Autoimmunity is considered as a key factor in vitiligo because of the clear association of the disorder with a personal or familial autoimmune diathesis, which targets in particular the thyroid gland in around 25% of patients (Taïeb and Picardo, 2007). The confusion which is frequently made in the published literature is to assume that vitiligo is itself a primary organ-specific autoimmune disorder, with cutaneous melanocytes being the target of a centrally dysregulated immune system (Basak et al., 2008). However, several observations suggest that a primary cutaneous cause should be considered with more attention and that immune disturbances may result in part of a primary skin anomaly (Gauthier et al., 2003). Along those lines, further challenging experiments of nature observed in the clinic suggest that segmental vitiligo (SV) may give clues to understand generalised (non-segmental) vitiligo (NSV) for several reasons developed thereafter.
Clinical findings in vitiligo challenge the widely accepted organ specific autoimmune pathomechanisms. We draw the attention to the fact that the distribution of segmental vitiligo (SV) fits in at least a subset of patients a pattern usually associated with cutaneous mosaicism. The association of SV to non-segmental vitiligo (NSV) now confirmed by several observations indicates a continuum between the two subsets with shared predisposing genetic factors, including genes operating specifically in the skin. Some pedigrees associating SV and NSV further suggest a mechanism of loss of heterozygosity for a dominant gene controlling part of the cutaneous phenotype. The mosaic hypothesis applies only to SV and to the rare SV-NSV association, but suggests that predisposing genetic factors in common NSV should also be searched directly in the skin. SV would be a good candidate disease to explore as a proof of principle of a new gene discovery strategy useful for multigenic disorders with organ specificity, applicable in priority to chronic inflammatory skin disorders.
The lines of Blaschko and segmental vitiligo
The lines described by Alfred Blaschko in 1901 (Blaschko’s lines BL) were based on drawings of cutaneous nevoid lesions (Blaschko, 1901). They have been revisited extensively by Happle (Happle, 1985) and considered to follow the dorso-ventral development of cellular components of the skin. When visible, BL reflect an underlying mosaicism demonstrated first in monogenic keratin disorders (Paller et al., 1994). BL have been postulated to correspond to a migration pattern restricted to cells of ectodermal or neuroectodermal origin (like melanocytes) (Taibjee et al., 2004), but several observations of purely dermal diseases following this pattern mitigate this view (Morice-Picard et al., 2007), and would rather favour the idea that BL reflect the development of the entire skin. The two types of lines (thin or large) may somewhat correspond to the underlying cellular origin of somatic mosaicism, dermal disorders tending to relate to broad bands, congenital melanocytic nevi (dermal and epidermal) being also exceptionally observed along this pattern (Figures 1 and 2). Other types of patterns which do not fit Blaschko’s lines (checkerboard, phylloid, garment-type) are also observed in the dermatology clinic (Happle, 2004), and may concern the melanocytic system or other types of cells.
The point was previously made that some blaschkolinear distribution patterns in SV are strikingling similar to that of epidermal nevi (Taieb, 2000), which are the established expression of somatic mosaicism for keratin (Paller et al., 1994) and other genes (Hafner et al., 2006) but may exceptionnally also involve germline cell mutations, a fact which explains the coexistence of segmental and generalized patterns in some rare pedigrees (Nazzaro et al., 1990). The absence of observed vitiligo inheritance among the offspring of segmental vitiligo patients provides also some additional support for the somatic mosaicism hypothesis of SV (Hann and Lee, 1996). The dermatomal distribution is usually considered as reflecting better the distribution of SV, but cases intersecting dermatomes without «filling» their theoretical territory of distribution are difficult to relate to this etiologic background. The sympathetic anomalies noted in favour of the neurogenic theory of SV (Wu et al., 2000) may also be a confounding factor related to the absence of melanocytes which have been considered as «neurons of the skin» and can release several neuromediators (Moellmann et al., 1973). Figure 3 shows how confusing can be the pattern on a limb if isolated. On the face, the lines of Blaschko have been drawn more recently (Happle and Assim, 2001), and some cases of SV fit clearly better those lines than dermatomal territories (Taieb, 2000) (Figure 4). This being stated, the hypothesis of a cutaneous mosaicism in SV based on distribution analysis does not rule out other triggering factors, especially neurogenic ones. It does not also exclude the possibility of an autoinflammatory component extending locally beyond the area demarcated by the developmental lines, but there is currently very limited histological evidence of inflammation in SV as opposed to NSV (Kim et al., 2008).
