Diphencyprone immunotherapy alters anti-hair follicle antibody status in patients with alopecia areata

ARTICLE

Alopecia areata (AA) is a common cause of reversible hair loss afflicting approximately 1-2% of the general population [1] and it is also expressed in several non-human mammals [2]. It is commonly manifested by patchy areas of hair loss on the scalp and other body parts, but can progress to complete loss of all body hair. While not life threatening, the disease is nonetheless serious because it is disfiguring and in humans can cause severe psychological and social problems including loss of employment.

The classic histology of AA - a dense peribulbar lymphocytic infiltrate primarily affecting early anagen hair follicles (HFs) - is one of the most consistent and reproducible immunological abnormalities in AA [3]. While it has been known for over a hundred years that dystrophic hair follicles (HF) in AA are associated with a leukocytic infiltrate, a paper by Van Scott [4] formally raised the possibility that AA was perhaps an autoimmune disorder. Much progress has been made recently in dissecting the etiologic factor(s) involved in AA and it is now generally assumed to be an autoimmune disease in which the HF is the primary target. Several of the requirements for ascribing autoimmune status [5] to AA now appear to be place. Not only can the disease be passively transferred by T cells [6], serum IgG may also disturb hair cycling [7]. Moreover, there is evidence that hair bulb melanocytes may express relevant target antigens [8, 9]. Despite this the epidermis (also containing melanocytes) is generally unaffected in AA. This may be explained by the observation that melanocytes of the skin and HF are antigenically distinct [10]. Moreover, it appears that the autoimmune response to HFs produces rather than follows the disease process [6, 11]. Other indirect clues for autoimmunity include the association of the disease with a particular HLA haplotype(s), other autoimmune diseases, and a responsitivity to immunosuppression therapies [3].

The biologic relevance of the HF antibodies in AA remains to be determined. A basic question is whether anti-HF antibodies participate in the pathogenesis of AA or are merely a result of AA-associated tissue damage. In this regard, it is of note that high titer antibodies are not observed in normal individuals despite the release of HFs antigens during the normal HF cycle or even in scarring alopecias [12, 13]. Moreover, the abnormal autoantibody response to HFs in the C3H/HeJ mouse model of AA is present both in affected mice and to a lesser degree in their, as yet, clinically unaffected littermates [14]. This may suggest that the presence of antibodies to HFs appear before the onset of hair loss and so may not, in this case, be produced as a secondary response to HF damage in AA. Finally, we have recently shown, in a preliminary study, that purified IgG from an AA-affected horse adversely affected hair regrowth when passively transferred to normal mice. While passively transferred "AA" IgG antibodies did not induce hair loss in this model, HFs in treated skin were retained in telogen while HFs located distant apparently cycled normally [7]. That preliminary study however, should be interpreted in light of an earlier study that reported the failure of passive transfer of whole serum from human patients with AA to inhibit hair growth in human scalp skin grafted onto nude mice [15].

Topical immunotheraphy with contact sensitizers is at present the most effective therapeutic modality in AA [16-22]. Squaric acid dibutylester (SADBE) and diphencyprone (DCP) are preferred as they are not mutagenic, infrequently used in industry, absent in the environment and do not cross react with other contact allergens [22]. Thus, it is of interest to elucidate whether immunotherapy-induced hair regrowth is associated with any alteration in titer of circulating serum anti-HF antibodies in AA patients that may be modulated during the course of the disease or more specifically, during hair regrowth. The latter scenario may be expected if the expression/presentation of antibody-eliciting HF antigens is modulated by the functional modification of the infiltrate during the immunotherapy. A reduction in anti-HF antibody titer and pattern of reactivity before/during early hair regrowth in treated AA patients may suggest a central role for anti-HF antibodies in the mediation of hair loss. Similarly, a reduction in anti-HF antibody titer and reactivity occurring only after more complete treatment may indicate a secondary role for anti-HF autoantibodies in the pathogenesis of AA.

