
	Version PDF

Alopecia areata: autoimmune basis of hair loss

European Journal of Dermatology. Volume 14, Number 6, 364-70, November-December 2004, Review article

Summary

Author(s) : Andrew F ALEXIS, Raghunandan DUDDA-SUBRAMANYA, Animesh A SINHA , Ronald O. Perelman Department of Dermatology, New York University School of Medicine, New York, NY, Department of Dermatology, Joan and Sanford Weill Medical College of Cornell University, New York, NY.

Summary : Alopecia areata (AA) is a heterogeneous disease characterized by nonscarring hair loss on the scalp or any hair-bearing surface [1]. A wide range of clinical presentations can occur - from a single patch of hair loss to complete loss of hair on the scalp (alopecia totalis) or the entire body (alopecia universalis). Particularly in severe or chronic cases, AA may cause considerable psychological and emotional distress for affected individuals [2]. The estimated lifetime risk of developing AA is 1.7% [3]. While the precise etiology of this common disorder has not been elucidated, a substantial body of evidence suggests that AA is an organ-specific, autoimmune disease, targeted to hair follicles. However, the antigenic target(s), mechanisms, and consequences of autoimmune attack in AA have yet to be determined. Here, we critically explore the evidence supporting the hypothesis that AA is an autoimmune disease and propose specific pathways by which self-directed immune responses are generated.

Keywords : alopecia areata

Pictures

ARTICLE

Auteur(s) :, Andrew F ALEXIS*¹, Raghunandan DUDDA-SUBRAMANYA*², Animesh A SINHA^2,*

¹Ronald O. Perelman Department of Dermatology, New York University School of Medicine, New York, NY
²Department of Dermatology, Joan and Sanford Weill Medical College of Cornell University, New York, NY

accepté le 20 Juillet 2004

Clinical features and pathophysiology of alopecia areata

The clinical hallmark of AA is the sudden development of a well circumscribed patch of hair loss on normal appearing skin [4]. The majority of patients present with limited patches of alopecia on the scalp that regrow spontaneously within 1 year [4]. However, an estimated 7% to 10% of patients may experience more extensive and chronic forms of the disease [4]. The factors that influence the course and extent of alopecia are not known, making it impossible to predict the disease outcome at the time of presentation.

The loss of hair in AA is related to alterations in the normal cycle of hair growth. The normal hair cycle consists of three distinct phases: (1) anagen (the phase in which hair growth occurs), (2) catagen (during which the hair follicle involutes), and (3) telogen (the resting phase). This cycle of growth, involution, and rest is regulated by complex interactions between the follicular epithelium and the adjacent dermal papilla [5]. However, the molecular signals that mediate the hair growth cycle are not yet fully understood and are the subject of ongoing investigation [5].

In AA, hair follicles enter telogen and late catagen prematurely [6, 7]. Tissue biopsies from the active border of AA lesions show a normal number of follicles, the majority of which are in telogen or late catagen [7]; on horizontal sections of lesional scalp biopsies, a decreased anagen-to-telogen ratio of hair follicles is observed [8]. With the exception of long-standing alopecia, hair follicles are preserved, even in clinically hairless lesions [9]. In long-standing lesions, miniaturization of anagen follicles is seen, accompanied by a decrease in follicle density [6, 8, 10].

Pathodynamically, involved hair follicles in AA are thought to undergo arrest in early anagen (anagen III/IV) – during which the inner root sheath of the hair follicle is conical and keratinized, and differentiation of the underlying hair cortex has just begun [7, 9, 11]. From this stage, arrested hair follicles may return prematurely to telogen – the resting phase – and continue to undergo shortened cycles [12].

The histopathological hallmark of AA is the presence of a perifollicular and intrafollicular lymphocytic infiltrate [12]. In particular, T-lymphocytes (predominantly CD4+ cells) as well as Langerhans cells and macrophages are found in the dermal papilla and matrix of involved hair follicles [13-15] Lymphocytic infiltration of the anagen hair bulb and dermal papilla is accompanied by aberrant expression of HLA class I and class II antigens, and intercellular adhesion molecule 1 (ICAM-1) on the follicular epithelium [16-19]. Of note, the inflammatory infiltrate in AA spares the isthmus of the hair follicle – the proposed site of hair follicle stem cells [20]; this is in contrast to the pattern seen in inflammatory scarring alopecias and may explain why hair follicles remain intact in AA [6, 12].

In many instances, animal models have been useful in helping elucidate mechanisms of disease. In particular, studies involving animal models with spontaneous or induced AA have helped provide greater insights into the pathogenesis, clinical behavior, and treatment modalities of AA. The C3H/HeJ mouse [21], Dundee experimental bald rat (DEBR) [22], and Smyth chicken [23] are models of spontaneous AA. In addition, spontaneous AA has been reported in several canine species [24]. AA can also be induced experimentally in C3H/HeJ mice [25]. Moreover, animal models may be particularly helpful in identifying genetic susceptibility and severity markers of AA [26].

