Interferon‐gamma in alopecia areata

ARTICLE

Auteur(s) : Ercan ARCA, Ugur MUSABAK, Ahmet AKAR, A. Hakan ERBIL, H. Blent TATAN

Department of Dermatology and ¹ Immunology, Glhane Military Medical Faculty, School of Medicine Etlik, 06018 Ankara, Turkey

Article accepted on 20/11/2003

Alopecia areata (AA) is a common type of hair loss with a lifetime risk of 1.7% in the general population [1]. The severity of the disease is variable and the clinical hallmark is nonscarring, patchy hair loss, which may recover spontaneously but may also progress to long lasting chronic alopecia totalis (AT) and alopecia universalis (AU) [2]. Although the pathogenesis of alopecia areata is poorly understood, evidence is accumulating to suggest that T cells and cytokines play an important role [3]. The immune response present in AA is associated with aberrant lesional expression of interferon-γ (IFN-γ), interleukin-2 (IL-2) and IL-1β, and overexpression of ICAM-1 and MHC molecules on hair follicle keratinocytes and dermal papilla cells [4]. If the patterns of cytokine release, for example IFN-γ, are causally related to the phenotype of disease expression, one would expect that the profile of a cytokine like IFN-γ profile in the localized form may be different from that in the extensive form [5].
In this study we measured the serum level of IFN-γ by enzyme immunoassay (EIA) technique in patients with the localized form and the extensive form (AT, AU or AT/AU) of AA, to investigate the role of IFN-γ in the pathogenesis of AA and to clarify if there was a different profile between localized and extensive form.

Material and methods

Forty patients with AA were included in this study. Nineteen patients had localized AA and twenty-one patients had AT, AU or AT/AU. The patients were characterized according to alopecia areata investigational assessment guidelines [6]. No patients had used any systemic medications for AA for at least 3 weeks or therapies such as PUVA, which could have a prolonged effect on IFN-γ levels, for at least 6 months before this study, and also patients who had other types of illness such as autoimmune diseases that could affect the outcome of the study and those who had received treatment with systemic steroids and other immunosuppressive medications. Twenty healthy controls who were age and sex matched were studied. The total group of AA patients as well as two subgroups according to scalp hair loss were compared to the control group. First group (n = 40) included all AA patients, second group (n = 19) was the patients with less than 99% scalp hair loss and third group (n = 21) was the patients with AT, AU or AT/AU.
The serum samples obtained for the assay of IFN-γ were stored at – 80°C before the assay. The levels of IFN-γ (Cytelisa, Human IFN-γ) in sera of patients and controls were measured by using enzyme immunoassay (EIA). The kit was from Cytimmune Sciences INC., Collage Park, Maryland. The range of detection was 8 pg/mL to 500 pg/mL for IFN-γ. The assay was performed in a blind fashion on coded samples by an investigator (UM) who was not informed of the patients' clinical status, after the collection of all samples had been completed.
The data are expressed as mean ± standard deviation. The test distribution was done by one-sample Kolmogorov-Smirnov test, and comparations were performed by independent-samples T-test. The data were considered statistically significant if p values were less than 0.05.

Results

The study group composed of 40 patients with AA (36 men and four female; the mean age of the patients was 23.4 years ranging from 19 to 37 years and 20 normal healthy individuals (18 men and two females; the mean age 22.4 years, ranging from 19 to 28). In the total of 40 patients with AA, 19 of them were LAA (16 men and 3 female; the mean age 23.8 years, ranging from 19 to 37) and the remaining 21 were AT, AU or AT/AU group (20 men and one female; the mean age 23 ranging from 19 to 36). The duration of hair loss in LAA and extensive groups varied from 2 to 156 months (mean 42.7 months) and 30 to 240 months (mean 111.14), respectively. The characteristic data of the patients is shown in Table 1.

The mean serum IFN-γ level in AA patients (n = 40) was 14.25 ± 8.76 pg/mL (mean ± SD), whereas that of LAA (n = 19) or extensive (AT, AU or AT/AU) (n = 21) was 13.45 ± 6.75 pg/mL or 14.98 ± 10.37 pg/mL, respectively. The mean serum IFN-γ level in controls was 9.95 ± 2.6 pg/mL.
Serum levels of IFN-γ in patients with AA were significantly higher than those in controls (t = 2.862, p = 0.006). Significant difference was observed in serum levels of IFN-γ between patients with LAA and control group (t = 2.112, p = 0.046). Serum levels of IFN-γ in patients with AT, AU or AT/AU were significantly higher than those in controls (t = 2.153, p = 0.042). There was no significant difference in levels of IFN-γ between patients with LAA and the extensive group (t = 0.559, p = 0.580).

