Primary cicatricial alopecia: Recent advances in understanding and management


Manabu Ohyama, M.D., Ph.D., Department of Dermatology, Keio University School of Medicine, 35 Shinanomachi Shinjuku-ku, Tokyo 160-8582, Japan. Email:


Primary cicatricial alopecias (PCA) are a rare group of disorders, in which the hair follicle is the main target of destructive inflammation resulting in irreversible hair loss with scarring of affected lesions. The most typical clinical manifestation of PCA is the loss of visible follicular ostia. The histopathological hallmark of a fully developed lesion is the replacement of the hair follicle structure by fibrous tissue. PCA could share similar clinical manifestations and eventually lead to “burn-out” alopecia. Some subsets are hardly distinguishable histopathologically and the mechanisms that elicit such a destructive reaction have not been fully elucidated. Thus, the management of PCA represents one of the most challenging clinical problems for dermatologists. The aim of this review is to provide a concise and comprehensive summary of recent advances in PCA management, especially focusing on novel methodologies to aid diagnosis, and updates on our understanding of the etiopathogenesis. Dermoscopy, a new pathological preparation technique and direct immunofluorescence analysis enable more accurate clinicopathological diagnosis of PCA. Microarray analysis may be beneficial to distinguish PCA subtypes. Currently suggested mechanisms underlying PCA include loss of immune protection of stem cells, impaired stem cell self-maintenance, enhanced autoimmunity by pro-inflammatory cytokines and environmental/genetic predispositions. Interestingly, recent data indicates the association between lipid metabolism dysregulation and PCA development, implying an important role of the sebaceous gland dysfunction in the etiopathogenesis. Based on that hypothesis and observations, novel therapeutic approaches have been proposed, including the use of peroxisome proliferator-activated receptor-γ agonist for lichen planopilaris.


Cicatricial alopecias are a group of intractable and uncommon hair loss disorders characterized by permanent hair follicle destruction.1–5 The most typical clinical manifestation of cicatricial alopecia is the loss of visible follicular ostia in a scarring area (Fig. 1a).4,5 The histopathological hallmark of a fully developed lesion is the replacement of the hair follicle structure by fibrous tissue (Fig. 1b).1,5,6 Cicatricial alopecia may result from trauma (burns, radiation, traction), infiltrative processes (sarcoidosis, carcinomas) or infection (dermatophyte).2,5 In those conditions, the hair follicle is a “by-stander” unfortunately involved in more global damage in the scalp; thus, permanent hair loss is a secondary event (secondary cicatricial alopecia).2,5 In contrast, primary cicatricial alopecias (PCA) are a group of disorders, in which the hair follicle is the main target of destructive inflammation resulting in irreversible hair loss.4,5,7–9 PCA include the conditions of varied clinical and pathological features. This, together with inconsistent use of terminology, has hampered a comprehensive definition of clinicopathological correlation,3 which has made the study of PCA pathophysiology difficult. A breakthrough in our understanding of PCA was made when hair follicle stem cells were identified in the bulge area of the hair follicle.10 Because the inflammation in PCA mostly involves the bulge region, it is now widely accepted that the loss of hair follicle stem cells is the main reason for permanent alopecia.9,11,12 Currently, PCA attract major interest from clinicians and also stem cell biologists as a model of organ-specific stem cell depletion.8,9,12 In the present paper, recent advances in understanding and management of PCA are reviewed, especially focusing on current insights into the etiopathogenesis. Detailed descriptions of the clinical and pathological features of individual PCA are out of the scope of this review. For those unfamiliar with each clinical entity, excellent review articles by the experts1,4–7,13 should help understanding.

Figure 1.

 Typical clinical and histopathological manifestation of primary cicatricial alopecia. (a) The loss of visible follicular ostia in a patch of scarring alopecia. (b) The replacement of the hair follicle structure by fibrous tissue (arrowheads) (hematoxylin–eosin, original magnification ×40).