Evidence for autoimmune basis of disease

General considerations

The earliest evidence that autoimmunity may play a role in the pathogenesis of AA stems from observed associations between classical autoimmune disorders and AA. In particular, vitiligo and thyroid diseases – including Hashimoto’s thyroiditis, have been found in increased frequency among patients with AA [27, 28]. Moreover, antithyroid antibodies and thyroid microsomal antibodies have been found in AA patients [29, 30]. Similar associations with other autoimmune diseases have been reported in numerous case reports, albeit less consistently. These include pernicious anemia, myasthenia gravis, and the candida-endocrinopathy syndrome [6, 31].

The efficacy of immunomodulatory agents in the treatment of AA further implicates immune mechanisms in the development of this disease. In particular, corticosteroids, photochemotherapy (PUVA), cyclosporin A, and topical sensitizers such as diphenylcyclopropenone have all been used successfully for treating AA. While all of the above therapies have in common a modulating effect on lymphocyte function, it remains unclear whether their effects are specifically related to suppression of lymphocytes or merely nonspecific effects on the inflammatory response.

Associations between specific human leukocyte antigen (HLA genes) and AA lend further circumstantial evidence that autoimmune mechanisms are involved in the pathogenesis of disease. In particular, HLA class II alleles – DR4, DR5 (DR11), [32-39] DR6 [40], and HLA-DQ3 [including both subtypes DQB1*0301 (DQ7) and DQB1*0302 (DQ8) ] [32-34, 40, 41] – have been most consistently associated with AA. Furthermore, specific HLA associations have been reported as markers of clinical sub-types of AA [42]. In the more severe forms of the disease – alopecia totalis (AT) and alopecia universalis (AU) – DRB1*0401 (DR4) and DQB1*0301(DQ7) are expressed with increased frequency [42].

Cellular and humoral responses

The characteristic presence of lymphocytes around lesional hair follicles has also served as evidence that abnormal immune responses may be involved in AA. As mentioned, the infiltrate consists predominantly of activated T lymphocytes along with macrophages and Langerhans cells [14, 15]. Both CD4+ and CD8+ T lymphocytes are observed peri- and intra-follicularly, with a predominance of CD4+ T cells [14, 43, 44].

The relative functional contribution of these T-cell subtypes in AA is not fully understood, but both cell types appear to be important mediators of disease [45]. Immunophenotyping studies in both humans and canine models of AA have demonstrated a predominance of CD8+ T cells in the intrafollicular infiltrate, while CD4+ cells are preferentially located peribulbarly [24]. As such, CD8+ T cells appear to directly attack the follicular epithelium and are likely key effectors of the pathogenic changes characteristic of AA. Interestingly, the extent of lymphocytic infiltration of the follicle may correlate with responsiveness to therapy – as increased lymphocytic infiltration has been associated with a poor response to treatment with contact immunotherapy (diphencyprone) [46].

The concomitant increased expression of HLA class I (HLA-A, B, and C) and class II (HLA-DR) antigens, and ICAM-1, further suggests a role for cell mediated immune mechanisms in AA [19, 47]. As these antigens (as well as Langerhans cells) are normally absent in the lower portions of the hair follicle, their presence in AA lesions suggests that antigen presentation and T cell activation are important factors in the development of this disease.

Moreover, cytokine release is also likely involved in the increased expression of HLA antigens and ICAM-1. This has been suggested by experiments demonstrating increased protein and mRNA expression of Th1-cytokines (IFN-γ, IL-2), and IL-1β in skin biopsies from patients with AA [48]. In addition, increased serum levels of IFN-γ in patients with AA compared to normal controls has been reported, further suggesting a role for this cytokine [49].

Additional evidence supporting T cell mediated autoimmune processes in the pathogenesis of AA is provided by Kalish et al. [50] who used limiting dilution analysis to determine the frequency of autoreactive T lymphocytes in scalp biopsy specimens and peripheral blood of AA patients. Statistically significant enrichment of T cells that proliferate in the context of autologous irradiated peripheral blood mononuclear cells was seen in four of seven patients studied [50]. However, enrichment of autoreactive T cells is not specific to AA; it has also been observed in allergic contact dermatitis [51].

Stronger evidence for circulating immune mechanisms involved in the development of AA come from studies in athymic (nude) mice. Lesional scalp skin from AA patients grafted onto nude mice shows regrowth of hair, suggesting a loss of inhibition by T cells. These experiments strongly suggest that the key mediators of AA reside outside the follicle itself; given that lesional skin grafts are able to regrow hair in an athymic host, functional T cell populations appear to be involved in the pathogenesis of this disease [52].