Discussion

Although the cause of AA is at present unknown, one can now define a common denominator of all clinical forms of AA occurring in all mammalian spices. This is the lymphocytic attack on the lower part of the anagen follicle and ectopic expression of MHC class I and II molecules on the epithelium of affected hair follicles, suggesting local release of cytokines. Hence, two key pathogenetic factors leading to hair loss in AA can be discriminated in the form of cytokines and T-cells. Histologically, the hair bulb is infiltrated and surrounded by mainly T-helper (Th) cells. A consistent feature was the presence of cytokines of the Th1 type (IFN-γ, IL-2, and IL-1β) [1].
The results presented here demonstrate that the mean serum levels of IFN-γ in the sera were significantly elevated in patients with AT or AU. And also the serum levels of IFN-γ in all AA patients were significantly increased. Our findings were similar to the study of Teraki et al. [5]. They compared the serum levels of cytokines, including IFN-γ, TNF-α, IL-1α, IL-2, IL-4, and IL-6 in patients with the localized form and the extensive form and found that the serum levels of IL-1α and IL-4 were significantly elevated in patients with the localized form. In contrast, the serum levels of IFN-γ and IL-2 were significantly elevated in patients with the extensive form. They said that these findings could be interpreted as an indication that Th1 type cytokines might be critical for the progression to the extensive form and that Th2 type cytokines may exert a more subtle influence on the inhibition of a cell-mediated attack on hair follicles [5]. But their results were drawn from a very low patient number (7 patients with LAA, and 7 patients with AU).
With data from rodent model characterization and functional studies, McElwee and Hoffmann [7] described a hypothetical disease development scenario. Organ-specific autoimmune diseases are frequently based on the activation of autoreactive Th1 lymphocyte cells, which may be accompanied by the activation of autoreactive T cytotoxic (Tc) lymphocytes. The cytokine expression profile in chronic AA affected mouse skin does not define an unequivocal Th1 vs. Th2 cytokine dominated state. The monocyte derived inflammatory cytokines TNF-α and IL-6 are expressed at an elevated level, and increased expression of Th1 cytokines IL-12 and IFN-γ as well as Th2 cytokines IL-4 and IL-10 were observed. They stated that both Th1 and Th2 mechanisms are active in chronic mouse AA [7]. Our finding that the serum levels of IFN-γ was significantly elevated in patients with AA and AT, AU, AT/AU indicates that Th1 type may play a role in the pathogenesis of AA, especially in the extensive form.
Reduced levels of IFN-γ and IL-1β and increased lesional expression of IL-10, TGF-β1 and TNF-α have recently been found in scalp biopsies after diphenylcyclopropenone treatment [4]. With regard to MHC molecules, by using dermal papilla cells derived from anagen hair follicles obtained from healthy donors, Konig et al. were able to imitate the in vivo situation of AA by incubating these cells with IFN-γ. They noted a significant increase in HLA-ABC as well as HLA-DR expression within 48 hours. Moreover they found that IFN-γ also led to an overexpression of ICAM-1 [4].
Concerning the expected upregulation of Th1 cytokines, Freyschmidt-Paul et al. found that IFN-γ was indeed significantly upregulated in draining LNC of IL-10^–/– as compared with IL-10^+/+ mice, which in turn accounted for the expression of the proinflammatory cytokine TNF-α [8].
In fact there are several pieces of evidence to indicate the involvement of Th1 type cytokines in inflammatory responses observed in AA; the presence of IFN-γ is inferred on the perifollicular infiltrates, and ICAM-1 and HLA-DR, which are shown to be induced by IFN-γ, are intensely expressed on follicular epithelium [5]. After that, Hoffmann et al. reported that Th1 type cytokine mRNA levels were increased in untreated AT [9]. It was shown that it is possible to transfer alopecia areata to human scalp grafts on CB-17 severe combined immunodeficiency (SCID) mice by the injection of scalp infiltrating T-lymphocytes [10]. T-lymphocytes must first be activated by in vitro culture with hair follicle homogenate in the presence of antigen presenting cell. Culture with nonfollicular scalp homogenate did not induce hair loss [11]. This suggests that AA is mediated by T-cells, which recognize a hair follicle antigen [12]. By using immunohistochemical and in situ hybridization studies to demonstrate the persistence of pro-inflammatory as well as apoptotic mechanisms in the skin biopsies from patients with chronic AA, Bodemer et al. have confirmed the presence of a cellular infiltrate in close contact with the hair follicle, producing IFN-γ in association with pro-inflammatory cytokine production (IL-1β) [2]. In the study of Gilhar et