BJD

British Journal of Dermatology

C U TA N E O U S B I O L O G Y

Possible patterns of epidermal melanocyte disappearance in nonsegmental vitiligo: a clinicopathological study* L. Benzekri,1,2 I. Hmamouchi3 and Y. Gauthier4 1

UFR of Dermatology, Mohammed V Souissi University Rabat, Rabat 10100, Morocco Department of Dermatology, Ibn Sina University Hospital, Rabat, Morocco 3 Faculty of Medicine, Laboratory of Biostatistic, Clinical Research and Epidemiology, Mohammed V Souissi University Rabat, Rabat 10100, Morocco 4 French Vitiligo Association, Bordeaux, France 2

Summary Correspondence Laila Benzekri. E-mail: [email protected]

Accepted for publication 28 May 2014

Funding sources None.

Conflicts of interest None declared. *Plain language summary available online. DOI 10.1111/bjd.13160

Background The depigmentation of vitiligo results in a progressive and chronic melanocyte loss with rare melanocytes occasionally remaining in the epidermis or the hair follicle reservoirs. Destruction by immune infiltrates in close contact with melanocytes within microvesicles and/or detachment of melanocytes followed by their transepidermal elimination should be regarded as possible mechanisms of chronic loss of pigment cells. Objectives To assess the frequency of these two histological findings and to establish a direct correlation with clinical features. Methods This was a prospective observational study that took place over 1 year. Each patient received a standardized evaluation that included daylight and Wood’s lamp examinations, pictures, biopsies performed on the marginal area, and histological and immunohistological studies. A second examination to assess the activity of the lesions was performed 1 year after inclusion in the study. Clinical changes associated with microvesicles were compared with those associated with detached melanocytes from the basal layer. Results This study included 50 patients. The histological findings were classified as inflammatory with isolated microvesicles (29 cases), noninflammatory with only detached melanocytes from the basal layer (12 cases) and a combination of coexisting microvesicles and detached melanocytes (six cases). Correlations were obtained between the histological findings and clinical features (aspect and activity of the lesions) and E-cadherin expression. Conclusions Our data suggest the existence of two patterns of melanocyte disappearance in nonsegmental vitiligo.

What’s already known about this topic?

• •

Vitiliginous skin shows a complete loss of melanin pigment and a marked reduction or absence of melanocytes along the basal layer of melanocytes. Histopathological findings in vitiligo skin are discordant according to the type of depigmented lesions.

What does this study add?



© 2014 British Association of Dermatologists

For the first time, two possible patterns of melanocyte disappearance that are correlated with histopathological findings and clinical features are proposed for nonsegmental vitiligo.

British Journal of Dermatology (2015) 172, pp331–336

331

332 Epidermal melanocyte disappearance in nonsegmental vitiligo, L. Benzekri et al.

Histopathological findings in nonsegmental vitiligo (NSV) differ according to the clinical aspect of the borders of the macules and their evolutivity. Hypomelanotic lesions with poorly defined borders present a superficial infiltrate of lymphocytes in the dermis and in the lower part of the epidermis, and they are usually extensive. In contrast, amelanotic lesions with sharply demarcated borders show apparently normal skin except for the absence of pigmentation of the basal layer, and are considered as stable.1 There is long-standing controversy regarding whether melanocytes in vitiligo macules are lost or still present but not detected with common stainings. In the majority of ultrastructural studies, the absence of melanocytes has been confirmed. If the progressive disappearance of melanocytes from the epidermis was generally admitted in the pathogenesis of vitiligo, the early initiating events of this phase would remain unexplained.2 The more interesting data were provided by peripheral biopsies of NSV lesions performed in sites of active depigmentation. Within perilesional skin epidermis, two main features were reported. Firstly, epidermal microvesicles containing T cells juxtaposed with the remaining melanocytes were observed in some cases of spreading and inflammatory vitiligo.3,4 In these cases of patients with active vitiligo, it was reported that interferon gamma levels, which lead to intercellular adhesion molecule (ICAM)-1 expression, were increased.5,6 ICAM-1, which is an important adhesion molecule for T-cell attachment, was expressed on basal melanocytes in active lesions but not in stable lesions.7 An imbalance of CD4+ : CD8+ ratio was reported and we observed an altered frequency of natural regulatory T cells (Tregs) in patients with NSV, which might be involved in the T-cell-mediated pathogenesis of NSV and its progression.8 Moreover, detachment of melanocytes followed by their transepidermal elimination in a melanocytorrhagic process was regarded in some cases as a possible mechanism of chronic loss of pigment cells.9–11 It was demonstrated that keratinocytes and melanocytes interact through adhesion systems that are much less organized than those that hold keratinocytes in place. It is assumed that melanocytes and keratinocytes interact mainly through the homophilic adhesion molecule E-cadherin, which is expressed on melanocytes and keratinocytes.12 A defect of E-cadherin, which is the major adhesion molecule, implicated in the cohesion of the epidermal melanin unit could be responsible for the melanocyte detachment.13,14 In the present clinicopathological study, we have tried to assess in all NSV skin sections the presence of some melanocytes detached from the basal layer and/or isolated microvesicles that were identified in the epidermis. Moreover, we have investigated possible correlations between the patterns of melanocyte disappearance, the global histological aspect including the E-cadherin expression, the clinical type and the activity of individual vitiligo lesions.