The present study was designed to examine the status of anti-HF antibodies during topical immunotherapy with the contact sensitizer DCP. The following questions were asked: 1) Does the pattern and titer of anti-HF IgG antibody reactivity to native HF proteins (i.e. in frozen scalp sections) alter during and after DCP-induced hair regrowth? 2) Is the pattern and titer of anti-HF antibody reactivity to 6 M urea extractable HF antigens altered as a result of DCP-induced hair regrowth. These questions were examined by use of indirect immunofluorescence and immunoblotting assays respectively.

Materials and methods

Patients: Eleven patients with AA of the scalp (AA Universalis n = 6; AA Totalis n = 1; AA Patchy n = 3; 9 females, 3 males, 10-54 yrs, mean 29 yrs) were included in this study (Table I). The duration of AA ranged from 5 months to 25 years (mean 9.6 yrs), while age of onset of AA ranged from 1 to 42 years of age (mean 16.8 yrs). Treatment was performed as previously described [17, 18, 22]. Patients were first sensitised with 2% diphencyprone (DCP) followed by weekly unilateral maintenance of 0.001-2% DPC in acetone on one side of the scalp until a mild contact dermatitis was obtained. All patients exhibited unilateral hair regrowth.

Serum: Serum was obtained by venipuncture from each patient before the start of therapy [BT; n = 11], after unilateral hair regrowth [ULR; n = 11], during continuing hair regrowth [CT; n = 6] and in some cases after complete regrowth [CP; n = 4] (Table I). Time intervals between BT and after ULR serum samples ranged from 3-9 months (mean, 5 months), between BT and CT serum samples ranged from 8-13 months (mean, 11 months), and between BT and CT ranged from 9-24 months (mean, 16 months).

Detection of anti-hair follicle antibodies

Indirect immunofluorescence assay: Seven micrometer cryosections of normal haired scalp (male, 35 yrs) were air-dried, blocked for 1 hr in 10% goat serum, incubated with test patient serum (diluted 1:100) in 1% goat serum for 1 hr at room temperature. Following washes in phosphate buffered saline (PBS), sections were incubated with fluorescein-conjugated goat IgG fraction to human IgG (ICN Biomedical Inc., Aurora, OH) diluted 1:200 in 1% goat serum for 1 hr. Following washes in PBS, the sections were mounted in Vectashield medium with 4,6-diamidino-2-phenylindole (DAPI) (Vector Labs. Ltd, Peterborough, UK). Negative control sections were incubated with PBS in place of the primary serum. The slides were read by fluorescence microscopy (Leica Microsystems, Germany). Staining intensity was measured on a scale of - (no staining) to +++ (strong staining). Staining was assessed for epidermis and individual HF components including; outer root sheath (ORS), matrix (Mx), inner root sheath (IRS), pre-cortex or keratogeneous zone (PC), hair shaft (HF) and follicular papilla (FP).

Western blotting assay: HF antigens were prepared as previously described [12]. Briefly, anagen VI HFs were plucked from occipital scalp of a normal healthy 36 years old male and immediately incubated over night in 6 M urea with a cocktail of protease inhibitors (Sigma, Dorset, UK). HF particulates were removed by centrifugation. The soluble HF antigens were then separated by SDS-8%-PAGE and electroblotted onto polyvinylidene difluoride microporous membranes. After blocking in 5% non-fat milk in PBS, individual membrane strips were incubated overnight with whole patient serum (diluted 1:100) incubated in biotin-conjugated goat affinity purified antibody to human IgG (diluted 1:100) (Organon Teknika Corp., Westchester, PA), followed by peroxidase-conjugated avidin (diluted 1:100) (Organon Teknika Corp., Westchester, PA) and developed with 4-chloronapthol (Sigma, Dorset, UK). The molecular weights of positive bands on the resulting immunoblots were calculated using a calibration curve drawn from the separation of prestained protein makers (New England BioLabs Inc.). Staining intensity was measured on a scale of - (no staining) to +++ strong staining.

Results

Anti-HF antibody reactivity to defined components of scalp anagen hair follicles is altered during DCP-induced hair regrowth in AA

Anti-HF antibody status before treatment: Circulating anti-HF IgG antibodies were detected in sera from all 11 patients before treatment with DCP (Table II). Serum antibody reactivity was directed predominantly to the lower HF and hair bulb in all patients and particularly to keratinocytes of the most proximal bulb (Figs. 1-3, Table II). In some cases, serum antibodies reacted weakly with keratinocytes in the epidermis but did not react with keratinocytes of the upper HF. In one patient (Tables I/II, patient #3), antibody reactivity was also detected to follicular papilla cells (Fig. 3).

