The epidermis generates an impressive set of critical defensive functions (Elias, 2005; Elias and Choi, 2005) that include the permeability barrier, allowing survival in a potentially desiccating external environment, and an antimicrobial barrier, which simultaneously encourages colonization by nonpathogenic “normal” flora while resisting growth of microbial pathogens (Elias, 2007). The permeability barrier resides in the stratum corneum (SC), where multiple layers of anucleate corneocytes are embedded in an extracellular matrix, enriched in ceramides (Cer), cholesterol, and free fatty acids, arranged as planar lamellar sheets (Elias and Menon, 1991). These lipids, as well as an assortment of hydrolases important for lipid processing and desquamation and at least two key antimicrobial peptides (Aberg et al., 2007, and references cited therein), are delivered to the extracellular matrix through secretion of epidermal lamellar body contents. Whereas lamellar body–derived proteases and their inhibitors orchestrate the orderly digestion of corneodesmosomes, allowing cells to shed invisibly at the skin surface (Brattsand et al., 2005; Stefansson et al., 2008), the lipid-processing enzymes (β-glucocerebrosidase, acidic sphingomyelinase, and secretory phospholipase A₂) generate the Cer and free fatty acids that, along with cholesterol, form the extracellular lamellar membrane (Elias and Menon, 1991). Thus, proteases as well as extracellular enzymes may ultimately be involved in the pathophysiology of eczema and pruritus.

Based primarily on the strong association of inherited abnormalities in filaggrin (FLG) expression, atopic dermatitis (AD), at least in Euro-Americans, is now increasingly considered a primary disorder of SC structure and function (Hudson, 2006; Irvine and McLean, 2006; Palmer et al., 2006; Smith et al., 2006; Weidinger et al., 2006). Thus, AD can be considered a disease of primary barrier failure, characterized by both a defective permeability (Proksch et al., 2006, and references therein) and antimicrobial function (Baker, 2006; Boguniewicz and Leung, 2006). Although both of these abnormalities are well-recognized features of AD, they have been widely assumed to reflect downstream consequences of a primary immunologic abnormality (the historical inside–outside view of AD pathogenesis). We and others have long proposed that the permeability-barrier abnormality in AD is not merely an epiphenomenon but rather the “driver” of disease activity (i.e., the reverse, outside–inside view of disease pathogenesis) (Elias and Feingold, 2001), because (1) the extent of the permeability-barrier abnormality parallels severity of disease phenotype in AD, (2) both clinically uninvolved skin sites and skin cleared of inflammation for as long as 5 years continue to display significant barrier abnormalities, (3) emollient therapy comprises effective ancillary therapy, and (4) specific replacement therapy, which targets the prominent lipid abnormalities that account for the barrier abnormality in AD, not only corrects the permeability-barrier abnormality but also comprises effective anti-inflammatory therapy for AD (Chamlin et al., 2002; Figure 1).

Open in a separate window

Figure 1

Secondary infections can further aggravate barrier abnormality in atopic dermatitis

AD, atopic dermatitis; AMP, adenosine monophosphate; Cer, ceramide; FLG, filaggrin; LEKTI, lymphoepithelial Kazal-type related trypsin inhibitor; PS, psychological stress; RH, relative humidity; Th1, T-helper 1; Th2, T-helper 2 (Modified from Elias et al., in press.)

Still, how loss of FLG (an intracellular protein) provokes a permeability-barrier abnormality is not known. In addition, it is not clear what drives barrier dysfunction in the skin of AD patients without a FLG mutation. Although it has been hypothesized that loss of FLG could cause corneocyte deformation, the barrier abnormality more likely is linked to lack of FLG as a substrate for proteolytic processing and further deimination into polycarboxylic acids, such as pyrrolidine carboxylic acid and trans-urocanic acid. These metabolites normally act as osmolytes (natural moisturizing factors), drawing water into corneocytes, partially accounting for corneocyte hydration. Hence, the most immediate result of FLG deficiency is decreased SC hydration, a well-recognized feature of AD. A steeper water gradient would inexorably accelerate transcutaneous water loss, particularly if a paucity of extracellular lipids results in decreased water-retaining ability. However, either corneocyte flattening or decreased SC hydration can suffice to enhance antigen penetration. We suspect that another mechanism is operative, because another inevitable consequence of decreased downstream production of polycarboxylic acid metabolites is an increase in the SC pH (Krien and Kermici, 2000), sufficient to increase the activities of the multiple serine proteases in SC, which all exhibit neutral-to-alkaline pH optima (Brattsand et al., 2005; Stefansson et al., 2008; Steinhoff et al., 2005). Increased serine protease activity could generate the active forms of the primary cytokines IL-1α and IL-1β from their 33-kDa pro-forms, which are stored in large quantities in the cytosol of corneocytes (Hachem et al., 2002; Nylander-Lundqvist et al., 1996), the first step in the cytokine cascade that we propose initiates inflammation in AD (Figure 2).