A role for humoral immunity in AA has also been investigated, producing disparate results. Autoantibodies reactive to perifollicular capillary endothelium were reported by Nunzi et al. [53], but this has not been confirmed by follow-up studies [6]. Antibodies to melanoma antigens in the sera of AA patients have been reported [54], although they were also found to be present in the sera of some normal individuals [54]. Similarly, antibodies that bound to human anagen hair follicle extracts in patients were also found in some normal individuals, albeit at lower titers [55]. The specific targets of follicular antibodies seen in AA include multiple structures [56]; as such, the significance of the above findings remains unclear.

Arguing against a pivotal role for humoral immune factors in disease is the finding of AA in patients with impaired antibody production as in common variable immunodeficiency [6, 57]. In addition, experiments involving the passive transfer of serum from AA and AU patients failed to induce lesions on human scalp skin transplanted onto nude mice [58], further undermining the relevance of humoral factors in the pathogenesis of disease. Autoantibodies may still, however, be of significance in disease perpetuation and progression [21, 59].

The strongest evidence implicating autoimmune mechanisms in the pathophysiology of alopecia areata has been provided by studies involving mice with severe combined immunodeficiency (SCID). In a set of experiments by Gilhar et al. [60], AA was induced on human scalp explants transplanted onto SCID mice by injection of autologous T lymphocytes from lesional skin. In particular, only T lymphocytes cultured with hair follicle homogenate and antigen presenting cells were capable of inducing AA in scalp explants. Both the clinical and histopathologic features of AA were reproduced in this model, including hair loss, perifollicular T lymphocyte infiltration, and expression of HLA-DR and ICAM-1 in the follicular epithelium. The induction of these changes was not a nonspecific effect of T cell activation, as lesions could not be produced by injection of IL-2 activated T cells from peripheral blood or scalp. Moreover, the fact that these changes could not be induced by lesional scalp T cells that were not cultured with follicular homogenate strongly suggests that the T lymphocytes involved in AA are reacting to specific follicular antigens [60]. As such, T cells targeted to follicular autoantigens appear to be key mediators in the pathogenesis of AA.

Autoantigenic targets

While the above studies strongly support the hypothesis that AA is an autoimmune disease, direct evidence is currently lacking. In particular, the pathogenic target of a putative autoimmune response has yet to be identified. Based on research to date, the potential targets of an autoimmune response in AA include follicular keratinocytes, melanocytes, and dermal papilla antigens [61].

The basis for the hypothesis that the autoantigen in AA may be melanocyte derived comes from clinical observations [62]. It is often observed that pigmented hairs are lost in greater measure than nonpigmented hairs in patients with AA [63-65]. Moreover, the initial regrowth of hair is often by nonpigmented (white) hairs [66]. Second, the hair bulb, which contains a large number of melanocytes, is a key site of infiltration by lymphocytes in AA [67]. Third, passive transfer of lymphocytes from mice immunized with melanoma associated antigens has been shown to induce clinical and histopathologic features of AA in naïve mice [68]. Most recently, Gilhar et al. [64] identified five melanocyte-derived peptides – all of which were HLA-A2-restricted – able to activate AA lesional scalp T cells when cultured together in vitro, and induce AA in human scalp explants on SCID mice. The authors concluded that multiple melanocyte epitopes can act as autoantigens in AA. One major weakness of this work is that a consistent association between HLA-A2 and AA has not been established.

Difficulties in identifying target protein and associated disease relevant epitopes abound. First, the autoimmune response could be triggered by one or more T cell epitopes. Relevant epitopes could then be derived from either one or more source proteins. Second, the epitope(s) responsible for disease induction may differ from those involved in disease progression. In particular, additional epitopes – distinct from those which induced the immune response – could be released over the course of the inflammatory process. This phenomenon is referred to as epitope spreading [69]. Third, certain epitopes may represent non-specific secondary targets of the inflammatory response – mere epiphenomena of the immune cascade. Fourth, the disease-inducing epitopes may be “cryptic” - derived from self-antigens that are normally sequestered from immune recognition and only exposed after injury to an immune privileged site [70]. As such, it will be important to distinguish antigen(s) or epitopes that are truly disease relevant – i.e. responsible for the induction and progression of disease – from those that are merely non-specific secondary targets of the inflammatory response. Moreover, differential immunodomains from the same or distinct target autoantigens may account for the variable phenotypes within the clinical spectrum of AA. Careful delineation of all the disease relevant target antigens and epitopes will be crucial to increasing the depth of our understanding of the clinical heterogeneity and pathogenesis of AA.

Putting it together - proposed mechanisms of disease

Conceptually, the key events involved in the development of AA evolve in three distinct physiologic environments: the thymus; the peripheral blood and skin-draining lymph nodes; and the hair follicle or target tissue. The initial prerequisite for an autoimmune response in AA is that immature T cells with the potential to react to self-antigens escape deletion in the thymus, and are positively selected to mature. These mature T cells with autoimmune potential migrate to the peripheral blood and lymph nodes, where after an encounter with relevant self or cross-reacting foreign antigen(s), they are activated to proliferate. Activated T cells would then have to escape peripheral tolerance mechanisms and proceed to the end-organ – the hair follicle – and induce the pathologic events that characterize AA. To that end, complex interactions between predisposing genetic and environmental factors likely play a role in the induction of immune-mediated responses in AA (( Figure 1 )).

