to be more critical determinants of the Th2 response characteristically observed in AD.

One caveat to consider when comparing antigen challenge models is that the purity of the antigen preparation is a critical determinant of the immunologic outcome. For example, many allergen preparations are contaminated with ligands that can trigger innate immune receptors (e.g. LPS or peptidoglycans) and therefore would have an adjuvant effect. Even more confusing is the growing literature suggesting that many allergens themselves activate innate immune receptors (e.g. house dust mite activating dectin-1 and TLR4; ragweed activating TLR4; nickel activating TLR4; cockroach activating TLR2) (Li et al., 2011; Page et al., 2008; Schmidt et al., 2010; Trompette et al., 2009). Several other variables that may be equally important include the dose of antigen used, age and sex of mice, dressing used to apply antigen to the skin surface, anatomic site where sensitization occurs, “cleanliness” of the vivarium (e.g. determinants of microbial skin flora) and type and magnitude of barrier disruption employed (Kondo et al., 1998; Oyoshi et al., 2009; Spergel et al., 1998; Strid et al., 2004).

In summary, murine studies strongly suggest that skin barrier disruption either by physical means or by targeted gene deletion, as in the FLG −/− mouse, significantly enhances immunologic responses to the epicutaneous application of allergens. Similarly, the mode of barrier disruption may determine the flavor of the immune response with some favoring Th2 and others favoring a Th17 response. A better understanding of the epithelial factors that regulate effector T cell maturation and polarization at the skin surface is crucial for the development of preventive as well as therapeutic strategies in antigen-driven skin disorders such as AD, PS and allergic contact dermatitis.

IV. Potential mechanisms for a Th2 response following epidermal barrier disruption

The first step in the development of any adaptive immune response is the activation of innate immune receptors or pattern-recognition receptors (PRRs) by highly conserved pathogen-associated molecular patterns (PAMPs) common to many classes of pathogens or danger-associated molecular patterns (DAMPs) which are cellular components released in the context of cell damage/necrosis (Janeway and Medzhitov, 2002). PRR activation results in the production of specific mediators (cytokines, chemokines, and antimicrobial peptides) as well as the activation and recruitment of immune cells (immature DC, natural killer cells, and neutrophils)(De Benedetto et al., 2009). Th2 responses are commonly elicited by extracellular parasites or environmental allergens (Finkelman and Urban, 1992; Paul and Zhu, 2010) and are mediated in part by the actions of a number of Th2 promoting cytokines including thymic stromal lymphopoietin [TSLP], IL33 and IL25 that are produced by tissue resident cells. By contrast, Th17 promoting cytokines include IL-1β, IL-6, IL-23 and TGFβ (Hammerich et al., 2011; Wilson et al., 2007). The specific innate immune pathways that mediate the effector T cell phenotype are not fully characterized (Paul and Zhu, 2010).

Immune mediators released after epidermal barrier disruption

In 1994, Nickoloff and Naidu (Nickoloff and Naidu, 1994) were the first to demonstrate the epidermal expression of several inflammatory cytokines (TNFα, CXCL8/IL8, IL10, IFNγ and TGFγ) and the Th17 promoting cytokine, TGFβ following tape-stripping of normal human skin. More recently, several groups have extended and confirmed these findings, showing enhanced epidermal mRNA expression of Th2 promoting cytokines, TSLP and IL33 as well as TNFα, heat shock protein (Hsp) 90, Hsp70, and CXCL8 after tape stripping performed on healthy subjects or in Notch deficient mice who have a chronic skin barrier defect (Angelova-Fischer et al., 2010; Briot et al., 2009; Demehri et al., 2008; Dickel et al., 2010). Importantly, this work introduced the notion that endogenous DAMPs or alarmins, such as Hsp70, Hsp90 and IL33, could activate innate receptors following epidermal injury.

TSLP is an IL7-like cytokine, that was originally characterized for its role in T and B cell development, but more recent literature strongly implicates its role in Th2 cell differentiation and in particular, in the pathogenesis of allergic inflammation (Ziegler and Artis, 2010). In 2002, Soumelis at al. (Soumelis et al., 2002) demonstrated that TSLP was highly expressed in the epidermis from AD subjects and that TSLP-activated DCs produced Th2 attracting chemokines (TARC and MDC) and primed naïve T cells to differentiate into Th2 cells. Importantly, TSLP was not detected in the skin from nickel-induced allergic contact dermatitis or systemic lupus erythematosus (Soumelis et al., 2002). Epithelium that is deficient in LEKT1, the protease inhibitor mutated in Netherton’s Syndrome, express more TSLP which is mediated by the enhanced activity of KLK5 in a proteinase-activated receptor-2 (PAR-2)-dependent fashion (Briot et al., 2009; Briot et al., 2010). This finding links a genetic barrier defect (e.g. mutation in SPINK5) that leads to overactivity of serine proteases to TSLP expression and has important implications for the pathogenesis of AD. In vitro studies have demonstrated that a number of proteases can act through PAR-2 and induce TSLP expression from keratinocytes or airway epithelial cells (Kouzaki et al., 2009; Zhang et al., 2009). Interestingly, a number of allergens including house dust mite, cockroach, fungi, and several pollens, contain proteases that may also trigger epithelial production of TSLP through a similar mechanism (Takai and Ikeda, 2011). This may also be true for Staphylococcus aureus which produces extracellular proteases and chronically colonizes the skin surface of most AD patients (Takai and Ikeda, 2011). Proteases may also disrupt epithelial TJs either by direct actions on TJ proteins or by activation of PAR-2 (Runswick et al., 2007; Tai et al., 2006; Takai and Ikeda, 2011; Wan et al., 1999). In summary, protease-mediated barrier disruption might have dual actions on the epithelium. They may induce the production of the Th2-promoting cytokine, TSLP and also facilitate allergen uptake by LC by disrupting TJs. It is intriguing to speculate that the more immunodominant allergens (e.g. house dust mite allergens) might be the ones that can both disrupt skin barriers and directly or indirectly act as Th2 adjuvants.

