HLA-DP antigens in patients with alopecia areata N. 0dum, N. Morling, J. Georgsen, B. K. Frentz, G. F. Jensen, L. Fugger, A. Svejgaard: HLA-DP antigens in patients with alopecia areata. Tissue Antigens 1990: 35: 114-117.
Abstract: The distribution of HLA-DP antigens were studied in 41 patients with alopecia areata (AA) and 188 ethnically matched controls. An increase of DR4 and possibly DR5 in 24 of these patients has previously been reported. HLA-DP typing for DPwl through w6 and the local specificity, CDP HEI,was performed using the primed lymphocyte typing (PLT) technique. The frequencies of DPw4 were 92.9% in AA patients and 71.3% in controls (relative risk, RR = 5.1, p =0.0019 and p < 0.03 when corrected). DPwl was decreased to 4.9% in patients compared to 14.9% in controls, but the difference was not statistically significant. The frequency of DR4 was 48.8% compared to 31.9% in controls (RR=2.0) whereas the frequency of DR5 did not differ from that in controls. DPw4 was not associated (in linkage disequilibrium)with DR4 or DR5 in patients or controls. Thus, in AA the association with DP and DR are independent of each other. However, individuals with both the DPw4 and DR4 alleles are at increased risk for AA, indicating that synergism between DP and DR gene products may play a role in the genetic susceptibility to AA.
Alopecia areata (AA) is a common hair disease of unknown etiology but of supposed (auto)immune nature. HLA-DR4 and D R 5 have been reported to confer susceptibility to AA (1, 2). DR5 may relate to early onset (2) and DR4 to longstanding, therapy-resistant, and widespread disease (1). Recently, attention has focused on D P antigens as disease markers and a number of associations have been reported, first in juvenile chronic arthritis (3,4) and later in aplastic anemia (9,coeliac disease (6, 7), multiple sclerosis (8), and acute lymphoblastic leukemia (9). Because 1) there is only minimal linkage disequilibrium between D P alleles and those of other loci and because 2) D P antigens functionally seem to resemble D R antigens, studies of D P polymorphism may provide new information on the relation between the HLA system and disease susceptibility, even in diseases with a previously established qssociation with DR. Thus, recent studies (4, 8, 7) have indicated that synergism between different HLA-class I1 (DP and DR/DQ) antigens may play an important role in disease predisposition. 114
N. O d d , N. Marling', J. Georgsen',
B. K. Jakohsea', 6. F r e e , 6. F. Jenton3, L Frggerl and A.Svejgaard' 'lissue Typing Laboratory, and 'Department of Dermatology Rigshospitalet (The National Hospi-
tal), Copenhagen, and 9eparbnent of Internal Medicine, Frederiksberg Hospital, Frederiksberg, Denmark
Received for publication 20 July 1989. revised, accepted 10 January 1990
Material and Methods Patients and controls: The study comprised 41 patients (16 males and 25 females), 24 of whom had been included in a previous study (1). The patients were characterized according to the following criteria: Family history of AA, duration of disease, resistance to therapy, previous atopia or allergic disease, presence of autoantibodies against thyeroglobulin, parietal cells, and intrinsic factors. With one exception, no patient had clinical manifestations of (other) autoimmune diseases. The controls comprised 188 randomly selected healthy Danes. HLA-DP typing:
Typing for HLA-DPwl-w6 and the local specificity, CDP-HEI, with the primed lymphocyte typing (PLT) technique was performed as described previously (10). Statistical methods: The comparisons of HLA antigen frequencies were done by Fisher's exact method and the strength of
HLA-DP antigens in patients with alopecia areata associations was estimated by the relative risk (RR) or odds ratio (OR), using Woolf s method. When no a priori hypothesis had been formulated, the pvalues were corrected by multiplication by (i) the number of DP antigens investigated and (ii) two to obtain a two-sided test (1 1).
Results
The frequency of HLA-DPw4 in AA patients was significantly (uncorrected p =0.002, p-corrected = 0.03) increased to 92.7% as compared to 71.3% in controls (RR=5.1, Table 1). Eight of 17 AA patients, who had not previously been DR typed, were DRCpositive (47.1 YOcompared to 3 1.9% in controls) whereas only 2 were DR5 positive (1 1.8% compared to 10.6% in controls). DR4 was increased to 48.8% (RR=2.03, p=0.032, Table 1) in the combined group of a total of 41 AA patients. In contrast, the frequencies of DR5 in patients and controls did not differ significantly (Table 1). There was no association between DPw4 and DR4 and/or DR5 in patients or controls (Table 2). An increase of DPw4 was seen in both DR4-positive and -negative patients (Table 2). Although there was no association between DPw4 and DR4 in either patient or control populations, the cooccurence of DPw4 and DR4 was significantly more frequent in AA patients (48.9%) compared to controls (23.4%, relative risk = 3.1, p < 0.001).
