British Journal of Dermatology (1976) 94, 131.
A comparison of histocompatibility antigens in dermatitis herpetiformis and adult coeliac disease P.P.SEAH, LIONEL FRY, J.W.KEARNEY, ELIZABETH CAMPBELL, J.F.MOWBRAY, J.S.STEWART AND A.V.HOEEBRAND Departments of Dermatology and Experimental Pathology, St Mary's Hospital, London, W.2., West Middlesex Hospital, Isleworth, Middlesex, and Department of Haematology, Royal Free Hospital, London, N.W.3 Accepted for publication 29 May 1975
SUMMARY
The incidence of histocompatibility antigens HL-A, 4a and 4b was studied in thirty-eight patients with dermatitis herpetiformis (DH) and thirty-six patients with adult coeliac disease (ACD). 3]he 4b antigen was found in all the DH and ACD patients. HL-A 8 was found in 89% ofpatients with ACD ^^^milar to the mcidence reported in previous studies—and in 79% of patients with DH, a higher incidence than in previous studies which may be due to stricter criteria being used here to diagnose DH. There was no significant difference in the incidence of HL-A 8 between those patients with DH whose small intestinal biopsies appeared macroscopically abnormal and those with a normal macroscopic appearance. These findings suggestthat patients with DH form a single disease group and do not support the concept previously postulated that there are two groups of patients with DH, one w i t h ^ increased incidence of HL-A 8 antigen similar to that in ACD who have a gluten sensitive enteropathy (GSE), and anothgr with a normal incidence of HL-A 8 antigen and without enteropathy.
The association between dermatitis herpetiformis (DH) and an enteropathy was first clearly demonstrated by Marks, Shuster & Watson (1966) and the enteropathy was soon shown to be gluten sensitive (Fry et al, 1967, 1968). However, the incidence and significance of gluten sensitive enteropathy (GSE) in DH is still disputed. Gebhard et al (1973) adhere to the earlier view that two-thirds of patients with DH have GSE and suggest that only in these patients is the incidence of HL-A 8 raised, as it is in adult coeliac disease (ACD) (Falchuk, Rogentine & Strober, 1972; Stokes et al, 1972). However, F£y & Seah-(i974) now believe that all patients with DH have GSE, and that earher reports of only two-thirds of patients having GSE were based on inadequate criteria for diagnosing DH and GSE. There are also two different views held on the relationship between the skin and intestine in DH. First, there are those who believe that the relationship is direct and that both the skin and gut lesions are gluten dependent (Fry et al, 1968, 1969, 1973; Barnetson, Heading & White, 1973). I jSecondly, there is the view that the relationship is indirect and the skin lesions are not gluten depeniment (Shuster, Watson & Marks, 1968; Weinstein et al, 1971; Gebhard et al, 1973). Reprint requests to: Dr L. Fry, Department of Dermatology, St Mary's Hospital, London W2 iNY. 131
132
P.P.Seah et al.