Working Classification of PCA

A decade ago, the North American Hair Research Society (NAHRS) sponsored a workshop and developed a working classification of PCA, which was mainly based on the most representative pathological finding of scalp biopsy samples.2,14 In this classification, PCA were divided into subgroups depending on the predominating inflammatory infiltrates (Fig. 2). Chronic cutaneous lupus erythematosus (CCLE), lichen planopilaris (LPP; Fig. 3a), classic pseudoplade of Brocq (CP), central centrifugal cicatricial alopecia (CCCA), alopecia mucinosa (AM) and keratosis follicularis spinulosa decalvans (KFSD) were categorized as “lymphocytic” PCA. Frontal fibrosing alopecia (FFA) and Graham–Little syndrome (GLS) were considered as LLP variants. The neutrophilic PCA group comprised folliculitis decalvans (FD; Fig. 3b) and dissecting cellulitis/folliculitis (perifolliculitis abscedens et suffodiens) (DC/DF). Acne keloidalis (AK), acne necrotica (AN) and eruptive pustular dermatosis (EPD) were classified as “mixed” cell infiltrate PCA. In addition, non-specific cicatricial alopecia was defined as “idiopathic scarring with inconclusive clinical and histopathological findings”.2 The end stage of various PCA may be included in this category as they are hardly distinguishable clinically and histopathologically.2 There have been debates whether this classification is satisfactory,4 however, it provides a practical and reasonable standard for clinical and basic studies and thus has been widely used.

Figure 2.

 Working classification of primary cicatricial alopecias. AK, acne keloidalis; AM, alopecia mucinosa; AN, acne necrotica; CCCA, central centrifugal cicatricial alopecia; CCLE, chronic cutaneous lupus erythematosus; CP, classic pseudopelade of Brocq; DC/DF, dissecting cellulitis/folliculitis (perifolliculitis abscedens et suffodiens); EPD, eruptive pustular dermatosis; FD, folliculitis decalvans; FFA, frontal fibrosing alopecia; GLS, Graham–Little syndrome; KFSD, keratosis follicularis spinulosa decalvans; LPP, lichen planopilaris.

Figure 3.

 Clinical presentations of lymphocytic and neutrophilic primary cicatricial alopecias. (a) Lichen planopilaris (LPP). (b) Folliculitis decalvans (FD). Note that FD is exudative with crust formation.

Advances in Diagnostic Procedures

Trichoscopy (dermoscopy)

The loss of follicular ostia, which is the most characteristic feature of PCA (Fig. 1a), may not be clinically evident in some cases, but could be clearly visualized under trichoscopy (dry dermoscopy). Indeed, trichoscopy significantly improves the accuracy of the diagnosis of PCA.4 Other PCA-associated signs, such as perifollicular erythema or scale hair tufting,15 are also detectable. Trichoscopy also helps clinicians assessing PCA disease activity. For instance, “follicular red dots”, erythematous polycyclic, concentric structures regularly distributed in and around the follicular ostia, are suggestive of active lupus erythematosus of the scalp.16 Thus, trichoscopy should be routinely performed when PCA are considered as differential diagnoses.

Histopathological examination

Scalp biopsy is required not only to confirm PCA diagnosis but also to determine the predominant infiltrates for classification (Fig. 4).4,5 The histological distinction between lymphocytic and neutrophilic PCA groups is mostly possible. However, a previous study suggested that some PCA cases within the same subgroups were hardly distinguishable histopathologically.3

Figure 4.

 Recommended sectioning for primary cicatricial alopecia biopsy samples. HoVert technique requires less samples compared to the combination of vertical and horizontal sections. DIF, direct immunofluorescence; HE, hematoxylin–eosin.

For an accurate histopathological diagnosis of PCA, biopsy samples should be obtained from active sites and carefully sectioned. Unlike in other skin diseases, the information obtained by vertical sections is limited in hair disorders.17,18 Transverse sections enable both qualitative (e.g. inflammatory change, fibrosis) and quantitative (e.g. hair follicle numbers, size, phase of hair cycle) examination of scalp biopsy samples.6 Potential disadvantages of transverse sectioning technique are limited visualization of gross reaction pattern and change in the dermoepidermal junction.7 Ideally, two 4-mm punch biopsies need to be performed to prepare both transverse and vertical sections (Fig. 4).7,19 However, multiple biopsies could burden the patients with greater medical costs and morbidity.20 Recently, the “HoVert” technique, a novel processing technique which produces transverse (horizontal) and vertical sections from a single biopsy, was described (Fig. 4).20 In this report, the authors adopted this technique for the diagnosis of alopecia cases, including discoid lupus erythematosus (DLE) and LPP, and concluded that the HoVert technique provided more pathological information than either vertical or horizontal sections alone.20 For those unfamiliar with the processing of scalp samples, HoVert preparation may be challenging. However, this new technique should be considered, especially in situations where only a single biopsy is allowed.