The association of SV and NSV
Following our review on the melanocytorrhagic hypothesis (Gauthier et al., 2003) which showed a pediatric case associating SV and NSV, other cases have been reported (Mulekar et al., 2006; Schallreuter et al., 2007, 2008). This phenomenon may have been neglected previously due to NSV masking SV, which indeed was the case in our princeps case (Gauthier et al., 2003) unveiled by UVB narrow band phototherapy. This could be the case when patients are examined in adulthood after a long lasting disease. In patients with mostly fair skin coming for SV, one should look more systematically at other minor symmetrical sites of involvement using Wood’s lamp in a completely dark room (Figure 5).
The usually more severe and earlier presentation of SV in the context of associated NSV favours a dosing effet of a common predisposing gene and a loss of heterozygosity (LOH) mechanism like in monogenic disorders caused by cutaneous somatic mosaicism (Happle, 2001). This corresponds to type II mosaicism according to Happle, as opposed to type I which results from only one mutation, causing an identifiable phenotype if dominant (Happle, 2001). This sequence (early SV, late NSV) may also reflect the role of a first cutaneous gene defect causing SV triggering a generalised immune response against cutaneous melanocytes supported by another immune-related gene defect. However, pedigrees showing the presence of cases with established SV in families with previous cases of NSV (Figure 6 illustrates one family out of four similar noted in the pediatric vitiligo clinic in 1 yr) reinforce the possibility of a mechanism close to type II mosaicism already demonstrated in a monogenic skin disorder using laser capture of mutated cells and subsequent DNA sequencing (Poblete-Gutiérrez et al., 2004).
Autologous cell therapy as a cure for SV: a further argument for a mosaic melanocyte gene defect
As already noted (Gauthier et al., 2003; Taieb, 2000), the good and long-term take of epidermal cell suspensions or autologous melanocyte cultures in SV as compared to NSV argues for the replacement of genetically abnormal cells by normal ones and concomitant absence of auto-inflammation/immunity This kind of in vivo experiments, which have been reported by several groups, suggest that cells coming from unaffected skin of SV patients cure a local genetic defect. They strengthen the case for melanocytes being the primary defective cell in SV since autologous pure melanocyte cultures have been used successfully in this setting (Chen et al., 2000). The poor success of autografts in NSV suggest that all epidermal melanocytes are genetically abnormal. However, there is at least ex vivo evidence of an additive effect of NSV keratinocytes (Cario-André et al., 2007), in favour of a more generalised cellular defect.