Clinical assessment and disease activity Patients were systematically examined both under daylight and under a Wood’s lamp. We proposed to identify two different clinical types of vitiligo separately: amelanotic with sharply demarcated borders and hypomelanotic with poorly defined lesion borders. Clinical pictures were taken at the time of the examination and a skin biopsy was performed concomitantly. All patients were examined 1 year after their first visit. The vitiligo was classified as stable if no new lesions had appeared, and as spreading or active if there had been an increase in the number and/or size of existing vitiligo lesions during this year. Skin biopsies Skin biopsies from the 50 patients were extended to involve the vitiligo lesions, the perilesional margin, and the nearby clinically normal skin. Biopsies of normal skin were provided by 10 control subjects with no apparent skin disease. Histology and immunohistology Skin samples with a 4-mm punch were fixed in formalin saline and embedded in paraffin wax. Eight to ten sections from each patient were stained with haematoxylin–eosin– safran and DOPA reaction. Cryostat sections were stained using monoclonal antibodies for the melanocytes (HMB45; Dako, Les Ulis, France). To assess the adhesivity of melanocytes and keratinocytes, the immunostaining was performed using a monoclonal antibody against E-cadherin (E-cadherin clone NCH38; Dako) and detected by the peroxidase–antiperoxidase technique. For semiquantitative evaluation, the immunostaining was graded as negative (0), weak (1+), moderate (2+) or strong (3+). The slides were assessed under an Olympus microscope (Olympus, Tokyo, Japan). We evaluated the melanocyte disappearance patterns (at least two detached melanocytes from the basal layer or two microvesicles containing melanocytes and lymphocytes for each section) and the adhesivity of epidermal keratinocytes and melanocytes with the E-cadherin monoclonal antibody. Evaluation of biopsy sections was blind and the evaluator was not aware of the clinical type of the lesion. Assessment method We tried to establish a relationship between the melanocyte disappearance patterns and the global histological aspect of the skin sample, including E-cadherin expression and the clinical aspect and activity of the lesions.

Material and methods

Statistical analysis

This study was approved by the Mohammed V Souissi University ethics committee, Rabat, Morocco. Informed consent was obtained from all patients before inclusion.

Values are expressed mean  SD or number (percentage). We compared the demographic and disease characteristics relating to the melanocyte disappearance patterns using Student’s t-test

British Journal of Dermatology (2015) 172, pp331–336

© 2014 British Association of Dermatologists

Epidermal melanocyte disappearance in nonsegmental vitiligo, L. Benzekri et al. 333

and v2-test as appropriate. Statistical analysis was performed with the Windows 130 version of SPSS software (IBM, Armonk, NY, U.S.A.). P-values < 005 were regarded as significant.

A relationship between melanocyte disturbances and vitiligo duration was assessed (Table 1). There was no statistical correlation between the vitiligo duration and the two disappearance patterns with a P-value of 041.