Primary cicatricial alopecia sections should be carefully investigated for microorganisms (bacteria and fungus), especially when neutrophilic infiltrates are predominant. Stains other than conventional hematoxylin–eosin, such as elastica Van Gieson, Alcian blue and periodic acid-Schiff may be beneficial.7 Previous studies reported that a distinctive elastic staining pattern helped differential diagnosis of advanced PCA, including pseudoplade of Brocq, LPP and DLE.7,21,22

Direct Immunofluorescence study

Although the systematic histopathological investigation described above provides a powerful diagnostic tool for PCA, inconclusive cases may still be encountered.23 Of note, distinction between LPP and CCLE is sometimes extremely difficult.6,13 Usefulness of direct immunofluorescence (DIF) studies for those tackling cases has been reported.23 The most characteristic DIF finding of CCLE is granular deposits of immunoglobulin (Ig) and C3 (Fig. 5) at the dermoepidermal junction, while that of LPP is globular deposits of IgM adjacent to the hair follicles or at the dermoepidermal junction.23 DIF study is valuable in the diagnosis of PCA especially in those with LPP or CCLE as differential diagnoses.

Figure 5.

 Direct immunofluorescence study is recommended for primary cicatricial alopecias. (a) Clinical presentation of chronic cutaneous lupus erythematosus. (b–d) Immunoglobulin (Ig) and complement (C) deposition visualized by immunofluorescence (original magnification ×200).

Microarray analysis

Microarray analysis of global gene expression profile may mark a new era in the diagnosis of PCA. There has been a debate whether LPP and pseudoplade of Brocq are distinct diseases or different presentation of the same pathogenic conditions, because they can share similar clinical and pathological features.24 Microarray technology allows the elucidation of a broad range of molecular aspects of the disease.24 In the report by Yu et al., microarrays were generated from total RNA isolated from active lesions of LPP and pseudoplade of Brocq and compared (Fig. 6).24 The analysis revealed that global gene expression profiles in LPP and pseudoplade of Brocq based on comparative intra-control scalp samples are different from each other with differential expression of specific genes, suggesting that the two conditions are biologically distinct.24 This finding demonstrated the usefulness of microarrays for the dissection of molecular mechanisms of PCA and implied the potentiality of global gene expression profiling as a diagnostic tool for clinically or pathologically indistinguishable PCA. Furthermore, microarray comparisons among different subtypes of PCA could enable the identification of definitive molecular markers of each condition, which are currently unknown.9

Figure 6.

 A schematic explanation of microarray comparison of lichen planopilaris and pseudopelade of Brocq by Yu et al.24. Total RNA was isolated from representative lesions of two primary cicatricial alopecia subtypes and used for microarray generation. Global gene expression analysis suggested that those conditions are biologically distinct with differential gene expression pattern (hematoxylin–eosin, original magnification ×100).

New Insights into the Etiopathogenesis of PCA

Involvement of hair follicle stem cells in PCA

Hair follicles regenerate themselves through the hair cycle, suggesting the presence of organ-specific stem cells.25,26 The landmark study by Cotsarelis et al. demonstrated that hair follicle epithelial stem cells reside in the bulge region, a contagious portion of the outer root sheath where the arrector pili muscle inserts.10 Series of lineage tracking experiments demonstrated that the bulge stem cells regenerate hair follicles in homeostasis.27–33 Indeed, permanent hair loss was observed in genetically engineered mice that were designed to specifically deplete bulge stem cells.32,33 Based on the observation that the bulge area, which is marked by the biomarkers such as keratin-15 and -19 and CD200,34–36 is preferentially destructed and replaced by fibrotic tissue in PCA (Fig. 7), the loss of bulge stem cells is considered to be a main reason for permanent hair loss in those conditions.9,11,12 Recent studies reported cell populations marked by MTS24,37Lrig1,38 Nestin,39Lgr540 and Lgr641 are also capable of reconstituting hair follicles, suggesting that those cells are endowed with some stem cell characteristics (Fig. 7). The inflammation and subsequent fibrosis in PCA also affect most of those populations (Fig. 7), except Lgr5 expressing cells. Thus, irreversible damage to hair follicle stem cells with resultant permanent hair loss still represents the most characteristic pathophysiology in PCA. However, it should be noted that the loss of stem cells alone could not explain other PCA manifestations, such as atrophy or follicular plugging.9 Some additional factors should contribute to those clinical phenotypes.

