HLA ASSOCIATIONS IN VITILIGO PATIENTS IN THE DUTCH POPULATION

GERARD T. VENNEKER* , LEO P. DE WAALt, WIETE WESTERHOF* , JOE D' AMARO§, GEZINA M.TH. SCHREUDER§, SYED S. ASGHAR* *Department of Dermatology, University of Amsterdam, Academisch Medisch Centrum, Amsterdam 'Department of Transplantation Immunology, Central Laboratory of the Red Cross Blood Transfusion Services, Amsterdam 'Department of Immunohaematology and Blood Bank, Leiden University Hospital, Leiden, The Netherlands


INTRODUCTION
Vitiligo is a patchy disease with depigmentation of the skin which has a variable age of onset, equal sex distribution and is occasionally familial with an autosomal dominant mode of inheritance with variable penetrance and expression (Finco et ai, 1991).This disease entity is commonly associated with several autoimmune diseases (McGregor et  a!., 1972; Cunliffe et a!., 1969; Bor et al., 1969; Lerner, 1959).The incidence of vitiligo in patients with autoimmune disease is 8% to 15 % compared to 1 % in the general population (Koransky, 1980).There are autoantibodies directed against melanocytes (Bystryn and Pfeffer, 1988; Norris et ai, 1988).Recent immunophenotyping studies suggest that cellular immunity against melanocytes may also be involved in vitiligo (Abdel-Nasser et a!., 1991).These observations suggest that there may be an autoimmune component in its etiology and pathogenesis.There is no consensus in previous reports of HLA-A and HLA-B associations with vitiligo in different caucasoid populations (Metzkeret al., 1980;Retornaz eta!., 1976;Gunther and Richter, 1976).For that reason, we decided to carry out a search for HLA associations in patients with vitiligo in the Dutch caucasoid population.We attempted to confirm the earlier reports of a significant positive association with HLA-DR4 (Foley et aI., 1983) and of a sig-nificant negative association with DR3 (Finco et ai, 1991), and to determine if there were any additional HLA-DR and DQ associations.

PATIENTS AND METHODS
Forty-eight unrelated Dutch caucasoid patients with generalized vitiligo without any associated disease were identified by the Department of Dermatology of the Academic Medical Centre of the University of Amsterdam.The control cases consisted of 703 healthy unrelated Dutch caucasoid blood donors provided by the Central Laboratory of the Blood Transfusion Service in Amsterdam.
HLA-A, B, C typings were performed with the standard NIH Iympho-cytotoxity method (van Rood, 1979) and the HLA-DR, DQ typings with the two-colour fluorescence test (van Rood et aI., 1976).All typings were performed in the HLA typing laboratory of the Central Laboratory of the Blood Transfusion Service, Amsterdam, The Netherlands.
Haldane's modification of Woolf's method was used to calculate the Relative Risk (RR) and its significance in this study (Woolf, 1955;Haldane, 1955).When indicated, P values were corrected for multiple comparisons using the formula suggested by J.
Edwards: Pc= I -(I -pu)n, where Pu is the uncorrected and Pc the corrected p value and n is the number of comparisons (Edwards, 1974).Antigenic splits were not included in the analyses because the validity of their estimated frequencies would be poor in the modest number of patients studied.

RESULTS
An overview of the borderline of clearly significantly different frequencies of the HLA-A, B, C, DR and DQ broad antigens in vitiligo patients and controls are set out in Table I.HLA-DR3 and DR4 are included in the table because other inve~gators have reported significant associations for those antigens (Finco et aI., 1991;Foley et al. , 1983).In the set of 48 patients HLA-A2, B 14, DR4 and DR6 were increased and HLA-B7, Cw7 and DR3 were decreased.Among these 7 antigens, only Cw7 and DR6 were still significant after correction for the 45 broad HLA antigens (8+ 16+ 7 + I 0+4) which were examined (p values of 0.04 and 0.02 respectively).No additional significant differences were found in the remaining 38 (45-7) broad HLA antigens.

DISCUSSION
Foley et al. reported a significant positive association between HLA-DR4 and vitiligo (RR=2.367,Puncorr=0.0049)and Finco et at., reported a negati ve association between HLA-DR3 and vitiligo (RR=0.4216).However, the p values in both studies were not corrected for the number of antigens examined.
In our results, HLA-Cw7 and DR6 are the only antigens which retain their significance after correction for the 45 antigens which were compared.That fact is interesting since vitiligo is commonly associated with several autoimmune diseases (McGregor et al., 1972;Cunliffe et ai., 1969;Bor et aI., 1969;Lerner, 1959), where its incidence in patients with autoimmune disease is 8% to 15% compared to only I % in the general population (Koransky, 1980), and there is evidence that DR6 may be a marker for high im'mune responsiveness (Hendriks et al., 1983;Hendriks et ai., 1986;Hendriks et ai., 1983).et al.. 1991;Foley et aI. , 1983).

Table 1 .
HLA antigen frequencies in vitiligo patients and their controls in the Dutch caucasoid population.HLA-DR3 and DR4 are included in TableIbecause other investigators have reported significant associations for those antigens (Finco