Epidermolysis bullosa (EB) is a heterogeneous group of congenital disorders characterized by skin blister formation. EB is subdivided into three main subtypes (EB simplex (EBS), junctional EB (JEB) and dystrophic EB (DEB)) and one minor subtype (Kindler syndrome (KS)), according to the level of skin split .
The EBS subtype can be defined as EBS with blisters within epidermal basal keratinocytes or above, and it is distinguished from other subtypes whose levels of blister formation are deeper (JEB and DEB) or variable (KS). Mutations in several genes have been identified as being responsible for EBS phenotypes. The clinical manifestations of EBS vary greatly depending on the causative genes. Some EBS subtypes are mild and tend to improve with age, whereas others are severe and often associated with early demise and/or other organ involvement. This chapter introduces the clinical and histological characteristics and classifications of EBS. Subsequently, each protein that is defective in EBS is discussed, as are animal models of the disease.
2. Overview of epidermolysis bullosa simplex
Mutations in genes encoding keratinocyte components involved in the organization of the cytoskeleton or cell-cell junctions are responsible for EBS. EBS can be subclassified into basal and suprabasal according to the level of skin split [1, 2] (Table 1).
Basal EBS is caused by defects in skin basement membrane (BMZ) proteins. Figure 1 diagrams the skin BMZ. Among the BMZ components, keratin 5/14 and plectin are the main targets in EBS [3, 4]. A few EBS cases have been reported to have mutations in ITGB4 and COL17, which encode β4 integrin and type XVII collagen, respectively [5, 6]. Recently, BPAG1-e was added to the list of basal EBS target proteins [7, 8].
In contrast, suprabasal EBS is associated with abnormalities in desmosomal proteins (Figure 2). So far, plakophilin-1, plakoglobin and desmoplakin are known to be the target proteins of suprabasal EBS [2, 9-11].
|Subtype||Target gene (protein)|
|EBS||Suprabasal EBS||PKP1 (plakophilin-1)|
|Basal EBS||KRT5 (keratin 5)|
|KRT14 (keratin 14)|
|COL17 (type XVII collagen)|
|ITGB4 (β4 integrin)|
Animal models have been used to clarify the function of some proteins and to develop new therapies for human diseases. Animal models of EB were reviewed recently [12, 13]. However, some new animal models have emerged since then [14, 15], and other transgenic mice with abnormalities in desmosomal proteins should be added to the list of EB animal models because of the introduction of the concept of “suprabasal EBS” . Table 2 summarizes animal models of EBS.
|KRT5||Cow||Naturally occurring (a heterozygous missense mutation)||Not mentioned|||
|KRT14||Mouse||Tg (expressing truncated protein)||Neonatal death|||
|KRT14||Mouse||KI (an inducible model)||Not mentioned|||
|PLEC||Mouse||Conditional KO||Neonatal death|||
|PLEC||Mouse||KI (expressing EBS-Ogna mutation)||Normal|||
|DSP||Mouse||Conditional KO||Not mentioned|||
|PKP1||Dog||Naturally occurring (a homozygous splice donor site mutation)||Neonatal death|
(6 of 9 affected dogs)
|ITGB4||Mouse||Partial ablation (expressing ectodomain of β4 integrin)||Neonatal death|||
|ITGB4||Mouse||Conditional KO||Not mentioned|||
|COL17A1||Mouse||KO||Prolonged survival in 20%|
3. Target proteins in basal EBS
3.1. Keratin 5/14
Recent brilliant reviews have addressed keratins and EBS [3, 32]. Here we focus on the history, mutation analysis, animal models and future therapeutics of keratin-associated EBS from the physician’s point of view.