Open in a separate window

Figure 2

“Outside–inside–outside” model of AD

Cer, ceramide; FLG, filaggrin; hBD2, human β-defensin-2; Th2, T-helper 2. (From Figure 2. Cer, ceramide; FLG, filaggrin; hBD2, human β-defensin-2; Th2, T-helper 2. (From Steinhoff et al., 2005; modified from Elias et al., in press.)

Another cause of inflammation in AD doubtless includes sustained antigen ingress through a defective barrier, leading to a T-helper 2 (Th2)-dominant infiltrate (Hudson, 2006). Yet FLG mutations alone do not provoke AD, as demonstrated in ichthyosis vulgaris, in which the same single- or double-allele FLG mutations reduce FLG content, but inflammation (i.e., AD) does not inevitably occur (Sandilands et al., 2007). We and others have suggested that certain acquired stressors could elicit disease by aggravating the barrier abnormality. Indeed, a barrier-dependent increase in pH (and serine protease activity) likely accounts for the precipitation of AD following the use of neutral-to-alkaline soaps (Figure 1) (Cork et al., 2006). Likewise, prolonged exposure to reduced environmental humidity, a well-known risk factor for AD, likely accelerates transcutaneous water loss rates across defective SC in AD, further aggravating the barrier abnormality. Finally, psychological stress, which aggravates permeability-barrier function in humans (Altemus et al., 2001), is both a well-known precipitant of AD and a cause of resistance to therapy. Moreover, a dysregulation of proteases, highly expressed by keratinocytes, protease inhibitors, and protease-activated receptors, may be important for neuronal–epidermal communication during barrier dysfunction (Demerjian et al., 2008; Hachem et al., 2006; Steinhoff et al., 2000, 2005).

Despite accumulating evidence in support of a barrier-initiated pathogenesis of AD, recent studies suggest specific mechanisms whereby Th2-generated cytokines could also further aggravate AD. As described in this issue by Kurahashi et al. (2008), exogenous applications of the Th2 cytokine, IL-4, impede permeability-barrier recovery after acute perturbations. The basis for the negative effects of IL-4 could include the prior observation by these authors that exogenous IL-4 inhibits Cer synthesis (Hatano et al., 2005). But IL-4 also inhibits expression of both FLG (Howell et al., 2007) and desmoglein 3 (Kobayashi et al., 2004), which also could further compromise barrier function, completing a potential outside–inside–outside pathogenic loop in AD (Figure 2). Furthermore, nerves may be activated by “barrier stressors” such as UV light, toxic or allergic agents, and microbial agents. Therefore, nerve-derived mediators could also modulate enzyme production, antimicrobial peptide (defensin) generation, pH changes, or SC humidity, thereby influencing keratinocyte function (Paus et al., 2006; Roosterman et al., 2006; Steinhoff et al., 2006).

Together, the converging pathogenic features described above create a strong rationale for the deployment of specific strategies to restore barrier function in AD. When used under nursing supervision, moisturizers have been shown to reduce topical steroid usage (Cork et al., 2003). Of various barrier-repair approaches, a mixture of the three SC lipids in a Cer-dominant formulation (TriCeram, Osmotics Cosmeceuticals) demonstrated improved clinical status, permeability-barrier function, and SC integrity when this technology was substituted for standard moisturizers in children with severe, recalcitrant AD (Chamlin et al., 2002). More recently, a higher-strength, FDA cleared formulation (EpiCeram cream, Ceragenix Pharmaceuticals) demonstrated efficacy that was comparable to that of a mid-potency steroid (fluticasone; Cutivate cream) in an investigator-blinded, multicenter clinical trial of pediatric patients with moderate to severe AD (Sugarman and Parish, in press). These studies, although still preliminary, suggest that correction of the lipid biochemical abnormality that is responsible for the barrier defect in AD could also downregulate the further insideto-outside alterations provoked by IL-4 and could comprise a new or ancillary approach to the therapy of AD.