Most cases of EBS are caused by heterozygous missense mutations in KRT5 and KRT14, encoding keratins mostly expressed in the epidermal basal layer [7, 21]. While the overwhelming proportion of EBS cases in the Western world are inherited in an autosomal dominant (a.d.) fashion, in countries where marriage within the same community/family is common, recessive cases are more prevalent. For example, in the Middle East, approximately 30 % of the cases are caused by bi-allelic recessive mutations in KRT14 [19, 22]. More complex patterns of inheritance may exist, as recently exemplified by a report on digenic inheritance in a child affected by a generalized form of EBS [23]. Mutations in other genes encoding basement membrane zone proteins have also been shown to account for a minority of EBS cases, as will be discussed later.

Most KRT5 and KRT14 mutations have been shown to disrupt the central alpha-helical segment of these keratin molecules, thereby compromising the structure and function of the cell cytoskeleton which becomes unable to accommodate even small amount of mechanical stress. As a consequence of keratin cytoskeleton dysfunction, the basal cell layer is prone to cytolysis when exposed to friction forces. At the ultrastructural level, keratin abnormal function translates into cell vacuolization, keratin filament clumping, and blister formation [15]. Phenotype-genotype analysis revealed that mutations affecting conserved areas at the beginning and end of the central rod segment are usually associated with a more severe phenotype (EBS Dowling-Meara, EBS-DM) than mutations affecting less conserved areas of the keratin molecules (EBS localized, EBS-loc) [24–26], although many exceptions to this rule have been reported [19, 22]. In addition, the nature of the amino acid substitution and not only its location can influence the severity of the disease as well [27, 28].

Most EBS-causing mutations exert a dominant-negative effect, namely, the mutant molecules interfere with the function of the normal keratins encoded by the wild-type allele. This situation has direct implications for the design of genetic therapies for EBS. Indeed, introduction of a wild-type allele is unlikely to benefit EBS patients; instead, effective therapies for EBS should be aimed at eliminating the deleterious keratin molecules encoded by the mutant allele (see below).

It should be noted that EBS phenotype usually evolves over time and generally show improvement as affected individuals get older. Reduced expression of mutant keratin genes and compensatory overexpression of keratins usually weakly expressed in the basal cell layers, such as KRT15 [29], have been invoked to explain this phenomenon. Somatic genetic events may also modify the course of the disease. Revertant mosaicism refers to a situation where a second mutation attenuates or abolishes the deleterious effect of the original mutation in certain areas of the skin. This phenomenon has been reported in a number of patients with EBS and may actually be more common than previously suspected [30, 31]. The phenotypic manifestations of EBS-causing keratin mutations can also be influenced by apparently silent sequence alterations [32]. Genetic background is also important as exemplified by the fact that phenotype-genotype correlations differ across populations and families [19] and by the fact keratin mutations are phenotypically expressed in a strain-dependent fashion in mice [33]. Finally, nongenetic factors may also determine the way a given sequence alteration manifests as illustrated by a transient EBS-like phenotype associated with bexarotene treatment in the presence of an otherwise silent polymorphism in KRT5 [34].

As reviewed so far, a wealth of evidence supports the view that structural fragility of basal keratinocytes accounts for the generation of skin blisters in individuals with EBS, reflecting a loss of epidermal integrity following incipient mechanical trauma [35]. Nevertheless, other mechanisms may also contribute to the disease phenotype. For example, recent data implicate excessive apoptotic activity, possibly induced by keratin clumps, and upregulation of the inflammatory response in the pathogenesis of EBS [36, 37]. In addition, some keratin mutations may affect the cytoskeletal dynamics or interfere with normal keratin post-translational modification [38, 39]. For instance, the presence of cytoplasmic aggregates containing mispolymerized mutant keratin proteins, a defining characteristic of EBS-DM, may contribute to the pathophysiology at the cellular and tissue levels. Conceivably, the failure of misfolded protein response to resolve these aggregates may lead to cellular and tissue stress [40, 41]. Also, gene expression analysis of an EBS-DM cell line has demonstrated upregulation of genes controlling epidermal development, migration, apoptosis, and wound healing as a molecular consequence of a KRT14 mutation [42].