Figure 7.

 Hair follicle stem cell populations (both established and putative) and anatomical levels of inflammatory change in primary cicatricial alopecia. Most populations are involved in destructive immune response.

Lipid metabolism dysregulation in PCA pathogenesis

The Asebia mouse, the most well-studied animal model for PCA,12 has a spontaneous mutation resulting in a defect of stearoyl-coenzyme A desaturase.42 The lack of this enzyme causes abnormal fatty acid composition in the sebaceous gland leading to its atrophy and defective secretion. This prohibits normal inner root sheath desquamation, enforces downward hair shaft growth penetrating into the bulb and elicits inflammatory response that eventually destroy the hair follicle structure (Fig. 8).43 Similar phenotypes have been described in other rodent models with respective mutations, including Defolliculated.9,44,45 These observations support the idea that inflammation in scarring alopecia is a secondary event resulting from a primary defect in the pilosebaceous unit.46 Consistent with this, Al-Zaid et al. reported that sebaceous gland loss was a common and early finding in PCA.47

Figure 8.

 Possible mechanisms underlying primary cicatricial alopecias. HF, hair follicle; PPAR-γ, peroxisome proliferator-activated receptor-γ.

Recently, Karnik et al. reported an intriguing data that implies a link between lipid metabolism dysregulation and hair follicle stem cell destruction.48 Microarray analysis identified downregulation of peroxisome proliferator-activated receptor-γ (PPAR-γ) signaling in LPP, suggesting a central role of defective lipid metabolism and peroxisome processing in LPP pathogenesis. Indeed, bulge stem cell-specific depletion of PPAR-γ caused scarring alopecia with focal inflammation and lipid deposition in mice,48 suggesting that PPAR-γ signaling is crucial for hair follicle stem cell maintenance (Fig. 8).

Impaired self-maintenance of hair follicle stem cells cause hair loss

Permanent hair loss can also be observed in “biphasic” alopecias (i.e. androgenetic alopecia, traction alopecia in which non-scarring alopecia is observed initially but scarring alopecia develops later).13 When compared with classic PCA, direct destruction of bulge stem cells by inflammation is less clear or hardly detectable in these conditions.9 Tanimura et al. reported that loss of interaction between Col17a1 and hair follicle stem cells impairs the self-renewal capacity of stem cells causing permanent hair loss in mice,49 demonstrating that the change in microenvironment, including decreased extracellular matrix expression, alone could cause scarring alopecia. Thus, it is possible that impaired self-maintenance or loss of self-regenerative potential due to environmental changes may also be responsible for PCA development (Fig. 8).

PCA and neurogenic inflammation

Harries and Paus proposed the potential involvement of neurogenic inflammation in PCA pathogenesis.9 Psycho-emotional stress upregulates nerve growth factor and substance P, an inducer of neurogenic inflammation via mast cell degranulation in the skin and affects hair growth and cycle in mice.50 Substance P positive nerve fibers are dense in the bulge area51 where stress-induced perifollicular inflammation and apoptosis are predominantly observed.52 In addition, given that substance P is a fibroblast growth factor,53 it could also promote scar formation in PCA (Fig. 8).9

Epithelial–mesenchymal transition (EMT) may contribute to fibrosis in FFA

Previous studies suggested a possible contribution of EMT in renal, liver and pulmonary fibrosis.54 Of note, lineage-tracking experiments clearly demonstrated epithelial cell to myofibroblast conversion, suggesting that EMT greatly contributed to the formation of the collagen network in renal fibrosis.55 Recently, Nakamura and Tokura demonstrated that an EMT marker, snail 1, was expressed in the dermal fibroblasts of FFA patients. The observation suggests a possible role of EMT conversion of hair follicle epithelial cells in the pathogenesis of PCA (Fig. 8).