Results

Discussion

Patient details The study included 50 patients with NSV, 33 of whom were women and 17 men. The mean age of the group was 36  15 years and the mean duration of vitiligo was 6 years. Thirty-four patients had a body surface area involvement < 10% and 16 patients had a body surface involvement > 10%. We found the presence of associated autoimmune diseases in nine patients: five cases of Hashimoto thyroiditis, one case of Graves disease, one case of rheumatoid arthritis, one case of psoriasis and one case of alopecia areata with type 1 diabetes and autoimmune thrombocytopenia and hypothyroidism. The documented clinical and biological findings included a family history of vitiligo (14 patients) and elevated autoantibody levels of antithyroid peroxidase (anti-TPO) antibodies in 11 patients and antithyroglobulin (anti-TG) antibodies in eight patients. Histological parameters Isolated microvesicles in the epidermis were observed in 29 patients. Isolated detachment of melanocytes from the basal layer was identified in 12 patients. Detachment of melanocytes combined with microvesicles was found in six patients. Three patients did not exhibit either microvesicles or detachment of melanocytes. In control skin, melanocyte abnormalities and other epidermal changes were not found. E-cadherin was expressed normally. No infiltrate was detected in the dermis. In vitiligo lesions, a relationship between melanocyte disturbances and global histological findings was demonstrated (Fig. 1). Inflammatory isolated microvesicles including melanocytes and lymphocytes (29 cases) were statistically correlated to inflammatory infiltrate of epidermis and dermis (29 cases) with a P-value < 0001. A relationship between melanocyte disturbances and E-cadherin expression was identified (Fig. 1). Isolated melanocyte detachments (12 cases) were statistically correlated to a downregulation of E-cadherin (weak and moderate in seven cases) with a P-value of 003. A relationship between melanocyte disturbances and clinical aspect was observed (Table 1). Isolated microvesicles (29 cases) were statistically correlated to hypomelanotic lesions with poorly defined borders (28 cases) with a P-value < 0001. A relationship between melanocyte disturbances and vitiligo lesion activity was found (Table 1). Isolated microvesicles (29 cases) were statistically correlated to spreading vitiligo lesions (21 cases) with a P-value of 001. © 2014 British Association of Dermatologists

Vitiligo is an enigmatic disease with different clinical presentations and variable histological aspects. In our study, we have clearly demonstrated two melanocyte disappearance patterns implicated in two specific pathomechanisms that corresponded to proper clinical features and histological findings. Isolated microvesicles were correlated with hypomelanotic and spreading lesions with inflammatory infiltrate in the skin, while melanocyte detachments were correlated with amelanotic and stable lesions with absence of any inflammatory infiltrate and with E-cadherin defect. In the marginal skin of 29 cases of vitiligo, inflammatory microvesicles could correspond to a necrosis process. So, in our study, the finding of microvesicles in the epidermis and dermal CD8 infiltrate associated with melanophages was consistent for the immune destruction of melanocytes.15 Small defects in the basement membrane zone, which could permit invading lymphocytes to insinuate into the epidermis but also perhaps could facilitate the dropping down of damaged melanocytes into the dermis, were demonstrated in histological study of vitiligo skin.16,17 Isolated microvesicles were statistically correlated to spreading vitiligo lesions. Immunoreactivity of E-cadherin at keratinocyte membranes was slightly decreased or normal, being around 965% for score (2+) and (3+). In fact, it was demonstrated that if tumour necrosis factor-a caused significant increase in the soluble ICAM-1 level, it did not cause any significant increase in soluble E-cadherin level.18 So, it seems that there are no evident links between the E-cadherin adhesion molecule, implicated in epidermal unit cohesion and ICAM-1, another important adhesion molecule for T-cell attachment that plays a crucial role in the destruction of melanocytes. In accordance with all these data, the responsibility at the genetic level of immunoregulatory proteins targeting the innate and adaptative immunity, including human leukocyte antigen, protein tyrosinase phosphatase nonreceptor type 22 (PTPN22), nuclear localization leucinerich-repeat protein 1 (NLRP1) and decreased Tregs and CD4+ : CD8+ ratio, was suspected in nonsegmental vitiligo.19–21 In the marginal skin of stable vitiligo with amelanotic lesions and clearly defined borders, serial sections revealed that several melanocytes had undergone detachment. They could be found in various suprabasal locations according to the melanocytorrhagic process previously described.4 Such melanocyte detachment, considered as an apoptotic process, was experimentally induced in reconstructed epidermis treated with nonimmunological triggering factors including hydrogen peroxide and epinephrine.22 This finding could suggest the possibility of a nonimmunological loss of melanocytes. E-cadherin, which is expressed on keratinocytes throughout the epidermis, is primarily responsible for adhesion of human melanocytes to British Journal of Dermatology (2015) 172, pp331–336