Anti-HF antibody status after unilateral hair regrowth: The pattern and titer of anti-HF antibody reactivity present in all sera after unilateral hair regrowth was similar to that before therapy (Table II). This second serum sample was taken at least 3 months, and in some cases 9 months, after the start of therapy. All patient samples revealed either no change in pattern/titer of anti-HF antibody reactivity or even exhibited an increased antibody titer to some HF components, especially the IRS and PC (Table II).

Anti-HF antibody status after continued but incomplete hair regrowth: Serum was available from the period of continued hair bilateral hair regrowth from 6 of 11 patients. Compared with anti-HF antibody status before therapy, three patients continued to exhibit a similar titer and pattern of anti-HF antibody reactivity. However, three others showed a reduction in these two antibody parameters. In this latter patient sub-group the reduction in anti-HF antibody titer was observed primarily in reactivity to the HF IRS and pre-cortex. Further reduction in anti-HF antibody reactivity was apparent in those patients for whom a second serum sample (taken 2 months after the first) was available from this period of continued, but incomplete, hair regrowth (Tables I/II patients #2 & #5).

Anti-HF antibody status after complete and sustained hair regrowth: Serum samples were available from 4 patients after they exhibited DCP-induced complete and sustained hair regrowth (Tables I/II, patients #1-4). Thus, these sera may provide more complete information on the modulation of HF autoreactivity and the related and resultant elicitation of anti-HF antibodies associated with the modulation of the HF infiltrate during DCP treatment. Anti-HF antibody reactivity in all of these samples was reduced compared to the level before therapy (Tables I/II, Fig. 3), although some low titer residual reactivity remained. Notably, residual anti-HF antibody reactivity after complete hair regrowth was in all cases restricted to the most proximal hair bulb matrix.

Anti-HF antibody reactivity to 6 M urea extractable HF antigens is altered during DCP-induced hair regrowth in AA

The HF antigen substrate used for the immunoblotting analysis of AA serum antibodies was prepared using 6 M urea extractable proteins of anagen scalp HFs [12]. Serum antibody reactivity to HF antigens were detected in all 11 patients (Table II), although there was significant inter-patient differences in both titer of anti-HF antibody titers and the pattern of HF antigens targeted. Confirming previous studies [7, 12, 14], antibody reactivities could be grouped to three major molecular weight regions; ± 40-60 kD, ± 60-75 kD and ± 90-120 kD. Strong reactivities were observed to particular antigen bands of 40, 47, 50, 57, 68, 90 and 102 kDa. Moreover, the immunoblotting data correlated closely with the indirect immunofluorescence data, i.e. sera titers on frozen scalp tissue and by immunoblotting were similar (Table II e.g. patient #3, Figs. 1-3).

Anti-HF antibody status before therapy and during continuing but incomplete DCP-induced hair regrowth: For the majority of patients there was little evidence that anti-HF antibody reactivities decreased as a function of time with continued, though incomplete, hair growth (Table II, Fig. 1). As with the indirect immunofluorescence, some patients exhibited similar or even higher antibody titers in sera taken during this period of continuing regrowth (Table II, Fig. 2). Notably however, a reduction in anti-HF antibody was seen when the before therapy sample was compared with the later of two serum samples taken during the period of continued, but incomplete, hair regrowth (Table II, patient # 2 and 5).

Anti-HF antibody status before therapy vs complete DCP-induced hair regrowth: All 4 of 4 patients exhibited a further reduction in the presence and titer of anti-HF antibodies by the time the final serum sample was taken during a period of sustained and complete hair regrowth. As was found by indirect immunofluorescence analysis, some residual reactivity remained in all cases. This was restricted mainly to antigen bands of 47, 50 and 57 kDa (Table II, Fig. 3).