The thymus: breach of central tolerance

Beginning in the thymus, bone marrow-derived T cell progenitors undergo the process of positive and negative selection based on their affinity for self peptide-MHC complexes [71-73]. The strength of the interaction between the T cell receptor and the self peptide-MHC complexes determines the ultimate fate of the developing T cell [71-73]. Weak interactions result in survival and differentiation of the T cell (positive selection), whereas high affinity binding results in apoptosis (negative selection) [71-73]. In this way, mature T cells that emigrate from the thymus are self-MHC restricted and are depleted of cells with a high affinity for self antigens.

The production of T cells that can distinguish between self and non-self antigens hinges on the ability of the thymus to present a comprehensive repertoire of self peptides. However, the thymus does not express many organ or tissue-specific antigens [71] and indeed, many hair follicle-specific antigens are most likely not present in the thymus for presentation to developing T cells [62]. As such, the pool of peripheral T cells may contain clones with potential specificity for hair follicle antigens that were not present in the thymus.

Underlying the affinity of T cell-self-peptide interactions, specific HLA molecules involved in presenting antigens during positive and negative selection in the thymus are critical to the shaping of the functional T cell repertoire. The aforementioned associations between specific HLA haplotypes and AA support this hypothesis. Individuals with specific HLA haplotypes (e.g. HLA-DQ3, DR6, DR11, DQ7) may be more likely to develop a T cell repertoire that encompasses autoreactive potential. Moreover, as different clinical subsets of AA have shown specific HLA associations (e.g. AA with DQ3, and AT/AU with DR4 and DQ7), an individual’s HLA haplotype may affect the T cell repertoire such that T cells recognizing unique follicular antigen-MHC complexes produce different clinical presentations.

The periphery

Escaping peripheral tolerance

While CD4+ and/or CD8+ T cells with autoreactive potential can escape into the periphery as a consequence of the inherent imperfections of central tolerance mechanisms in the thymus, their ability to induce an autoimmune response requires antigen-specific activation and escape from peripheral tolerance. In the periphery, T cells that recognize and are activated by self peptides generally undergo clonal expansion, followed by deletion or the induction of anergy [72]. Defects in clonal deletion or anergy induction could lead to the accumulation of autoreactive T cells and hence autoimmune disease.

Clonal deletion, which involves activation-induced cell death (apoptosis) of self-reactive T cell clones, is largely mediated through the interaction of the Fas death receptor (CD95) with its ligand [71, 72]. Theoretically, decreased expression or defective function of Fas and/or Fas ligand on follicular T cells could lead to prolonged survival and activity of autoreactive T cells in AA.

Anergy, or functional inactivation, involves the interaction of T cell adhesion molecules, CTLA-4 and PD-1 (programmed cell death 1), with their respective ligands (CD80/86, PDL1/2) [71]. Defective expression of these molecules could also lead to autoimmunity. Interestingly, recent work has demonstrated increased expression of CTLA-4 and Fas ligand in the spleens of AA resistant mice [74]. Further studies are necessary to elucidate the precise roles of these adhesion molecules in the development of AA.

In addition to clonal deletion and anergy, regulatory T cells are likely to play a role in preventing autoimmune reactions in the periphery [71]. CD 4+ CD25+ T cells are thought to suppress immune responses through down-regulatory adhesion molecule interactions (e.g. CTLA-4) or by cytokine release - transforming growth factor (TGF)-β or interleukin (IL) –10 [71-73]. Decreases in regulatory T cell number or function can promote the development of autoimmune reactions [71]. Indeed, in AA, there is recent evidence in murine models of AA demonstrating a marked decrease in CD4+CD25+ regulatory T cells [75].

Molecular mimicry

An autoimmune attack of hair follicle antigens could arise from T cells that are activated by foreign (e.g. microbial) antigens, but also cross-react with follicular self proteins [71]. This concept, known as molecular mimicry, could explain the reported associations between microbial factors, such as CMV, and AA [76]. In this way, microbes expressing epitopes that resemble hair follicle peptides may trigger the activation of autoreactive T cells leading to AA. While confirmatory experimental evidence for this hypothesis is lacking, reported links between infectious environmental factors and AA [76] hints that molecular mimicry may play a role in this disease.