334 Epidermal melanocyte disappearance in nonsegmental vitiligo, L. Benzekri et al.

(a)

(b)

(c)

(d)

(e)

(f)

(g)

(h)

(i)

Fig 1. Melanocyte disturbances and other histological findings. (a–c) Control skin. (a) Normal distribution of melanocytes (arrows), HMB45 staining 9400. (b) Homogeneous expression of E-cadherin at the surface of keratinocytes and melanocytes (arrow), E-cadherin immunolabelling 9400. (c) Absence of lymphocytic infiltrate in the dermis, CD8 T-lymphocyte immunolabelling 9400. (d–f) Melanocyte detachment. (d) Melanocytes (arrows), detachment of melanocytes (MD) from the basement membrane, HMB45 9400. (e) Heterogeneous and reduced expression of E-cadherin, mainly at the membrane of the keratinocytes and melanocytes (arrow) in the lower part of the epidermis, E-cadherin immunolabelling 9400. (f) Absence of inflammatory infiltrate, rare T8 lymphocyte (arrow) in the dermis, CD8 T-lymphocyte immunolabelling 9400. (g–i) Microvesicle. (g) Lymphocytes surrounding a melanocyte within an epidermal microvesicle (MV), lymphocytic infiltrate in the dermis, HMB45 staining 9400. (h) Heterogeneous and weak reduced expression of E-cadherin located around a microvesicle (MV), E-cadherin immunolabelling 9400. (i) clustered CD8 T lymphocytes (T8) in an epidermal microvesicle (MV), CD 8 T-lymphocyte immunolabelling 9400. Table 1 Comparison between patients with microvesicles and those with detached melanocytes

Age (years), mean  SD Vitiligo duration (years), mean  SD Family history, n (%) Autoimmune diseases associated, n (%) Positivity TPO antibodies, n (%) Positivity TG antibodies, n (%) Activity of lesions, n (%) E-cadherin immunological labelling, n (%) Weak Moderate Strong Hypomelanotic lesions with poorly defined borders, n (%) Inflammatory infiltrate, n (%) Body surface involvement < 10%, n (%)

Patients with microvesicles (n = 29)

Patients with detached melanocytes (n = 12)

P-value

38  16 111  101 5 (17) 4 (14) 21 (75) 23 (82) 21 (78)

37  11 134  125 5 (42) 4 (33) 7 (64) 8 (73) 4 (33)

078 041 019 034 047 051 001

1 13 15 28

(3) (45) (52) (97)

29 (100) 17 (59)

4 3 5 2

(33) (25) (42) (17)

003

< 0001

0 (0) 9 (75)

< 0001 048

TG, thyroglobulin; TPO, thyroid peroxidase.

British Journal of Dermatology (2015) 172, pp331–336

© 2014 British Association of Dermatologists

Epidermal melanocyte disappearance in nonsegmental vitiligo, L. Benzekri et al. 335