al., the aim was to characterize the cytokine response of infiltrating lymphocytes active in transferring hair loss. The biopsies that were taken from human scalp grafts on SCID mice that had been injected with lymphocytes were analyzed for expression of cytokines by both follicular epithelium and infiltrating T-cells. They found in grafts with hair loss that the predominant cytokine of infiltrating T-cells was IFN-γ, while the follicular epithelium expressed IFN-γ-inducible protein-10kDa, suggesting a Th1 pathogenesis for AA [13]. By examining the expressions in cryostat sections of scalp skin obtained from a total of 28 patients with AA and from five normal control subjects, Sato-Kawamura et al. reported that Cd3⁺ T-cell infiltrates with dominance of CD4⁺ over CD8⁺ T-cells were present perifolliculary and perivasculary in AA lesions; molecules induced by IFN-γ were expressed on the hair follicle in AA lesions, and IFN-γ-producing cells were detected only in AA lesions. They said that these findings suggest that infiltrating T-cells interact with perifollicular or follicular antigen-presenting cells to produce IFN-γ, which deprives dermal papilla cells of their ability to maintain anagen hair growth [14]. Our study provides data supporting the hypothesis that AA is mediated by a Th1 immune response, marked by production of IFN-γ.
Since AA and other autoimmune diseases have similar pathogenesis, there is some information from other autoimmune diseases where the levels of IFN-γ correspond with the disease severity. Psoriasis has similarities to AA in that it also involves a Th1 cell response. Since T cells alone do not produce psoriasis, they must produce signals that induce keratinocyte proliferation. There is a predominance of Th1 cytokines, namely IFN-γ and IL-2 in psoriatic plaques. Treatment with rh IFN-γ administered subcutaneously proved ineffective in psoriatic arthritis, but approximately 25% of treated patients developed foci of psoriasis at the injection sites. This reaction did not occur at the sites of saline injection, implying that IFN-γ is a key cytokine in the pathogenesis process of psoriasis [15]. In the study of Abdel-Naser et al., they studied the punch biopsies from depigmented vitiliginous skin, normal-looking pigmented skin, and marginal skin and found that IL-2 and IFN-γ receptors were clearly expressed by the cellular infiltrate [16].
Potential targets for therapy of AA include interference with CD4⁺ or CD8⁺ T-cell antigen recognition, effectors function, cytokine profiles, antigen presentation, homing and costimulation. There are multiple immunomodulatory agents under development that have a potential application to AA [17]. IFN-γ is an important immune regulator in normal immunity. When IFN-γ production is disturbed, various autoimmune diseases can develop. Anti-IFN-γ has been tested in several Th1 cell autoimmune diseases, including rheumatoid arthritis, multiple sclerosis, uveitis, Type I diabetes, and various autoimmune skin diseases (alopecia areata, psoriasis vulgaris, vitiligo, pemphigus vulgaris and epidermolysis bullosa). A strong, sometimes striking, therapeutic response followed administration of anti-IFN-γ, indicating that it may be a promising therapy for Th1 autoimmune diseases [18]. Although imiquimod stimulates IFN-γ, IL-12, and this Th1-favouring branch of immune response is used to treat AA, the results are controversial. In a study of D'Ovidio et al., imiquimod did not work for extensive AT and AU cases [19]. On the other hand, imiquimod may be more effective in less extensive and chronic forms of AA with a similar mechanism as other topical immunotherapies, although AA is potentially associated with IFN-γ and IL-1, a Th1 cytokine profile [20].
In conclusion, while AA is a common disease, treatment of its extensive form is difficult and its outcome is not easily predicted. In this study we assume that the elevated serum levels of IFN-γ may reflect the inflammatory symptoms in AA, especially in extensive forms and that control of IFN-γ production may be important to the management of this disease. And also the measurement of serum IFN-γ in patients with AA may be useful in discriminating those likely to progress to AU from the remaining LAA, or as a prognostic indicator. In future studies it could be interesting to evaluate the evolution of IFN-γ levels in a patient with spontaneous regression, or progressive extension such as from LAA to AT or AU. Although we mainly focused on IFN-γ as a crucial cytokine in the pathogenesis of AA, the involvement of IL-1 and TNF-α, which has been suggested by Bodemer et al. [2], should also be investigated in future projects. n

References

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

2. Bodemer C, Peuchmaur M, Fraitaig S, Chatenoud L, Brousse N, de Prost Y. Role of cytotoxic T-cells in chronic alopecia areata. J Invest Dermatol 2000; 114: 112-6.