Increased apoptosis in CCLE and LPP

In PCA, apoptotic keratinocytes are frequently observed in hair follicles, implying that apoptosis may play a role in PCA pathogenesis (Fig. 8).8,9 In line with this, upregulation of p53 and Fas in keratinocytes, together with increase in Fas ligand-positive infiltrating cells, were observed in CCLE.56–59 In addition, global gene expression profiling of LPP lesions elucidated upregulation of genes involved in apoptosis.48

Environmental and genetic factors of PCA

A recent large-scale survey failed to demonstrate an obvious association between central hair loss and the use of a hot comb or relaxer in an African-American woman,60 however, habitual traumatic hair care has been implicated in CCCA pathogenesis.61 Scalp trauma has also been considered to play roles in the pathogenesis of other PCA subtypes, including AK, FD and EPD.9 Of note, a possible link with EPD has been repeatedly reported.62–66 Although evidence is still not sufficient to support a definitive conclusion, the possibility still remains that trauma is a trigger of some forms of PCA.

Hypersensitivity reaction to Staphylococcus aureus infection has been implicated in the pathogenesis of neutrophilic PCA, especially FD.9,67 In theory, superficial S. aureus infection alone is not likely to destruct bulge stem cells and cause permanent hair loss.68 In addition, unlike lymphocytes, neutrophils are not antigen-specific and short-lived.12 Accordingly, without continuous stimuli, chronic inflammation observed in neutrophilic PCA could not be sustained. Thus, S. aureus infection could induce neutrophilic PCA only when it is coupled with some anomaly in host defense or additional events including secondary infection.

Drugs and vaccination may induce PCA.9 Acne keloidalis may be induced by cyclosporine.69,70 Association between imatinib and follicular mucinosis was also described.71 In addition, Graham Little–Piccardi–Lasseur syndrome following hepatitis B virus vaccination has been reported.72

Several reports of the familial PCA cases suggested the existence of genetic factors.73–76 X-linked keratosis follicularis spinulosa decalvans represents genetic scarring alopecias, which is caused by a missense mutation in the MBTSP2 gene.77 Further accumulation of familial PCA pedigrees could enable the identification of gene mutations that predispose affected individuals to permanent hair loss.

Impaired immunological stem cell protection and PCA

The loss of immunological protection of bulge stem cells is an attractive hypothesis for preferential involvement of the bulge area in PCA pathology.8,12 Two major possible mechanisms that defend stem cells from unwanted immunological insults are “immune privilege” and “non-danger” signal (Fig. 8).8,12,78

The term “immune privilege” describes intrinsic machineries to avoid unwanted immune responses.78,79 It has been reported that the hair follicle bulge possesses several characteristics common to immune privilege sites, including decreased expression of major histocompatibility complex (MHC) class I and β2-microglobulin molecules, reduced number and impaired Langerhans cell function, increased production and secretion of immunosuppressive molecules (indoleamine-2,3- dioxygenase, macrophage migration inhibitory factor, α-melanocyte-stimulating hormone, transforming growth factor and human leukocyte antigen E).12,78,80 Interestingly, Harries et al. demonstrated the upregulation of MHC class I and II and β2-microglobulin in the bulge in PCA affected lesions, suggesting immune privilege collapse (Fig. 8).78 Although this observation seems to provide a comprehensive explanation for inflammatory destruction of bulge in PCA, an exact mechanism that elicits an auto-inflammatory response has not been elucidated.8 In addition, the possibility that those findings were secondary to inflammation needs to be excluded.

A recent study identified CD200 as a cell surface marker of human hair follicle bulge cells.36 CD200 is a type 1 transmembrane glycoprotein that transmits an immunosuppressive signal through the CD200 receptor (CD200R).81,82 Constitutive expression of CD200 in bulge cells is thought to be a “non-danger” signal,83 which maintains immunocytes in a quiescent state via CD200–CD200R interaction between the bulge cells and Langerhans cells or other dendritic cells.12 The role of disrupted CD200–CD200R interaction in the development of tissue-specific autoimmunity has been proposed.84 In line with this, when CD200 null skin was grafted onto the wild-type mice, severe cell infiltration attacking hair follicles was provoked and permanent hair loss similar to PCA eventually developed.85 Thus, loss of CD200 expression in the bulge is a possible mechanism that underlies PCA pathogenesis (Fig. 8). Decreased CD200 expression was detected in the bulge area of unusual alopecia areata patients with a bulge involving cell infiltration,86 further supporting this hypothesis.