Table 1. Frequencies of HLA-DP and OR antigens in alopecia areata (AA) and in controls
HLA
DPwl DPw2 OPw3
Oh4 OPw5 OPw6 COP-HE1 OR1 OR2 OR3 OR4 OR5 OR6 OR7
Frequencies Patients Controls (N=41) (N=188) % %
4.9 22.0 19.5 92.7 4.9 4.9 4.9 24.4 19.5 17.1 48.8 22.0 14.6 24.4
14.9 20.2 19.1 71.3 6.4 5.3 5.3 19.6 25.4 28.8 31.9 10.6 16.2 19.6
RR = relative risk 'p = 0.06; i.e. not statistically significant 'p = 0,0019; p < 0.03 atter correction (cf. Methods) +p = 0.032 'p = 0.049; not statistically significant after correction
RR
When comparing the groups of (i) DFW4 and DR4positive, (ii) DPw4-positive DRbnegative, and (iii) DPw4-negative and DRCpositive, patients and controls with the group of DPw4- and DRCnegative patients and controls, significantly higher ORs were obtained for DPw4- and DRllpositive patients (OR=8.6, pKO.05) than for the two other groups (OR=3.8 and 1.2, respectively, Table 2). In addition to the DPw4 association, the only deviation of DP antigen frequencies in A4 was a decreased frequency of DPw 1 which, however, was not statistically significant. The low number of patients investigated did not allow any conclusive analysis of DP antigen frequencies in subgroups of AA. However, the DPw4 association seemed to be most pronounced in patients with long-lasting, therapy-resistant disease and/or patients with autoantibodies, but these differences were not statistically significant. Discussion
The present study suggests that D h 4 is associated with AA, and possibly most pronounced with patients suffering from a severe form. The strength of the association was of the same magnitude as those reported for DR4 and DR5 (1, 2). In addition, we observed an insignificantly decreased frequency of DPw 1, which has also been observed in other diseases of possible autoimmune etiology, e.g. in pauciarticular onset, juvenile chronic arthritis (4) and rheumatoid arthritis (12). It is unlikely that the increased frequency of DPw4 is due to linkage disequilibrium with DR4 because there was no association between these antigens, either in patients or controls. This is in agreement with previous studies of healthy Danes (10,13) and of other northern European populations (14). Moreover, there was no evidence that other, non-HLA related factors, e.g. poor stimu-
O.2gb
5.10'
Table 2 Odds ratios for various phenotypic groups. Number of DR4andlor DPw4
+
-k
-
-k
2.03' 2.36'
+-
~~
patients
controls
OR'
20
44 16
8.6'
1
18 2
90
1.2 3.8
38
~
'Odds ratios (ORs) were calculated using the group of individuals who were negative for both OR4 and DPw4 as reference. 'p
Abstract: The distribution of HLA-DP antigens were studied in 41 patients with alopecia areata (AA) and 188 ethnically matched controls. An increase of DR4 and possibly DR5 in 24 of these patients has previously been reported. HLA-DP typing for DPwl through w6 and the local specificity, CDP HEI,was performed using the primed lymphocyte typing (PLT) technique. The frequencies of DPw4 were 92.9% in AA patients and 71.3% in controls (relative risk, RR = 5.1, p =0.0019 and p < 0.03 when corrected). DPwl was decreased to 4.9% in patients compared to 14.9% in controls, but the difference was not statistically significant. The frequency of DR4 was 48.8% compared to 31.9% in controls (RR=2.0) whereas the frequency of DR5 did not differ from that in controls. DPw4 was not associated (in linkage disequilibrium)with DR4 or DR5 in patients or controls. Thus, in AA the association with DP and DR are independent of each other. However, individuals with both the DPw4 and DR4 alleles are at increased risk for AA, indicating that synergism between DP and DR gene products may play a role in the genetic susceptibility to AA.
Alopecia areata (AA) is a common hair disease of unknown etiology but of supposed (auto)immune nature. HLA-DR4 and D R 5 have been reported to confer susceptibility to AA (1, 2). DR5 may relate to early onset (2) and DR4 to longstanding, therapy-resistant, and widespread disease (1). Recently, attention has focused on D P antigens as disease markers and a number of associations have been reported, first in juvenile chronic arthritis (3,4) and later in aplastic anemia (9,coeliac disease (6, 7), multiple sclerosis (8), and acute lymphoblastic leukemia (9). Because 1) there is only minimal linkage disequilibrium between D P alleles and those of other loci and because 2) D P antigens functionally seem to resemble D R antigens, studies of D P polymorphism may provide new information on the relation between the HLA system and disease susceptibility, even in diseases with a previously established qssociation with DR. Thus, recent studies (4, 8, 7) have indicated that synergism between different HLA-class I1 (DP and DR/DQ) antigens may play an important role in disease predisposition. 114
N. O d d , N. Marling', J. Georgsen',
B. K. Jakohsea', 6. F r e e , 6. F. Jenton3, L Frggerl and A.Svejgaard' 'lissue Typing Laboratory, and 'Department of Dermatology Rigshospitalet (The National Hospi-
tal), Copenhagen, and 9eparbnent of Internal Medicine, Frederiksberg Hospital, Frederiksberg, Denmark
Received for publication 20 July 1989. revised, accepted 10 January 1990
Material and Methods Patients and controls: The study comprised 41 patients (16 males and 25 females), 24 of whom had been included in a previous study (1). The patients were characterized according to the following criteria: Family history of AA, duration of disease, resistance to therapy, previous atopia or allergic disease, presence of autoantibodies against thyeroglobulin, parietal cells, and intrinsic factors. With one exception, no patient had clinical manifestations of (other) autoimmune diseases. The controls comprised 188 randomly selected healthy Danes. HLA-DP typing:
Typing for HLA-DPwl-w6 and the local specificity, CDP-HEI, with the primed lymphocyte typing (PLT) technique was performed as described previously (10). Statistical methods: The comparisons of HLA antigen frequencies were done by Fisher's exact method and the strength of
HLA-DP antigens in patients with alopecia areata associations was estimated by the relative risk (RR) or odds ratio (OR), using Woolf s method. When no a priori hypothesis had been formulated, the pvalues were corrected by multiplication by (i) the number of DP antigens investigated and (ii) two to obtain a two-sided test (1 1).