Epithelial TSLP is also induced through a TLR3 and 5 mediated mechanism in response to microbial products (Kinoshita et al., 2009; Le et al., 2010; Ma et al., 2009) as well as in response to ragweed allergen which is a TLR4 mediated event (Li et al., 2011). In a human skin injury model, the epidermal TSLP immunostaining was observed primarily at later time points (e.g. 48 hr) after the skin barrier perturbation (Angelova-Fischer et al., 2010). This delayed expression of TSLP suggests that TSLP may not be as critical in the sensitization phase as in the elicitation phase of an AD lesion. Therefore, Th2 adjuvants that are induced rapidly would be more likely candidates to initiate a Th2 immune response in the context of barrier disruption.

In contrast, IL33, a novel member of the IL1 family, is rapidly released in response to tissue injury and a more likely candidate to initiate Th2 polarization during the elicitation phase. IL33 was identified in 2005 using a computational database approach to search for the ligand of the orphan receptor ST2, a member of the Toll-like/IL-1R superfamily (Schmitz et al., 2005). It is recognized as an alarmin or DAMP because it is released during epithelial cell death, is associated with infection or tissue injury and is induced by microbial ligands through a TLR-mediated pathway (Moussion et al., 2008). Strong evidence demonstrates that IL33 plays an important role in allergic diseases (Schmitz et al., 2005). Firstly, IL33 is markedly elevated in the serum of patients during anaphylactic shock (Pushparaj et al., 2009), in asthmatic subjects and in the skin of subjects with atopic dermatitis (Oboki et al., 2011; Pushparaj et al., 2009). Secondly, IL33 plays an important role in the early phase of the Th2 immune response to intestinal helminths, acting as a bridge between innate and adaptive immunity (Humphreys et al., 2008). Lastly, mast cells produce IL33 in response to IgE-dependent activation but also amplify the inflammation resulting from both IgE-dependent and independent mast cell and basophil activation (Hsu et al., 2010; Silver et al., 2010).

IL25 (or IL17E), a member of the IL-17 cytokine family, when overexpressed in murine models results in the production of Th2 cytokines IL-4, IL-5, and IL-13, eosinophilia, and elevated serum IgE, similar to what has been observed with TSLP and IL33 (Fort et al., 2001). IL25 is expressed by mouse epithelial cells following allergen stimulation (Angkasekwinai et al., 2007) and in the human skin of AD patients (Wang et al., 2007). Whether barrier disruption/tissue injury induces IL25 production by human keratinocytes has not been studied. But, IL25 might contribute to barrier dysfunction in AD subjects. It was recently shown that IL25 inhibits FLG synthesis by keratinocytes (Hvid et al., 2011) and therefore IL25 is capable of promoting AD both by its effects on the adaptive immune response and on epidermal barrier.

Langerhans cell activation after barrier disruption

The skin is protected by a variety of antigen presenting cells. In the epidermis, LCs form a dense network that covers the whole body. A large number of other DC subpopulations are found in the dermis and under inflammatory conditions migrate into the epidermis. LCs and dermal DCs thus form several layers of immunological defense; however, their specific roles in skin immunity are still controversial (Merad et al., 2008). Skin DCs take up antigens encountered in the skin and migrate to draining lymph nodes, where they present the antigens to T cells. Mechanical or chemical stress on the epidermis has been reported to activate LCs which is thought to be mediated by the secretion of proinflammatory cytokines from keratinocytes (Lessard et al., 1966, 1968; Nickoloff and Naidu, 1994; Nishijima et al., 1997; Streilein et al., 1982; Wood et al., 1992) (Figure 3). Under steady-state conditions, LCs are on standby with their dendrites aimed outwards, positioned close to, but never crossing, the TJ barrier (Kubo et al., 2009), even though they express the key TJ component, claudin-1 (Zimmerli and Hauser, 2007) (Figure 1 & 2). Once they are activated, which can occur in response to mere tape-stripping of the skin surface, LCs extend their dendrites through TJ barriers, become activated and take up antigens from the extra-TJ environment (Kubo et al., 2009). Whether the tape stripping causes transient openings of TJs, the release of epidermal mediators that activate LC or direct dendrite migration to the epidermal surface or all of the above has yet to be determined. Interestingly, the epidermal TJ integrity is maintained by the de novo formation of TJs between keratinocytes and LCs, suggesting that LCs are able to take up foreign antigens from outside TJ barriers without penetration of the antigen through the TJ barriers (Kubo et al., 2009). As TJs are a size-limiting barrier (Tsukita et al., 2001), it is possible that some smaller antigens, haptens or chemicals may penetrate the TJ barrier and through this mechanism encounter LC or DC. Thus, we need to re-evaluate each antigen and allergen individually to determine the limits of their penetration in the context of relevant “barrier disrupted” conditions.