Cell-mediated autoimmune response and pro-inflammatory cytokines in CCLE

It is widely accepted that PCA are categorized as autoimmune diseases.8,12 In particular, predominance of chemokine receptor 4-expressing activated T cells87 and increase in γδ-T cells88 in CCLE lesions suggested a key role of cell-mediated autoimmunity in the pathogenesis (Fig. 8). High levels of interferon (IFN)-α and subsequent enhancement of T-helper (Th)1 type immune response is a characteristic feature of active cutaneous LE, which is reflected in IFN-α inducible MxA protein expression.89 In scarring CCLE, lesional expression of MxA was closely associated with increase in granzyme B expression in skin-homing T cells expressing cutaneous lymphocyte antigen (CLA), providing evidence that IFN and cytotoxic lymphocytes targeting adnexal structures are responsible for scarring processes in CCLE.90 Other pro-inflammatory cytokines, such as INF-γ, interleukin-2 and tumor necrosis factor-α, have been reported to be potentially involved in the pathogenesis of PCA, including CCLE91 and LPP (Fig. 8).92

New Horizon in the Management of PCA

In PCA, the regrowth of once severely affected hair follicles can hardly be expected. Because the stable bioengineering of human hair follicles, especially for the clinical application, have not been achieved,93 the goals of PCA treatment are currently limited to relieve symptoms (not only hair loss but also itching, pain and discomfort) and to better control/block further spread of lesions. Given the central roles of cytotoxic autoimmune response and host predisposition to S. aureus infection in the pathogenesis of lymphocytic and neutrophilic PCA, respectively, the first-line medications generally selected for the former subtype are immunosuppressive agents and the latter antimicrobials or dapsone.4,94,95 However, currently available remedies are sometimes of limited efficacy. Accordingly, there is a clear demand for novel therapies.

For those patients whose PCA are well controlled and without remaining inflammation, surgical treatments including the removal of alopecic scar or hair transplantation may be performed and beneficial.4,96 Because inflammatory cicatricial alopecias, even in a stable “burn-out” stage, could be reactivated by surgery,97 the criterion for eligibility needs to be strictly defined. For example, a minimum 2-year disease-free period is required to undergo a surgical treatment at the University of British Columbia Hair Clinic.94

New insights into the etiopathogenesis of PCA imply the use of novel medications for the treatment of PCA. As described above, abnormal functioning of PPAR-γ leads to aberrant lipid metabolism in the sebaceous gland and subsequently elicits the inflammation involving the bulge stem cells in LPP.48 Accordingly, pioglitazone hydrochloride, an oral PPAR-γ agonist, was administrated to an intractable male case of LPP.98 Interestingly, a scalp biopsy taken after 6 months of treatment demonstrated significant decrease in infiltrating cells.98 He was administrated pioglitazone hydrochloride for 14 months. One year after the therapy, the patient was symptom free with no sign of hair loss,98 suggesting that PPAR-γ agonists may provide promising remedies for LPP, including its subtypes. Also, loss of CD200 in the bulge area elicits stem cell involving inflammation and leads to permanent hair loss in mice.85 Interestingly, CD200-fc administration to experimental arthritis model mice successfully prevented pro-inflammatory cytokine production without any immunosuppressive events.99 This observation suggested that CD200 agonists might be beneficial in PCA treatment.

Closing Remarks

Despite recent advances in our understanding of the etiopathogenesis and the pathophysiology of PCA, the diagnosis is still challenging and the treatment may be frustrating with limited efficacy in some cases. Once inflammation irreversibly damages the hair follicle stem cell compartment, permanent loss of affected hair follicles is inevitable. To more efficiently prevent this, treatments need to directly target the master regulators in PCA pathogenesis. As microarray analysis of LPP elucidated PPAR-γ dysregulation in LPP and led to the clinical trial of PPAR-γ agonist,48,98 current progress in high-throughput screening methodology may enable the molecular dissection of the etiopathogenesis to specify pivotal targets and, eventually, allow the development of small molecules or biologics for the treatment of other types of PCA. It should be emphasized that the success of such a scenario totally depends on accurate diagnosis, appropriate experimental design and, finally, the enthusiasm of clinician and patients to conquer PCA.


I thank Dr Ophelia Veraitch (Department of Dermatology, Keio University School of Medicine) for her help in the preparation of the manuscript. I am grateful for Dr Masayuki Amagai (Professor and Chairman, Department of Dermatology, Keio University School of Medicine) for his critical reading of the manuscript. Writing this manuscript was made possible in part by the Keio Gakuji Academic Development Funds.