In six patients, we observed, in the same section of vitiligo epidermis, the coexistence of one inflammatory microvesicle and a few detached melanocytes from the basal layer with a downregulation of E-cadherin at the cell membrane of the keratinocytes (Fig. 2). Consequently, we can propose, as it has already been reported, that the loss of melanocytes in vitiligo could arise through a combination of different or successive pathogenic mechanisms that act in concert. In relation to the histopathology of perilesional skin in vitiligo, it was suggested that epidermal infiltrate of lymphocytes could be present in early stage lesions and disappear later when the disease became stable.2 It has also been proposed that the inflammatory element could be present in all cases, higher in active lesions than in long-lasting lesions, where it could be neglected and misdiagnosed.4 However, we demonstrated that there was no statistical correlation between the vitiligo duration and the two disappearance patterns. In conclusion, these data lead us to consider the possibility that inflammatory microvesicles or detached melanocytes could be the markers of two different pathomechanisms, which could be related in some cases. Vitiligo is probably a polygenic and multifactorial disease characterized in all cases by melanocyte loss. Fig 2. Coexistence in the same section of microvesicle (MV) of melanocyte detachment (MD) and melanocyte in basal location (m) (double staining, immunolabelling of E-cadherin and tyrosine-related protein (TRP)1 immunostaining of melanocytes 9400).

keratinocytes in vitro. At the margin of the lesion, as previously reported, we have identified, in the lower layers of the epidermis, a significant downregulation of E-cadherin in 58% of cases ranging from score (1) to score (2) coexisting with some melanocyte detachment. So, in the vitiligo epidermis, these adhesivity defects and some keratinocyte damage and dysfunction could certainly facilitate melanocyte detachment following mechanical trauma and/or biochemical stress and, consequently, a passive melanocyte death. After their detachment, the melanocytes could lose communication with proximal cells and extracellular matrix and would consequently be deprived of essential signals for their survival as was described in the ‘anoikis’ theory.23 In agreement with these adhesivity defects, genetic variants of the gene for discoidin domain receptor (DDR)1, which is a regulator of cell adhesion molecules, were identified in patients with vitiligo.24 Its reduced immunohistochemical expression was recently reported in lesional skin of vitiligo.25 DDR1 is a receptor tyrosine kinase that acts as a collagen IV adhesion receptor but could also interact with E-cadherin through the cell-junction complexes formed between DDR1 and E-cadherin.26 In these stable lesions, characterized by the absence of inflammatory reaction at the histological level, it was demonstrated that ICAM-1 is not usually expressed on basal melanocytes.7 This type of vitiligo lesion could be treated with success by melanocyte transplantation. © 2014 British Association of Dermatologists

References 1 Benzekri L, Gauthier Y, Hamada S, Hassam B. Clinical features and histological findings are potential indicators of activity in lesions of common vitiligo. Br J Dermatol 2013; 168:265–71. 2 Pretti Aslanian FM, Noe RA, Cuzzi T, Filgueira AL. Abnormal histological findings in active vitiligo include the normal appearing skin. Pigment Cell Res 2007; 20:144–5. 3 Hann SK, Park YK, Lee KG et al. Epidermal changes in active vitiligo. J Dermatol 1992; 19:217–22. 4 Le Poole IC, van den Wijnsgaard RM, Westerhof W, Das PK. Presence of T cells and macrophages in inflammatory vitiligo skin parallels melanocyte disappearance. Am J Pathol 1996; 148:1219– 28. 5 Dwivedi M, Laddha NC, Shah K et al. Involvement of interferongamma genetic variants and intercellular adhesion molecule-1 in onset and progression of generalized vitiligo. J Interferon Cytokine Res 2013; 33:646–59. 6 Laddha NC, Dwivedi M, Gani AR et al. Tumor necrosis factor B (TNFB) genetic variants and its increased expression are associated with vitiligo susceptibility. PLoS ONE 2013; 8:0081736. 7 Ahn SK, Choi EH, Lee SH et al. Immunohistochemical studies from vitiligo–comparison between active and inactive lesions. Yonsei Med J 1994; 35:404–10. 8 Dwivedi M, Laddha NC, Mansuri MS et al. Decreased regulatory Tcells and CD4(+)/CD8(+) ratio correlate with disease onset and progression in patients with generalized vitiligo. Pigment Cell Melanoma Res 2013; 26:586–91. 9 Gauthier Y, Carrio Andre M, Ta€ıeb A. A critical appraisal of vitiligo etiologic theories. Is melanocyte loss a melanocytorrhagy? Pigment Cell Res 2003; 16:322–32. 10 Gauthier Y, Carrio-Andre M, Lepreux S et al. Melanocyte detachment after skin friction in non lesional skin of patients with generalized vitiligo. Br J Dermatol 2003; 148:95–101.