Breach of immune privilege

The anagen hair bulb represents an immune privileged site as do the anterior chamber of the eye, testes, and trophoblast [77]. Such sites are characterized by the absence of MHC class I expression and the presence of immunosuppressive cytokines (e.g TGF-β) [77]. A breakdown in this immune privilege may be a triggering factor in AA. Aberrant expression of MHC and ICAM-1 molecules in hair follicles could play a key role in initiating AA. Paus et al. [62] suggest that, in the context of an inflammatory insult, the anagen hair bulb can be induced to express MHC class I molecules. In particular, infections, follicular microtrauma, or microbial superantigens may induce the release of the proinflammatory cytokine, IFN-γ, inducing the ectopic expression of MHC class I and class II molecules on follicular bulb cells. This, in turn, could lead to the induction of CD8+ and CD4+ T cells targeted to newly exposed follicular antigens – normally sequestered from immune recognition. In this model, either environmental or genetic factors may contribute to the collapse of immune privilege in the hair bulb leading to disease induction.

The target tissue

After escaping negative selection in the thymus, activation by self- or foreign antigens presented in skin-draining lymph nodes, and evasion of peripheral tolerance mechanisms, autoreactive T cells that reach the target tissue are poised to induce the cascade of autoimmune events in AA. The previously discussed involvement of both CD4+ and CD8+ T lymphocytes, follicular autoantibodies, differential Fas ligand expression, and cytokine release in AA strongly suggest that many of the above factors may contribute to the pathologic events within the hair follicle. However, the precise factors that link immune dysregulation to the induction of hair loss in AA remain unclear. Possibilities include: (i) direct cytotoxicity by CD8+ T cells, natural killer (NK) cells, or NK-T cell activity; (ii) antibody dependent cell-mediated cytotoxicity (ADCC); (iii) apoptosis of hair follicle keratinocytes via Fas-Fas ligand interactions; or (iv) cytokine-induced inhibition of the hair cycle. Future work will be required to distinguish which factors are responsible for disease induction, progression, and resolution of this disease.

Conceptually, the culmination of the aforementioned immunologic events should result in the premature arrest of the hair cycle in early anagen and thus, incomplete keratinization of the developing hair cortex [6, 7, 11]. There is evidence to suggest that the link between lymphocytic infiltration of the follicle and the disruption of the hair follicle cycle in AA may be provided by a combination of factors, including cytokine release, cytotoxic T cell activity, and apoptosis. First, increased IL-1β, (along with IFN-γ, IL-2, and TNF-α) expression has been demonstrated in AA lesions [44, 48, 78, 79]; as IL-1β has been shown to potently inhibit hair growth in vitro [78], it may play a role in hair cycle dysregulation in AA. Second, a role for direct T cell cytoxicity has been suggested by the demonstration (by in situ hybridization) of granzyme B expression within the peri- and intrafollicular mononuclear cell infiltrate of AA lesions [44]. This serine proteinase, contained within cytoxic T lymphocytes, is thought to be critical to direct T cell cytotoxicity via apoptosis. Third, apoptosis via Fas-Fas ligand interactions is suggested by immunohistochemical studies demonstrating Fas protein expression on hair follicle keratinocytes in lesional skin biopsies of AA patients [44].

Conclusions and outlook

Taken together, autoimmune destruction in AA likely involves interwoven immunologic pathways, influenced at least in part by genetic and environmental events. Based on existing knowledge, the currently accepted mechanism of the pathogenic cascade that leads to disease development is illustrated in ( Figure 2 ). The key concepts illustrated in the figure have been well described in recent reviews of the Fourth International Research Workshop on Alopecia Areata [80, 81].

Despite major advances in our understanding of the immunological mechanisms in AA, several unanswered questions remain. What is the nature of the autoantigen(s) in AA? What is the specificity and phenotype of autoreactive cells targeted to the hair follicle? What factors (perhaps genetic and/or environmental) lead to a breach in the immunologic tolerance of hair follicle-associated antigens? Answers to these questions will not only confirm the hypothesis that AA is an organ-specific autoimmune disease, but will also set the stage for the development of novel, disease-specific therapies. In fact, new biologic therapies currently approved or in development for other inflammatory diseases (including psoriasis, rheumatoid arthritis, and Crohn’s disease), may have potential applications in the treatment of AA. Possible therapeutic strategies using these agents include inhibition of T cell activation and/or migration, depletion of memory effector T-cells, as well as inhibition or modulation of inflammatory cytokine release. Alternatively, identifying the autoantigen(s) involved in AA could potentially lead to the development of immunotherapies using disease-relevant peptides to induce tolerance. Undoubtedly, continued progress in elucidating the dysregulation of immune pathways that lead to AA can be expected to identify specific and effective therapies for this putative organ-specific autoimmune disease.

References

1 McDonagh AJ, Tazi-Ahnini R. Epidemiology and genetics of alopecia areata. Clin Exp Dermatol 2002; 27: 405-9.

2 Schmidt S, Fischer TW, Chren MM, et al. Strategies of coping and quality of life in women with alopecia. Br J Dermatol 2001; 144: 1038-43.

3 Safavi KH, Muller SA, Suman VJ, et al. Incidence of alopecia areata in Olmsted County, Minnesota, 1975 through 1989. Mayo Clin Proc 1995; 70: 628-33.