Discussion

This study has shown that anti-HF antibody reactivity in AA patients is significantly altered after DCP-induced hair regrowth. A striking reduction was detected in both the titer and range of HF components/antigens targeted by anti-HF IgG antibodies particularly in those patients that exhibited complete and sustained hair regrowth after DCP-treatment. This change represents a significant modulation of the AA immune response to antigens expressed on HFs during hair regrowth. Whether this downregulation of antibody-production is a cause or only a result of hair regrowth may be answered by a closer look at the changes in anti-HF antibody-titer and range of HF components/antigens targeted during the course of treatment. Unilateral hair regrowth was associated with no change, or even an increase, in anti-HF antibody titer and reactivity was observed in all patients. This finding may reflect the fact that partial hair regrowth is unlikely to be associated with the complete removal of the stimulus for the elicitation of the anti-HF antibodies, i.e. AA being still active in the un-haired part of the scalp. Therefore, we can conclude that the downregulation of anti-HF antibodies is likely to be a result rather than the cause of hair regrowth induction by topical immunotherapy.

The clinical observation that part of the scalp may exhibit active hair loss while hair close by may be regrowing, strongly suggests that the immune responses that drive active hair loss and regrowth are likely to be very local to the affected HFs themselves. Immunohistochemical investigations in humans and mice [23-25] revealed that topical immunotherapy does indeed change this local immune response: CD4⁺ and CD8⁺ T cells are not removed but their composition is rather changed after treatment with DCP or SADBE. In humans the CD4/CD8 ratio changed from 4:1 or 2:1 before treatment to 1:1 after treatment. Furthermore, the aberrant expression of MHC class I and II molecules on hair follicle epithelium is strikingly reduced after treatment in both C3H/HeJ "AA" mice and humans with AA [24]. The observation of an oligoclonal T-cell receptor repertoire in AA [26, 27] has indicated that classical activation of T cells with antigen-presentation by MHC class molecules may be involved in pathogenesis. Hence, the reduced MHC class I and II expression after topical immunotherapy may lead to a decreased antigen-presentation and subsequent reduction in the production of anti-HF antibodies by B cells.

The pattern of anti-HF antibody reactivity and antibody titer exhibited significant change over the duration of the current study. Hair regrowth is likely to be associated with a reduction in the expression or presentation of particular HF epitopes that induce the production of anti-HF antibodies. Notably, a most striking change was apparent in antibody reactivity to hair bulb tissue components and to antigens in the 45-65 kDa molecular weight range. These latter antigens include HF-specific keratins and keratin-related proteins [28]. This alteration may be explained in part by the fact that the half-life of pre-formed IgG antibodies is 2-3 months and that sera in this study was collected over a period of up to 16 months. Thus, altered anti-HF autoreactivity is likely to be due to changes in the presentation of HF antigens during the DCP-induced immuno-modulation and hair regrowth.

As long as there is a T cell infiltrate targeting the HF at any part of the scalp or body and MHC class expression on HF-epithelium is not sufficiently down-regulated by treatment, B cells will be stimulated to produce anti-HF antibodies which consequently circulate and will be detectable. Even in the 4 patients with complete and sustained hair regrowth treatment needed to be continued to prevent a relapse. Although the disease activity was not measured in those patients with complete hair regrowth, and continued to be treated by topical immunotherapy, clinical experience has shown that a sudden stop of treatment immediately after complete hair regrowth leads to new hair loss in most cases. Therefore, it can be assumed that a T cell infiltrate with concomitant MHC class expression is still present in some HF, but is kept under control by continuous topical immunotherapy. These residual infiltrates may be responsible for low level antigen-presentation that gives rise to the decreased, but still present, anti-HF antibody titer and reactivity in our patients after complete hair regrowth.