3. Yoshino T, Asada H, Ando Y, Fujii H, Yamaguchi Y, Yoshikawa K, Itami S. Impaired responses of peripheral blood mononuclear cells to T-cell stimulants in alopecia areata patients with a poor response to topical immunotherapy. Br J Dermatol 2001; 145: 415-21.

4. König A, Happle R, Hoffmann R. IFN-γ-induced HLA-DR but not ICAM-I expression on cultured dermal papillacells is downregulated by TNF-α. Arch Dermatol Res 1997; 289: 466-70.

5. Teraki Y, Imanishi K, Shiohara T. Cytokines in alopecia areata: contrasting cytokine profiles in localized form and extensive form (alopecia universalis). Acta Derm Venereol (Stockh) 1996; 76: 421-3.

6. Olsen E, Hordinsky M, McDonald-Hull S, Price V, Roberts J, Shapiro J, Stenn K. Alopecia areata investigational assessment guidelines. J Am Acad Dermatol 1999; 40: 242-6.

7. McElwee KJ, Hoffmann R. Alopecia areata-animal models. Clin Exp Dermatol 2002; 27: 410-7.

8. Freyschmidth-Paul P, McElwee KJ, Happle R, Kissling S, Wenzel E, Sundberg JP, Zöller M, Hoffmann R. Interleukin-10-deficient mice are less susceptible to the induction of alopecia areata. J Invest Dermatol 2002; 119: 980-2.

9. Hoffmann R, Wenzel E, Huth A, Steen P, Scaufele M, Henninger HP. Cytokine MRNA levels in alopecia areata before and after treatment with the contact allergen diphenylcyclopropenone. J Invest Dermatol 1994; 103: 530-3.

10. 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.

11. Gilhar A, Shalaginov R, Assy B, Serafimovich A, Kalish RS. Alopecia areata is a T-lymphocyte mediated autoimmune disease: Lesional human T-lymphocytes transfer alopecia areata to human skin grafts on SCID mice. J Investig Dermatol Symp Proc 1999; 4: 207-10.

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

13. Gilhar A, Landau M, Assy B, Ullmann Y, Shalaginov R, Serafimovich S, Kalish RS. Transfer of alopecia areata in the human scalp graft/prkdc^scid mouse system is characterized by a Th1 response. Clinical Immunology 2003; 106: 181-7.

14. Sato-Kawamura M, Aiba S, Tagami H. Strong expression of CD40, CD54, and HLA-DR antigen and lack of evidence for direct cellular cytotoxicity are unique immunohistopathological features in alopecia areata. Arch Dermatol Res 2003; 294: 536-43.

15. Griffiths CEM. The immunological basis of psoriasis. JEADV 2003; 17 (Suppl. 2): 1-5.

16. Abdel-Naser MB, Kruger-Krasagakis S, Krasagakis K, Gollnick H, Abdel-Fattah A, Orfanos CE. Further evidence for involment of both cell mediated and humoral immunity in generalized vitiligo. Pigment Cell Res 1994; 7: 1-8.

17. Kalish RS, Gilhar A. Alopecia areata: autoimmunity - the evidence is compelling. JID Symposium Proceedings 2003; 8: 164-7.

18. Skurkovich B, Skurkovich S. Anti-interferon-gamma antibodiesin the treatment of autoimmune diseases. Curr Opin Mol Ther 2003; 5: 52-7 (abstract).

19. D'Ovidio R, Claudatus J, Prima TD. Ineffectiveness of imiquimod therapy for alopecia totalis/universalis. JEADV 2002; 16: 416-17 (Letter).

20. Sommerfeld B, Poppva I. Does imiquimod normalize hair growth in alopecia areata? 10^th Congress of EADV. Munich, Germany, October 10-14, 2001. P32-25 (Abstract).