Conclusions

In this review we have highlighted several new concepts that demonstrate the dynamic interaction between the epidermis and the immune system. Many of these concepts were the direct result of early observations made by cutaneous biologists who noted barrier defects in inflammatory skin disorders over 100 years ago. The recognition that keratinocytes are not just important to maintain tissue structure, polarity and to provide a fence-like function but also respond dynamically to environmental perturbations is still a relatively novel concept. In their role as “first-responders” they have clearly emerged as a key cell in the innate immune system. We have highlighted several human skin diseases with mutations that lead to a defective skin barrier and are often associated with markers of a Th2 response (IgE and eosinophilia) and have clinical features characteristic of an atopic disorder, most commonly, AD. It is suspected, but not known whether disorders characterized by a defect in TJ would also be associated with Th2 responses and atopic features. Interestingly, there is a monogenetic disorder called neonatal ichthyosis-sclerosing cholangitis (NISCH) syndrome, which is caused by a mutation in claudin-1, and has a skin phenotype notable for ichthyosis, as the name implies, and pruritus. Unfortunately, we do not know if these subjects have a skin barrier defect and whether they have greater allergen sensitization or atopic diseases. A number of genetically-altered mouse models have highlighted the complex network of proteins that are important for a healthy skin barrier. When this barrier is disrupted by mechanical means or genetically altered as in the case of the complete absence of FLG (ft/ft mice), the epidermis becomes permissive to allergen sensitization. The adaptive immune response that predominates when the skin barrier is physically disrupted is Th2 polarized; whereas, in the ft/ft mouse, a mixture of T helper responses (Th1, Th2 and Th17) is observed. This finding coupled with the evidence that the enhanced epidermal water loss (e.g. ↑ TEWL) characteristic of AD patients with moderate to severe disease is not fully explained by FLG mutations suggests that these mutations may not be the sole barrier defect in this disease. Many other abnormalities have been found in AD subjects that could affect barrier and include defects in lipid metabolism, altered expression of antimicrobial peptides, dysregulated EDC genes, altered activity of endogenous proteases, reduced activity of endogenous antiproteases, TJ defects and simple trauma from the persistent itch-scratch cycle. Some of these defects are likely genetic (FLG, SPINK5) while others may be acquired (e.g. trauma from scratching, in response to the actions of Th2 and Th17 cytokines or on an epigenetic basis).

The mechanism(s) by which epidermal barrier disruption informs and directs the adaptive immune response to antigens is becoming clearer. We have presented the data that implicate several epidermal-derived alarmins (TSLP, IL25 and IL33) in directing a Th2 immune response and would likely be operative when the barrier is compromised. In addition, transient compromise in skin barrier provides the signals that lead to movement of LC dendrites through TJ and ultimately the engulfment of antigens present on the epidermal surface. We have not discussed the various subsets of skin-resident DCs or their relative effects on the adaptive immune system as this remains a highly controversial area. In conclusion, although we have summarized important evidence suggesting that skin barrier impairment appears to be an important determinant of allergen sensitization, we did not conclusively answer our initial question, is “Skin barrier disruption a requisite for allergen sensitization?” Sorting out the relative importance of specific epidermal barrier defects and the pathways by which they affect immune functions will be the charge for cutaneous biologists over the next 75 years.

Abbreviations

AD

atopic dermatitis

CLDN

claudin

CDSN

corneodesmosin

DAMPs

danger associated molecular patterns

DC

dendritic cells

FLG

filaggrin

Hsp

heat shock proteins

IV

ichthyosis vulgaris

KLK

kallikrein

LC

Langerhans cells

LEKTI

lymphoepithelial Kazal-type 5 serine protease inhibitor

NS

Netherton’s syndrome

OVA

ovalbumin

PAMPs

pathogen associated molecular patterns

PRR

pattern recognition receptors

PS

psoriasis

SPINK5

serine peptidase inhibitor Kazal type 5

SC

stratum corneum

SG

stratum granulosum

Th

T helper

TJ

Tight junction

TLR

toll-like receptor

TSLP

thymic stromal lymphopoietin