It is thought that the immune response in AA is directed to antigen(s) expressed by normal HF [3]. This expression of antigen appears to be hair cycle-dependent, as mid to full anagen HFs appear to be most vulnerable, i.e. are preferentially targeted [29]. This hair cycle-specific feature of AA activity has lead to the prediction that target proteins may be expressed in the reconstructing and activating HF pigmentary unit in anagen HF [30]. Indeed, it appears that T lymphocytes and antibodies in AA patients may target melanocytic antigens [8, 9]. However, the exact nature of the antigen(s) and epitopes, presented inappropriately to T lymphocytes in the abnormal AA immune response, has yet to be elucidated. Anti-HF antibody specificities may serve as useful probes for the eventual identification and characterization of the AA target antigen(s). That immuno-modulation does not reduce the titer of the full spectrum of anti-HF antibodies detectable during the untreated phase may suggest that the HF express some immunodominant "AA-relevant" epitopes. Whether circulating anti-HF antibodies similarly target all scalp HFs, i.e. those undergoing active hair loss as well as those in active regrowth is unclear, but this is likely in the absence of evidence for antibody partition in AA scalp. In any event, the co-expression of hair regrowth and continued presence/production of anti-HF antibodies suggests that hair regrowth is possible despite the presence of some antibody specificities, and so it is unlikely that the antibodies mediate hair loss in the absence of HF-reactive T lymphocytes.

The current study supports the view that that DCP-induced modification of the peri/intrafollicular infiltrate may not ultimately alter the presentation of HF antigens that elicit antibody production, i.e. HF target antigens are still presented even after successful treatment and after hair regrowth. One hypothesis to explain these findings may be that AA is mediated by a T cell inflammation with classical activation of T cells by presentation of an HF-antigen via MHC class I or II molecules. This response may generate cytokine production and Fas-FasL interaction [31]. The later interactions may then damage the HF or at least alter its cycling pattern. With time the activated T-cells then switch from a Th1 to Th2 profile during the course of the disease and induce production of antibodies by B cells. The time sequence of these events may be important, given that anti-HF antibodies may be detected before clinical evidence of hair loss. Antibodies may remain for as long as the HF antigens continue to be presented i.e. independent of presence or absence of the T cell infiltrate.

CONCLUSION

In summary, this study has shown that immuno-therapy in AA is associated with a reduction in the titer and pattern of reactivity of anti-HF antibodies. These antibodies may therefore hold the key to the identity of HF antigen targets in AA, as they are likely to be similar (if not identical) with those targeted by the autoreactive T lymphocytes [32]. While anti-HF antibodies may not "cause" AA, their presence does appear to correlate with the presentation of pathogenic HF antigens. Moreover, the presence of anti-HF antibodies may also correlate with the status/activity of anti-HF autoreactive T-cells, in that these antibody reactivities are reduced or disappear after DCP-induced modulation of the T cell infiltrate. While there is increasing evidence that the presence of T cells correlates with acute disease [6], the presence/titer of anti-HF antibodies may be a marker of clinical disease activity or opportunity for spontaneous regrowth. Why the expression of HF antigens, or the stimulus that elicits the production of anti-HF antigen antibodies, may disappear in some patients (indicated by disappearance of antibodies) is unknown. We are currently attempting to identify the HF antigens targeted by the immune response in AA and are studying different strategies including screening cDNA HF libraries with sera from AA-affected C3H/HeJ mice. These studies may help to elucidate the nature of the involvement of antibodies in AA and identify the pathogenic potential of specific T cell clones that react with the identified HF target antigens.

Article accepted on 4/1/01

REFERENCES

1. Safavi KH, Muller SA, Suman VJ, Moshell AN, Melton LJ. Incidence of alopecia areata in Olmsted County, Minnesota, 1975-1989. Mayo Clin Proc 1995; 70: 628-33.

2. McElwee KJ, Boggess D, Olivry T, Oliver RF, Whiting D, Tobin DJ, Bystryn JC, King LE, Sundberg JP. Comparison of alopecia areata in human and non-human mammalian species. Pathobiol 1998; 66: 90-107.

3. Tobin DJ, Bystryn JC Immunology of alopecia areata. In: Hair and hair disorders: research, pathology and management. FM Camacho, VA Randall, Price V (eds), Martin Dunitz, London, 2000, 187-201.

4. Van Scott EJ. Morphologic changes in pilosebaceous units and anagen hairs in alopecia areata. J Invest Dermatol 1958; 31: 35-43.

5. Rose NR. Characteristics of autoimmune disease. J Invest Dermatol 1991; 96; 87S.

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

7. Tobin DJ, Alhaidari Z, Olivry T. Equine alopecia areata autoantibodies target multiple hair follicle antigens and may alter hair growth in passive transfer studies. Exp Dermatol 1998; 7: 289-97.

8. 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.