Transgenic mouse models have also revealed several novel functions for keratin proteins in skin epithelia [43] which may all be of relevance to the pathogenesis of EBS, including regulation of cell and tissue growth in the epidermis and hair follicle [10, 44] and promotion of keratinocyte proliferation correlating with the expression of proinflammatory and/or mitogenic cytokines and chemokines in keratinocytes [45]. A recent genome-wide screen has implicated KRT6, KRT16, and KRT17 in a genetic network defining the key interrelationship between barrier function, inflammation, and tumor susceptibility in mouse skin [46]. In keeping with the theme of inflammation and chemokines, Roth et al. [47] have reported an increased density of Langerhans cells in mice null for KRT5 and in individuals with EBS resulting from mutations at the KRT5 locus. This phenomenon correlates with an upregulation in chemokines (e.g., CCL2, CCL19, and CCL20) known to attract Langerhans cell precursors to the skin and otherwise adds to the emerging concept that keratins may act as key immune modulators in the skin [45].

It seems that epidermal fragility is almost exclusively associated with defective function of the conserved central regions of keratin molecules since skin blistering is unusual in disorders resulting from mutations affecting the head or tail domain of keratins. For example, in Dowling-Degos disease (DDD; MIM179850) resulting from mutations in the KRT5 gene region encoding the protein head domain [48], blistering is not observed (although acantholysis is often seen [49]); in contrast, melanosome transport and epithelia growth are abnormal, resulting in reticulate hyperpigmentation of the flexures, comedo-like lesions on the neck, and pitted perioral acneiform scars [50]. In Naegeli-Franceschetti-Jadassohn syndrome (NFJS; 161000) and dermatopathia pigmentosa reticularis (MIM125595), which are caused by mutations affecting the head domain of KRT14 [51], blisters are very unusual; instead the patients display reticulate hyperpigmentation and lack dermatoglyphics [52], due to deranged regulation of apoptotic activity in the basal cell layer [53]. EBS with mottled pigmentation (MIM131960) is characterized by skin blistering associated with reticulate skin pigmentation and is often caused by a recurrent missense mutation (p.P24L) affecting KRT5 head domain [54]. Altogether, these data suggest a role for the non-helical domain of basal keratins in regulating skin pigmentation. No mechanism has been established to account for these pigmentation phenotypes, although there is preliminary evidence pointing to an interaction between KRT5 and components of microtubule-dependent motors, which are involved in melanin pigment transport [48, 55].

A number of subtypes of EBS are not caused by mutations in keratin genes per se but rather result from defective function of molecules associated with keratins. These can be subclassified as basal vs. suprabasal forms of EBS.

Mutations affecting intracellular components of the hemidesmosomes result in basal EBS, underscoring the interdependency between hemidesmosomal junction and epidermal cell cytoskeleton functions. For example, mutations in the genes encoding integrin β[beta]4 and collagen type XVII have been found to cause EBS [56, 57]. A new form of autosomal recessive (AR) EBS was recently described as due to biallelic mutations in the DST gene that encodes the coiled-coil domain of the epithelial isoform of bullous pemphigoid antigen 1, BPAG1-e (also known as BP230) [58, 59]. A number of relatively rare forms of EBS associated with suprabasal blistering are now recognized (EBS-SB). Ectodermal dysplasia with skin fragility (EDSF) (MIM604536) is caused by mutations in PKP1, encoding plakophilin 1, a component of the desmosomal plaque [60–62]. Keratin intermediate filaments binding to the desmosomal plaque, at least in lower suprabasal epidermal cells, are critically dependent upon normal PKF1 function [63], which may explain the common occurrence of blistering in EBS and in EDSF. Lethal acantholytic epidermolysis bullosa (LAEB) is an AR disorder caused by mutations in DSP encoding the desmosomal protein, desmoplakin (DSP). LAEB-causing mutations result in truncated DSP polypeptides lacking the tail domain of the protein, which play a pivotal role in binding to keratin filaments [64]. Histology shows suprabasal clefting and acantholysis throughout the spinous layer, mimicking pemphigus. A second form of lethal congenital EB is caused by a homozygous nonsense mutation in JUP encoding plakoglobin [65]; the aforesaid is a constituent of desmosomes and adherens junctions, and the complete loss of the protein affects expression and distribution of desmosomal components which highlights the fundamental role of plakoglobin in epidermal cohesion.