4 Madani S, Shapiro J. Alopecia areata update. J Am Acad Dermatol 2000; 42: 549-66; quiz 67-70.

5 Paus R, Cotsarelis G. The biology of hair follicles. N Engl J Med 1999; 341: 491-7.

6 McDonagh AJ, Messenger AG. The pathogenesis of alopecia areata. Dermatol Clin 1996; 14: 661-70.

7 Messenger AG, Slater DN, Bleehen SS. Alopecia areata: alterations in the hair growth cycle and correlation with the follicular pathology. Br J Dermatol 1986; 114: 337-47.

8 Whiting DA. Histopathology of alopecia areata in horizontal sections of scalp biopsies. J Invest Dermatol 1995; 104: 26S-27S.

9 McDonagh AJ, Messenger AG. Alopecia areata. Clin Dermatol 2001; 19: 141-7.

10 Hair Research. 3rd Annual meeting of the European Hair Research Society. Berlin, October 2-3, 1992. Skin Pharmacol 1994; 7: 1-108.

11 Van Scott EJ. Morphologic changes in the pilosebaceous units and anagen hairs in alopecia areata. Journal of Investigative Dermatology 1958; 31: 35.

12 Messenger AG, McDonagh AJG. In: Francisco M, Camacho VAR, Price Vera H, editors. Hair and its disorders: biology, pathology and management. London: Dunitz, 2000: 181.

13 Messenger AG, Simpson NB. In: Dawber R, editor. Diseases of the hair and scalp. 3rd edn, Oxford; Maldern, MA: Blackwell Science, 1997: 344.

14 Perret C, Wiesner-Menzel L, Happle R. Immunohistochemical analysis of T-cell subsets in the peribulbar and intrabulbar infiltrates of alopecia areata. Acta Derm Venereol 1984; 64: 26-30.

15 Wiesner-Menzel L, Happle R. Intrabulbar and peribulbar accumulation of dendritic OKT 6-positive cells in alopecia areata. Arch Dermatol Res 1984; 276: 333-4.

16 Brocker EB, Echternacht-Happle K, Hamm H, et al. Abnormal expression of class I and class II major histocompatibility antigens in alopecia areata: modulation by topical immunotherapy. J Invest Dermatol 1987; 88: 564-8.

17 Gupta AK, Ellis CN, Cooper KD, et al. Oral cyclosporine for the treatment of alopecia areata. A clinical and immunohistochemical analysis. J Am Acad Dermatol 1990; 22: 242-50.

18 McDonagh AJ, Snowden JA, Stierle C, et al. HLA and ICAM-1 expression in alopecia areata in vivo and in vitro: the role of cytokines. Br J Dermatol 1993; 129: 250-6.

19 Messenger AG, Bleehen SS. Expression of HLA-DR by anagen hair follicles in alopecia areata. J Invest Dermatol 1985; 85: 569-72.

20 Cotsarelis G, Sun TT, Lavker RM. Label-retaining cells reside in the bulge area of pilosebaceous unit: implications for follicular stem cells, hair cycle, and skin carcinogenesis. Cell 1990; 61: 1329-37.

21 Sundberg JP, Cordy WR, King Jr LE. Alopecia areata in aging C3H/HeJ mice. J Invest Dermatol 1994; 102: 847-56.

22 Michie HJ, Jahoda CA, Oliver RF, et al. The DEBR rat: an animal model of human alopecia areata. Br J Dermatol 1991; 125: 94-100.

23 Smyth Jr JR, McNeil M. Alopecia areata and universalis in the Smyth chicken model for spontaneous autoimmune vitiligo. J Investig Dermatol Symp Proc 1999; 4: 211-5.

24 Chan LS. Animal models of human inflammatory skin diseases. Boca Raton: CRC Press, 2004.

25 McElwee KJ, Boggess D, King Jr LE, et al. Experimental induction of alopecia areata-like hair loss in C3H/HeJ mice using full-thickness skin grafts. J Invest Dermatol 1998; 111: 797-803.

26 McElwee K, Freyschmidt-Paul P, Ziegler A, et al. Genetic susceptibility and severity of alopecia areata in human and animal models. Eur J Dermatol 2001; 11: 11-6.

27 Shellow WV, Edwards JE, Koo JY. Profile of alopecia areata: a questionnaire analysis of patient and family. Int J Dermatol 1992; 31: 186-9.

28 Muller SAWR. Alopecia areata: an evaluation of 736 patients. Arch dermatol 1963; 88: 290-7.

29 Milgraum SS, Mitchell AJ, Bacon GE, et al. Alopecia areata, endocrine function, and autoantibodies in patients 16 years of age or younger. J Am Acad Dermatol 1987; 17: 57-61.