Keratin is one of the most abundant components of the epithelial cytoskeleton . Typically, type I and type II keratins form heteropolymers that function in cells . Keratin 5 (K5) and keratin 14 (K14) are specifically expressed in epidermal basal cells [34, 35] (Figure 1). In the 1980’s, disorganization of those keratins was recognized in the basal keratinocytes of EBS patients [36, 37]. From those findings, it had been hypothesized that EBS patients have mutations in KRT5 or KRT14, which encodes K5 or K14, respectively. In the early 1990’s, transgenic mice overexpressing mutated K14 were reported to have severe skin fragility . Soon after this discovery, two groups of researchers identified EBS cases with heterozygosity for KRT14 missense mutations [38, 39], which were followed by the identification of the first EBS family with a heterozygous KRT5 mutation . Since then, several hundreds of EBS patients have been described as having KRT5 or KRT14 mutations and have been summarized in the Human Intermediate Filament Database (http://www.interfil.org/) .
There are several subtypes of keratin-associated EBS, as described in Table 3 . Classical and common EBS subtypes, in which traits are autosomal-dominantly inherited, are Dowling-Meara type EBS (EBS-DM), non Dowling-Meara type (EBS-gen-non-DM) and localized type (EBS-loc), from the severest to the mildest. Ultrastructurally, basal keratinocytes of EBS-DM are characterized by keratin aggregates . Hot spots of the mutations in KRT5 or KRT14 are located within the helix-boundary motifs of each keratin . A missense mutation in one allele of those regions (which leads to an amino acid alteration) typically exerts a dominant-negative effect on keratin organization. The severity of the clinical manifestations among EBS-DM, EBS-gen-non-DM and EBS-loc is generally determined by the site of the mutations and the difference between the original and the mutated amino acids . However, it is not always easy to predict the phenotype from the underlying mutations and, in some cases, two different amino acid substitutions at the same codon result in different clinical manifestations [43, 44]. As a single amino-acid alteration does not necessarily cause a pathological change, in vitro and in silico systems to validate mutational effects have been proposed where keratin organization is visualized in cells transfected with mutated or wild-type keratins [44, 45].
The pathogenesis of EBS development through keratin mutations has also been demonstrated in animal models (Table 2). Following the discovery of transgenic mice overexpressing mutated K14 described above , Krt5-null and Krt14-null mice were reported to have a skin fragility phenotype [16, 19], although the condition of those mice was different from that of most EBS patients, where altered amino acids yield dominant-negative effects. Instead, those Krt5-null and Krt14-null mice show the phenotype of autosomal recessive EBS (EBS-AR) whose K5 or K14 is null . To reproduce dominant-negative effects of mutated keratins in human EBS (EBS-DM, EBS-gen-non-DM and EBS-loc), inducible knock-in EBS model mice were generated, in which a Krt14 missense mutation equivalent to human EBS mutation was introduced . This inducible EBS model recapitulates the skin fragility seen in human patients with autosomal dominant EBS. Furthermore, there is one naturally occurring bovine with a heterozygous KRT5 mutation . This Friesian-Jersey crossbred bull exhibits the EBS phenotype.
Therapeutic interventions for EBS have been confined to palliative modalities. However, recent innovations in RNA interference have led to therapeutic strategies for dominant-negative disorders including keratin-associated EBS, where aberrant mutated keratin is knocked down while normal keratin synthesis on another allele is left intact . This RNAi strategy is promising and will be further validated in clinical trials.
Plectin is a cross-linking protein between the cytoskeleton and membranous proteins including hemidesmosomal components (Figure 1). Plectin has been known to have many transcript isoforms that differ from each other in N-terminal sequences at the protein level . Among the many transcript isoforms, plectin 1a is the one that is mainly expressed in epidermal keratinocytes . In addition to 5’ transcript complexity, plectin has a rodless splicing variant . There are several EBS subtypes that are caused by plectin deficiencies (Table 4).
In the mid-1990’s, mutations in the gene encoding plectin (PLEC) were discovered in patients with EBS with muscular dystrophy (EBS-MD) [53, 54]. Since then, many PLEC mutations, mostly located in the region encoding the rod domain of plectin, have been reported in EBS-MD patients [4, 47, 55].