British Journal of Dermatology (2015) 172, pp331–336

336 Epidermal melanocyte disappearance in nonsegmental vitiligo, L. Benzekri et al. 11 Kumar R, Parsad D, Kanwar AJ. Role of apoptosis and melanocytorrhagy: a comparative study of melanocyte adhesion in stable and unstable vitiligo. Br J Dermatol 2011; 164:187–91. 12 Tang AM, Eller M, Hara M et al. E-cadherin is the major mediator of human melanocyte adhesion to keratinocytes in vitro. J Cell Sci 1994; 107:983–92. 13 Kim NH, Lee AY. Reduced aquaporin 3 expression and survival of keratinocytes in the depigmented epidermis of vitiligo. J Invest Dermatol 2010; 130:2231–9. 14 Delmas V, Rubod A, Luciani F et al. E-cadherin is required for melanocyte maintenance in the basal layer and is altered in vitiligo melanocytes. J Invest Dermatol 2012; 132:S44–9. 15 Badri AM, Todd PM, Garioch JJ et al. An immunological study of cutaneous lymphocytes in vitiligo. J Pathol 1993; 170:149–55. 16 Bose SK, Ortonne JP. Focal gaps in the basement membrane of involved and uninvolved skin of vitiligo: are they normal? J Dermatol 1994; 21:152–9. 17 Westerhof W, Manini P, Napolitano A, d’Ischia M. The haptenation theory of vitiligo and melanoma rejection: a close-up. Exp Dermatol 2011; 20:92–6. 18 Parfiniewicz B, Pendzich J, Kapral M et al. The influence of TNFalpha on concentration of soluble adhesion molecules in cultures of HT-29 cells exposed to inositol hexaphosphate. Acta Pol Pharm 2012; 69:1291–7.

British Journal of Dermatology (2015) 172, pp331–336

19 Singh A, Sharma P, Kar HK et al. HLA alleles and amino-acid signatures of the peptide-binding pockets of HLA molecules in vitiligo. J Invest Dermatol 2012; 132:124–34. 20 Birlea SA, Ahmad FJ, Uddin RM et al. Association of generalized vitiligo with MHC class II loci in patients from the Indian subcontinent. J Invest Dermatol 2013; 133:1369–72. 21 Jin Y, Birlea SA, Fain PR et al. Genome-wide analysis identifies a quantitative trait locus in the MHC class II region associated with generalized vitiligo. J Invest Dermatol 2011; 131:1308–12. 22 Carrio-Andre M, Pain C, Gauthier Y, Ta€ıeb A. The melanocytorrhagic hypothesis of vitiligo tested on pigmented stressed, reconstructed epidermis. Pigment Cell Res 2007; 20:385–93. 23 Frisch SM, Screaton RA. Anoikis mechanism. Curr Opin Cell Biol 2001; 13:555–62. 24 Ricard AS, Pain C, Daubos A et al. Study of CCN3 (NOV) and DDR1 in normal melanocytes and vitiligo skin. Exp Dermatol 2012; 21:411–16. 25 Reichert-Faura A, Jung JE, Moreschi Neto V et al. Reduced immunohistochemical expression of Discoidin Domain Receptor 1 (DDR1) in vitiligo. J Eur Acad Dermatol Venereol 2013; 27:1057– 9. 26 Wang CZ, Yeh YC, Tang MJ. DDR1/E-Cadherin complex regulates the activation of DDR1 and cell spreading. Am J Physiol Cell Physiol 2009; 297:C419–29.

© 2014 British Association of Dermatologists

Possible patterns of epidermal melanocyte disappearance in nonsegmental vitiligo: a clinicopathological study.

The depigmentation of vitiligo results in a progressive and chronic melanocyte loss with rare melanocytes occasionally remaining in the epidermis or t...
655KB Sizes 0 Downloads 3 Views