30 Friedmann PS. Alopecia areata and auto-immunity. Br J Dermatol 1981; 105: 153-7.

31 Main RA, Robbie RB, Gray ES, et al. Smooth muscle antibodies and alopecia areata. Br J Dermatol 1975; 92: 389-93.

32 Colombe BW, Price VH, Khoury EL, et al. HLA class II antigen associations help to define two types of alopecia areata. J Am Acad Dermatol 1995; 33: 757-64.

33 Welsh EA, Clark HH, Epstein SZ, et al. Human leukocyte antigen-DQB1*03 alleles are associated with alopecia areata. J Invest Dermatol 1994; 103: 758-63.

34 Morling N, Frentz G, Fugger L, et al. DNA polymorphism of HLA class II genes in alopecia areata. Dis Markers 1991; 9: 35-42.

35 Friedmann. Clinical and immunologic associations of alopecia areata. Semin Dermatol 1985; 4: 9-15.

36 Frentz G, Thomsen K, Jakobsen BK, et al. HLA-DR4 in alopecia areata. J Am Acad Dermatol 1986; 14: 129-30.

37 Odum N, Morling N, Georgsen J, et al. HLA-DP antigens in patients with alopecia areata. Tissue Antigens 1990; 35: 114-7.

38 Orecchia G, Belvedere MC, Martinetti M, et al. Human leukocyte antigen region involvement in the genetic predisposition to alopecia areata. Dermatologica 1987; 175: 10-4.

39 Duvic M, Hordinsky MK, Fiedler VC, et al. HLA-D locus associations in alopecia areata. DRw52a may confer disease resistance. Arch Dermatol 1991; 127: 64-8.

40 de Andrade M, Jackow CM, Dahm N, et al. Alopecia areata in families: association with the HLA locus. J Investig Dermatol Symp Proc 1999; 4: 220-3.

41 Akar A, Orkunoglu E, Sengul A, et al. HLA class II alleles in patients with alopecia areata. Eur J Dermatol 2002; 12: 236-9.

42 Colombe BW, Lou CD, Price VH. The genetic basis of alopecia areata: HLA associations with patchy alopecia areata versus alopecia totalis and alopecia universalis. J Investig Dermatol Symp Proc 1999; 4: 216-9.

43 Baadsgaard O, Lindskov R, Clemmensen OJ. In situ lymphocyte subsets in alopecia areata before and during treatment with a contact allergen. Clin Exp Dermatol 1987; 12: 260-4.

44 Bodemer C, Peuchmaur M, Fraitaig S, et al. Role of cytotoxic T cells in chronic alopecia areata. J Invest Dermatol 2000; 114: 112-6.

45 Gilhar A, Landau M, Assy B, et al. Mediation of alopecia areata by cooperation between CD4+ and CD8+ T lymphocytes: transfer to human scalp explants on Prkdc(scid) mice. Arch Dermatol 2002; 138: 916-22.

46 Freyschmidt-Paul P, Hamm H, Happle R, et al. Pronounced perifollicular lymphocytic infiltrates in alopecia areata are associated with poor treatment response to diphencyprone. Eur J Dermatol 1999; 9: 111-4.

47 Bystryn JC, Tamesis J. Immunologic aspects of hair loss. J Invest Dermatol 1991; 96: 88S-89S.

48 Hoffmann R. The potential role of cytokines and T cells in alopecia areata. J Investig Dermatol Symp Proc 1999; 4: 235-8.

49 Arca E, Musabak U, Akar A, et al. Interferon-gamma in alopecia areata. Eur J Dermatol 2004; 14: 33-6.

50 Kalish RS, Johnson KL, Hordinsky MK. Alopecia areata. Autoreactive T cells are variably enriched in scalp lesions relative to peripheral blood. Arch Dermatol 1992; 128: 1072-7.

51 Kalish RS, Johnson KL. Enrichment and function of urushiol (poison ivy)-specific T lymphocytes in lesions of allergic contact dermatitis to urushiol. J Immunol 1990; 145: 3706-13.

52 Gilhar A, Krueger GG. Hair growth in scalp grafts from patients with alopecia areata and alopecia universalis grafted onto nude mice. Arch Dermatol 1987; 123: 44-50.

53 Nunzi E, Hamerlinck F, Cormane RH. Immunopathological studies on alopecia areata. Arch Dermatol Res 1980; 269: 1-11.

54 Galbraith GM, Miller D, Emerson DL. Western blot analysis of serum antibody reactivity with human melanoma cell antigens in alopecia areata and vitiligo. Clin Immunol Immunopathol 1988; 48: 317-24.

55 Tobin DJ, Orentreich N, Fenton DA, et al. Antibodies to hair follicles in alopecia areata. J Invest Dermatol 1994; 102: 721-4.

56 Tobin DJ, Hann SK, Song MS, et al. Hair follicle structures targeted by antibodies in patients with alopecia areata. Arch Dermatol 1997; 133: 57-61.

57 Spickett G, Prentice AG, Wallington T, et al. Alopecia totalis and vitiligo in common variable immunodeficiency. Postgrad Med J 1991; 67: 291-4.