9. Gilhar A, Landau M, Assy B, Shalaginov R, Serafimovich S, Kalish RS. 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.

10. Tobin DJ, Bystryn JC. Different populations of melanocytes are present in hair follicles and epidermis. Pigment Cell Res 1996; 9: 304-10.

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

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

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

14. Tobin DJ, Sundberg JP, King LE Jr, Boggess D, Bystryn JC. Autoantibodies to hair follicles in C3H/HeJ mice with alopecia areata-like hair loss. J Invest Dermatol 1997; 109: 329-33.

15. Gilhar A, Pillar T, Assay B, David M. 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.

16. Shapiro J, Price V. Hair regrowth. Therapeutic agents. Dermatol Clin 1998; 16: 341-56.

17. Hoffmann R, Happle R. Topical immunotherapy in alopecia areata: What, How and Why? Dermatol Clin 1996; 14: 739-44.

18. Happle R, Kalveram KJ, Buchner U, Echternacht-Happle K, Göggelmann W, Summer KH. Contact allergy as a therapeutic tool for alopecia areata: application of squaric acid dibutylester. Dermatologica 1980a; 161: 289-97.

19. Case PC, Mitchell AJ, Swanson NA, van der Veen EE, Ellis CN, Headington JT. Topical therapy of alopecia areata with squaric acid dibutylester. J Am Acad Dermatol 1984; 3: 447-51.

20. Caserio RJ. Treatment of alopecia areata with squaric acid dibutylester. Arch Dermatol 1987; 123: 1036-41.

21. Micali G, Cicero RL, Nasca MR, Sapuppo A. Treatment of alopecia areata with squaric acid dibutylester. Pharmacol Therapeut 1996; 35: 52-6.

22. Freyschmidt-Paul P, Hoffmann R, Levine E, Sundberg JP, Happle R, McElwee KJ. Current and potential agents for the treatment of alopecia areata. Curr Pharm Des 2001; 7: 213-30.

23. Happle R, Klein HM, Macher E. Topical immunotherapy changes the composition of the peribulbar infiltrate in alopecia areata. Arch Dermatol Res 1986; 278: 214-8.

24. Bröcker EB, Echternacht-Happle K, Hamm H, Happle R. 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.

25. Freyschmidt-Paul P, Sundberg JP, Happle R, McElwee KJ, Metz S, Boggess D, Hoffmann R. Successful treatment of alopecia areata-like hair loss with the contact sensitizer squaric acid dibutylester (SADBE) in C3H/HeJ mice. J Invest Dermatol 1999; 113: 61-8.

26. Dressel D, Brutt CH, Manfras B, Zollner TM, Wunderlich A, Bohm BO, Boehncke WH. Alopecia areata but not androgenetic alopecia is characterized by a restricted and oligoclonal T cell receptor-repertoire among infiltrating lymphocytes. J Cutan Pathol 1997; 24: 164-8.

27. Hoffmann RHA, Wenzel E, Steen PHM, Schäufele M, Raykowski S, Happle R. T-cell receptor Vb chain variable gene use in alopecia areata totalis. Eur J Dermatol 1996; 6: 65-9.

28. Tobin DJ, Bystryn JC. Alopecia areata is associated with antibodies to hair follicle-specific antigens located predominantly in the proliferative region of hair follicles. In: Neste DJJ, Randall VA, eds. Hair research for the next millennium. Amsterdam : Elsevier Science B.V., 1996: 237-41.

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

30. 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.

31. Freyschmidt-Paul P, McElwee KJ, Botchkarev V, Kissling S, Wenzel E, Sundberg JP, Happle R, Hoffmann R. Fas-deficient C3MRL-Tnfrsf6^lpr mice and FasL-deficient C3H/HeJ-Tnfsf6^gld mice are relatively resistant to the induction of alopecia areata by grafting of alopeic areata-affected skin from C3H/HeJ mice. J Invest Dermatol Symp Proc (in press).

32. Falcone M, Lee J, Patkstone G, Yeung B, Sarvetnick N. B lymphocytes are crucial antigen-presenting cells in the pathogenic autoimmune response to GAD65 antigen in non obese diabetic mice. J Immunol 1998; 161: 1163-8.