Results
The frequency of HLA-DPw4 in AA patients was significantly (uncorrected p =0.002, p-corrected = 0.03) increased to 92.7% as compared to 71.3% in controls (RR=5.1, Table 1). Eight of 17 AA patients, who had not previously been DR typed, were DRCpositive (47.1 YOcompared to 3 1.9% in controls) whereas only 2 were DR5 positive (1 1.8% compared to 10.6% in controls). DR4 was increased to 48.8% (RR=2.03, p=0.032, Table 1) in the combined group of a total of 41 AA patients. In contrast, the frequencies of DR5 in patients and controls did not differ significantly (Table 1). There was no association between DPw4 and DR4 and/or DR5 in patients or controls (Table 2). An increase of DPw4 was seen in both DR4-positive and -negative patients (Table 2). Although there was no association between DPw4 and DR4 in either patient or control populations, the cooccurence of DPw4 and DR4 was significantly more frequent in AA patients (48.9%) compared to controls (23.4%, relative risk = 3.1, p < 0.001).
Table 1. Frequencies of HLA-DP and OR antigens in alopecia areata (AA) and in controls
HLA
DPwl DPw2 OPw3
Oh4 OPw5 OPw6 COP-HE1 OR1 OR2 OR3 OR4 OR5 OR6 OR7
Frequencies Patients Controls (N=41) (N=188) % %
4.9 22.0 19.5 92.7 4.9 4.9 4.9 24.4 19.5 17.1 48.8 22.0 14.6 24.4
14.9 20.2 19.1 71.3 6.4 5.3 5.3 19.6 25.4 28.8 31.9 10.6 16.2 19.6
RR = relative risk 'p = 0.06; i.e. not statistically significant 'p = 0,0019; p < 0.03 atter correction (cf. Methods) +p = 0.032 'p = 0.049; not statistically significant after correction
RR
When comparing the groups of (i) DFW4 and DR4positive, (ii) DPw4-positive DRbnegative, and (iii) DPw4-negative and DRCpositive, patients and controls with the group of DPw4- and DRCnegative patients and controls, significantly higher ORs were obtained for DPw4- and DRllpositive patients (OR=8.6, pKO.05) than for the two other groups (OR=3.8 and 1.2, respectively, Table 2). In addition to the DPw4 association, the only deviation of DP antigen frequencies in A4 was a decreased frequency of DPw 1 which, however, was not statistically significant. The low number of patients investigated did not allow any conclusive analysis of DP antigen frequencies in subgroups of AA. However, the DPw4 association seemed to be most pronounced in patients with long-lasting, therapy-resistant disease and/or patients with autoantibodies, but these differences were not statistically significant. Discussion
The present study suggests that D h 4 is associated with AA, and possibly most pronounced with patients suffering from a severe form. The strength of the association was of the same magnitude as those reported for DR4 and DR5 (1, 2). In addition, we observed an insignificantly decreased frequency of DPw 1, which has also been observed in other diseases of possible autoimmune etiology, e.g. in pauciarticular onset, juvenile chronic arthritis (4) and rheumatoid arthritis (12). It is unlikely that the increased frequency of DPw4 is due to linkage disequilibrium with DR4 because there was no association between these antigens, either in patients or controls. This is in agreement with previous studies of healthy Danes (10,13) and of other northern European populations (14). Moreover, there was no evidence that other, non-HLA related factors, e.g. poor stimu-
O.2gb
5.10'
Table 2 Odds ratios for various phenotypic groups. Number of DR4andlor DPw4
+
-k
-
-k
2.03' 2.36'
+-
~~
patients
controls
OR'
20
44 16
8.6'
1
18 2
90
1.2 3.8
38
~
'Odds ratios (ORs) were calculated using the group of individuals who were negative for both OR4 and DPw4 as reference. 'p