58 Gilhar A, Pillar T, Assay B, et al. Failure of passive transfer of serum from patients with alopecia areata and alopecia universalis to inhibit hair growth in transplants of human scalp skin grafted on to nude mice. Br J Dermatol 1992; 126: 166-71.

59 Dudda Subramanya RX, Luettich Z, King K, Sinha H. Transcriptional profiling in alopeia areata defines disease specific signature, and implicates immune response, cell cycle control, and apoptosis related genes in disease pathogenesis. 2003; AA Submitted for publication.

60 Gilhar A, Ullmann Y, Berkutzki T, et al. Autoimmune hair loss (alopecia areata) transferred by T lymphocytes to human scalp explants on SCID mice. J Clin Invest 1998; 101: 62-7.

61 Randall VA. Is alopecia areata an autoimmune disease? Lancet 2001; 358: 1922-4.

62 Paus R, Slominski A, Czarnetzki BM. Is alopecia areata an autoimmune-response against melanogenesis-related proteins, exposed by abnormal MHC class I expression in the anagen hair bulb? Yale J Biol Med 1993; 66: 541-54.

63 Guin JD, Kumar V, Petersen BH. Immunofluorescence findings in rapid whitening of scalp hair. Arch Dermatol 1981; 117: 576-8.

64 Gilhar A, Landau M, Assy B, et al. Melanocyte-associated T cell epitopes can function as autoantigens for transfer of alopecia areata to human scalp explants on Prkdc(scid) mice. J Invest Dermatol 2001; 117: 1357-62.

65 Tobin DJ, Fenton DA, Kendall MD. Ultrastructural observations on the hair bulb melanocytes and melanosomes in acute alopecia areata. J Invest Dermatol 1990; 94: 803-7.

66 Hordinsky M. Alopecia Areata. In: Olsen E, editor. Disorders of Hair Growth: Diagnosis and Treatment. 1st edn, New York: McGraw-Hill, Inc, 1994: 195-222.

67 Tobin DJB. J. Immunobiology of alopecia areata. In: Camacho Francisco MRVA, Price Vera H, editors. Hair and its Disorders: Biology, Pathology and Management. London: Martin Dunitz Ltd, 2000: 187-201.

68 Becker JC, Varki N, Brocker EB, et al. Lymphocyte-mediated alopecia in C57BL/6 mice following successful immunotherapy for melanoma. J Invest Dermatol 1996; 107: 627-32.

69 Chan LS, Vanderlugt CJ, Hashimoto T, et al. Epitope spreading: lessons from autoimmune skin diseases. J Invest Dermatol 1998; 110: 103-9.

70 Warnock MG, Goodacre JA. Cryptic T-cell epitopes and their role in the pathogenesis of autoimmune diseases. Br J Rheumatol 1997; 36: 1144-50.

71 Walker LS, Abbas AK. The enemy within: keeping self-reactive T cells at bay in the periphery. Nat Rev Immunol 2002; 2: 11-9.

72 Ohashi PS. T-cell signalling and autoimmunity: molecular mechanisms of disease. Nat Rev Immunol 2002; 2: 427-38.

73 Rose NR. Mechanisms of autoimmunity. Semin Liver Dis 2002; 22: 387-94.

74 McElwee KJ, Hoffmann R, Freyschmidt-Paul P, et al. Resistance to alopecia areata in C3H/HeJ mice is associated with increased expression of regulatory cytokines and a failure to recruit CD4+ and CD8+ cells. J Invest Dermatol 2002; 119: 1426-33.

75 Zoller M, McElwee KJ, Engel P, et al. Transient CD44 variant isoform expression and reduction in CD4(+)/CD25(+) regulatory T cells in C3H/HeJ mice with alopecia areata. J Invest Dermatol 2002; 118: 983-92.

76 Skinner Jr RB, Light WH, Bale GF, et al. Alopecia areata and presence of cytomegalovirus DNA. Jama 1995; 273: 1419-20.

77 Christoph T, Muller-Rover S, Audring H, et al. The human hair follicle immune system: cellular composition and immune privilege. Br J Dermatol 2000; 142: 862-73.

78 Hoffmann R, Eicheler W, Huth A, et al. Cytokines and growth factors influence hair growth in vitro. Possible implications for the pathogenesis and treatment of alopecia areata. Arch Dermatol Res 1996; 288: 153-6.

79 Hoffmann R, Wenzel E, Huth A, et al. Cytokine mRNA levels in Alopecia areata before and after treatment with the contact allergen diphenylcyclopropenone. J Invest Dermatol 1994; 103: 530-3.

80 Kalish RS, Gilhar A. Alopecia areata: autoimmunity--the evidence is compelling. J Investig Dermatol Symp Proc 2003; 8: 164-7.

81 Norris DA. How close are we to solving the puzzle? Review of the Alopecia Areata Research Workshop David Norris. J Investig Dermatol Symp Proc 2003; 8: 222-5.