In 2005, two groups independently reported a new EBS subtype with PLEC mutations: EBS with pyloric atresia (EBS-PA) [56, 57]. EB with pyloric atresia (PA) had been known in patients with ITGA6 or ITGB4 mutations [58, 59]. However, skin specimens from those patients with integrin mutations show skin-split at the level of the lamina lucida, leading to the diagnosis of junctional EB (JEB). In contrast, EBS-PA cases with PLEC mutations were characterized by skin-split within epidermal basal cells .
The reason PLEC mutations lead to two distinct subtypes of EBS was clarified only recently. The development of monoclonal antibodies against several portions of plectin allowed us to understand the plectin expression patterns that distinguish between EBS-MD and EBS-PA . EBS-MD skin typically shows the expression of rodless plectin without that of full-length plectin, whereas neither rodless nor full-length plectin is present in EBS-PA skin .
The next big question was whether EBS-MD and EBS-PA can occur simultaneously in a single patient or those two distinct EBS subtypes are mutually exclusive. Recently, one case was reported to have the phenotype of both EBS-MD and EBS-PA (EBS-MD-PA) . The patient had truncation mutations at the last exon of PLEC, which resulted in the expression of diminished and shortened full-length and rodless plectin without the intermediate filament binding domain .
Apart from autosomal recessive EBS subtypes associated with PLEC mutations (EBS-MD, EBS-MD and EBS-MD-PA), there is one distinct autosomal dominant EBS with a PLEC mutation: EBS, Ogna (EBS-Og). EBS-Og is caused by a heterogeneous mutation of p. Arg2000Trp and is characterized by mild blister formation without MD or PA phenotype [4, 60]. To date, 5 unrelated families of EBS-Og have been reported to have the same mutation [49, 60].
Animal models of plectin-deficient EBS have been generated (Table 2). Plec-null mice show severe blistering phenotype and neonatal death , although gastrointestinal tracts were not investigated to confirm PA or PA-like lesions. Myofibril integrity is impaired in the skeletal and heart muscle of those mice . Epidermis-specific ablation of plectin also elicits a severe blistering phenotype and early lethality in mice . Furthermore, mice knocked-in with the murine equivalent mutation of EBS-Og show skin fragility due to epidermal-specific proteolysis of mutated plectin .
Dystonin, encoded by DST, has various isoforms in neural, muscle and epithelial tissue. BPAG1-e, also called BP230, is a major skin isoform of dystonin and a component of hemidesmosomes (Figure 1). BPAG1-e is known to be an autoantigen in bullous pemphigoid as well as type XVII collagen (C17) [61-63]. Since COL17, which encodes C17, was identified as a causative gene for non-Herlitz JEB , DST, which encodes BPAG1-e, had also been hypothesized for decades to be a target gene in other EB subtypes. However, it was only recently that mutations in DST were identified in autosomal recessive EBS patients [7, 8]. Those two patients typically had a mild acral blistering phenotype and had truncation mutations in the coiled-coil rod domain of BPAG1-e. Electron microscopy observation revealed loss of the inner plaque of hemidesmosomes in both cases [7, 8]. Dst-null mice show neural degeneration and mild skin fragility upon mechanical stress  (Table 2).
Mutations in COL17 have been known to be responsible for non-Herlitz JEB (nH-JEB), in which the lamina lucida is the location of the skin-split as described above  (Figure 1). However, one case was reported to show a phenotype of EBS with COL17 mutations . The mutations found in that case caused a loss of intracellular C17 . Furthermore, Col17-null mice were reported to show a reduced number of hypoplastic hemidesmosomal inner and outer attachment plaques with poor keratin filament attachment . These findings suggest that COL17 mutations can cause not only nH-JEB but also EBS, depending on the mutational sites.
α6/β4 integrins are hemidesmosomal components that are encoded by ITGA6/ITGB4, respectively. (Figure 1). Those genes are also target genes in JEB (with or without PA), just as COL17 is a target gene in nH-JEB. There is one autosomal recessive EBS case where the intracellular portion of β4 integrin was deleted .