Comments and proposals
Until recently, the target theory for skin involvement in vitiligo (the skin pigmentary system as the target of autoimmunity) has underemphasized information coming from the skin itself. If somatic mosaicism holds true for SV, it seems reasonable to speculate that differences in affected and unaffected skin in SV, as well as differences in NSV associated with SV should provide relevant information on the cause of the disease. Technically this might be difficult if cultures need to be implemented in lesional skin for melanocytes (Cario-André et al., 2007). CGH arrays might be used for detection of mosaicism in conjunction with microarrays from RNA extracted from cell cultures of various cell types (keratinocytes, melanocytes, or fibroblasts). We have so far limited evidence that NSV is caused primarily by melanocytes, keratinocytes seeming also involved. An impaired behaviour of vitiligo keratinocytes has been confirmed in vitro (Bondanza et al., 2007; Cario-André et al., 2007; Schallreuter and Pittelkow, 1988). Especially, lesional keratinocytes have been shown to have a shorter life span associated with modification of proliferation and senescence marker expression (p16, p53, p21) compared to keratinocytes from clinically non-involved skin (Bondanza et al., 2007). Recent data in NSV indicate abnormal expression profiles of genes involved in melanocyte development, intracellular processing and trafficking of tyrosinase gene family proteins, packing and transportation of melanosomes, cell adhesion and antigen processing and presentation (Strömberg et al., 2008). Further focusing either on one gene or on a subset of genes might be difficult. As compared to NSV, SV would probably be more informative in terms of skin causative genes, because the background of inflammation/autoimmunity would be probably more limited. Autoimmunity manifested by associated autoimmune disorders is rare in SV (Iacovelli et al., 2005). Autoantibodies to melanocytic antigens are also not common (Kemp et al., 1997). The only shared «immune» phenotype is the presence of halo nevi (Barona et al., 1995), which seem nearly as frequent in SV and NSV (6.4 versus 8.6%), and quite high as compared to controls without vitiligo – less than 1%- (Larsson and Liden, 1980). This finding, if confirmed, may suggest that halo nevi-associated SV have some counteracting protective mechanisms respective to the development of NSV.
The mosaic hypothesis applies only to SV and to the rare SV-NSV association, but suggests that predisposing genetic factors in common NSV should also be searched directly in the skin. Based on our clinical observations of SV and mixed SV-NSV, we would like to propose an alternative scenario for NSV. Still unknown genetic predisposing factors would affect first the skin pigmentary system, and might secondly activate skin immune/inflammatory responses leading to a more severe expression of the disease. This scenario is somewhat derived from that proposed for another common skin disorder, atopic dermatitis (Taieb, 1999), for which a skin barrier genetic dysfunction comes first (Palmer et al., 2006) and may produce direct inflammation, and can engage in a facultative second step the immune system into an «allergic» TH2 dependent pathway in a subset of patients. This two-step model is consistent with the common presentation of NS vitiligo as an isolated chronic cutaneous disorder in the majority of patients. AIS 3 which segregates with «pure» NSV phenotypes is already a good candidate gene on chromosome 8 (Spritz et al., 2004). Since friction/pressure-prone areas are frequently initially affected, it would be tempting to speculate that the primary skin anomaly affects the upper layers of the epidermis to explain the Koebner’s phenomenon (Gauthier, 1995); it would explain nicely the subsequent epidermal activation of innate immunity-based mechanisms via stratum corneum activation of inactive Il1beta precursors (Taieb, 2007) in the context of the recent identification of NALP1 as a predisposing gene in NSV associated with auto-immune disorders (Jin et al., 2007). Pressure may also affect directly basal epidermal layer biology if an abnormal adhesion of melanocytes comes first to explain their detachment (Gauthier et al., 2003). Also envisaged is a genetic cause limiting survival and self renewal of epidermal and follicular melanocytes, the latter being hit earlier by the disease in SV as compared to NSV (Taieb, 2000). The clear association of NSV to familial hair greying (Halder et al., 1987; Taïeb and Picardo, 2007) also favours a genetic background affecting melanocyte survival. The expression of several genes influencing such pathways is indeed modified in NSV (Strömberg et al., 2008).
Beyond pigment cell disorders, the finding of superimposed segmental lesions in acquired chronic skin disorders (Happle, 2007) point to more general pathomechanisms which highlight the causative role of specific skin-associated predisposing genes. Non-skin disorders have probably a mosaic gene expression but are less accessible to observation and sampling. Because of its frequency, SV would be a good candidate disease to explore as a proof of principle for this gene discovery strategy in multigenic disorders with organ specificity, applicable in priority to chronic inflammatory skin disorders. The ultimate advantage of a more skin-focused gene approach in vitiligo would be the development of targeted topical corrective treatments, safer to use and overall more logical to treat early if the disease is primarily a cutaneous one.