4. Target proteins in suprabasal EBS
Desmoplakin is a plakin family protein located in desmosome  (Figure 2). Two isoforms (desmoplakins I and II) are generated through alternative splicing . Desmoplakin I is mainly expressed in the heart, whereas desmoplakin II is abundant in the skin . In the early 1990’s, desmoplakin was determined as a major autoantigen in paraneoplastic pemphigus [67, 68]. Mutations in the gene encoding desmoplakin, DSP, have been reported in several genodermatoses, mostly with cardiac manifestations [11, 69]. In 2005, a very severe EB case, referred to as lethal acantholytic epidermolysis bullosa (LAEB), was reported to have a homozygous deletion mutation in DSP . The patient showed severe skin blistering and early demise. There have been only three reports on LAEB with DSP mutations [70-72]. Skin specimens in all the cases revealed acantholytic features in histopathology. From the correlation of clinical manifestations and mutational sites, it seems that complete or almost complete loss of desmoplakin might lead to LAEB . However, at least one full-length desmoplakin (either isoform I or II) may be enough to prevent the development of LAEB .
There are two desmoplakin-associated EBS model animals (Table 2). The fact that Dsp knockout mice show embryonic lethality confirms that desmoplakin is essential in the early development of tissue architecture through embryogenesis . Epidermis-specific ablation of Dsp elicits severe skin defects in newborn mice .
Plakophilin-deficient EBS is listed in the newest classification of EB . This entity has also been called ectodermal dysplasia-skin fragility syndrome (ED-SF). An excellent review on this EBS subtype was published recently . The first case of ED-SF and the mutations in the gene encoding plakophilin-1, PKP1, were reported in 1997 . Since then, many cases of ED-SF with PKP1 mutations have been published. The clinical manifestations of ED-SF include skin fragility, perioral cracking, alopecia and palmoplantar keratoderma .
The desmosomal expression of plakophilin-1 (Figure 2) accounts for skin fragility and histological features of skin specimens characterized by widening of spaces between keratinocytes. However, the phenotype of ectodermal dysplasia may not be explained solely by desmosomal proteins. Recently, plakophilin-1 has been identified as a regulator of protein synthesis and proliferation through a pathway associated with eIF4A1 . It is speculated that the role of plakophilin-1 in translation and proliferation is involved in abnormalities in skin appendages of ED-SF patients .
Mice models in which plakophilin-1 is defective have not been reported. However, there is a naturally occurring canine model with PKP1 mutations that recapitulates human ED-SF  (Table 2). This family of Chesapeake Bay retriever dogs typically shows skin fragility and some ectodermal dysplasiac manifestations such as hair loss.
JUP, which encodes plakoglobin, was not listed as a causative gene of EB in the report of the Third International Consensus Meeting on Diagnosis and Classification of EB . It was only recently that a homozygous nonsense mutation of this gene, leading to complete loss of plakoglobin, was revealed to be responsible for one subtype of suprabasal EBS . Lethal congenital EB (LCEB), named by the authors, has manifestations similar to those of LAEB, which is caused by DSP mutations . This similarity is accounted for by the expression pattern of plakoglobin and desmoplakin in desmosomes (Figure 2). This new entity is expected to be included in future classifications of EB .
Many genes are involved in the manifestations of EBS, as described in this chapter. The most common subtype is keratin-associated EBS caused by dominant-negative effects of aberrant mutated protein. RNAi strategies will be used in future clinical trials, although it is not easy to apply such therapies for all patients, because each patient has a different mutation. Tailor-made strategies will be required to correct each EBS mutation.
Other EBS subtypes are generally complicated with organ malfunction. The task of clinicians is to predict the prognosis of each EBS cases based on the causative genes. It is imperative to clarify what organs, other than the skin, will suffer dysfunction in each EBS case.