There is now increasing evidence that immunological processes may be involved in the pathogenesis of both DH and ACD, and that the^irSiunological disorder may be reflected"by a EglTtrequency of certain Kistocompatibility antigens. Using more precise criteria for the diagnosis of DH and GSE, we have studied the incidence of histocompatibility antigens, including for the first time, the complex of antigens 4a and 4b of van Rood in patients with DH and ACD, with particular reference to the presence and severity of the intestinal lesion in DH. The results of this study clearly demonstrate that DH patients fall into a single group and cannot be divided according to the presence of a macroscopic gut lesion and increased incidence of HL-A 8 on the one hand, and absence of both these factors on the other. PATIENTS AND METHODS Patients and controls
Thirty-eight patients with DH and thirty-six patients with ACD were investigated. The diagnosis of DH was made on clinical and histological features, the response of the skin lesions to dapsone and reappearance of the rash on withdrawal of the drug, and the presence of IgA in the uninvoived skin (Fry & Seah, 1974). The diagnosis of ACD was made in patients with a flat, or flat-with-mosaic proximal small bowel mucosal appearance (Holmes, Hourihane & Booth, 1961), and the characteristic histological features, particularly decreased cell height and increased lymphocytic infiltration of the small intestinal epithelium. Controls were provided from the Tissue Typing Laboratory at St Mary's Hospital. Data from 180 individuals, mainly pregnant women, blood donors and laboratory staflF were used, representing a cross-section of a London population, from which the patients were derived. For the 4a and 4b antigen, data from 923 similar controls were used. These controls were typed with the same anti 4a and anti 4b sera as the disease population. Histocompatibility
typing
Lymphocytes from defibrinated blood were used in a modified two-stage Kissmeyer-Neilsen lymphocyte micro-toxicity test. Antisera from multiple pregnancy and sensitized kidney transplant patients were used: a total of 171 antisera were employed, and the range of antisera to each of the histocompatibihty antigens varied from 1:17 different sera. The specificities tested in the first and second segregant series were HL-A i, 2, 3, 5,7, 8,9,10,11,12,13,14,17,27,28 and W5, Wio, W15, W21, W22, W29, W32, Da25. The sera used for typing were a mixture of locally obtained sera and those obtained from a number of typing centres. The specificities have been validated by parallel typing with the National Tissue Typing Reference Laboratory and London Hospital plates. Patients were also tested for the histocompatibility antigens 4a and 4b of van Rood whose relationship to the HL-A arjigprif; remains, at the present time, uncertain, it was decided to test for these antigens in view of the~ known frequent association of 4b with the HL-A i, 8 chromosome. The 4a and 4b antisera were in part a gift from Professor J.J.van Rood and some local sera which consistently gave similar reactions to these. Small intestinal biopsies
Small intestinal biopsies were taken before treatment with a gluten-free diet. The biopsies were taken from the duodeno-jejunal fiexure under radiographic control, using a Crosby capsule and polythene guide cuff as described by Evans et al (1970). Specimens were examined under a dissecting microscope for their macroscopic appearances and then processed routinely for histology. Quantification of intra-epithelial lymphocytic infiltration was also performed in patients with DH according to the method of Fry et al. (1972).
Histocompatibility antigens in dermatitis herpetiformis in
p
o •a
a
O
M
a
I 3
CO
•g & m
•a
•s s 8 .S
-S
o\
Ov 00
6
o\ 00 00
a
00
l^ M r»
M M M
ov t^ M
•* ov
m c»
p w^
M N
tvb
p m
N t^ •* t-
Ov 60 m Ov
p
o 00
00
^1^ '5 "^ II
133
134
P.P.Seah et al. RESULTS
Histocompatibility antigens The frequency of detection of the HL-A antigens in control subjects and those with ACD and DH are shown in Table i. In the controls HL-A i was present in 32%, HL-A 8 in 26% and HL-A i and 8 existing together in 19%. In the patients, HL-A i was detected in twenty-three (61%) of the thirty-eight DH patients and in twenty-seven (75%) of the thirty-six ACD patients. HL-A 8 was detected in thirty (79%) of the DH patients, and in thirty-two (89%) of the ACD patients. HL-A i and 8 were found together in TABLE 2. Tissue typing results in thirty-eight patients with dermatitis herpetiformis HL-A I 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38
+ — + + + + + + ++ + + + + + + + + + + + + + + — +
HL-A 8
Others
+ 3,12 + 2, 12 + 2 , 12, W28 + W18 2, 5, 12, W29 + 12, W29 3, 5,W5 + 2, 3, 7 + 1 2 , W29 + 11, W22 2, 12, W29, W18 + 2, W5 + + + + + + + + + + + + + + + + + + + +
2 , 2 ,
3 , 3 , 2 ,
2 ,
2, 7 ,
WIO
W27 2, W18, W27 3, W I O 7 3, 12, W22, W29 II, W5 12, W29 7 3, 7, 9, W18 7 2, 9, 12 9, W21 12, W29 9 7 9, W32, W18 I I , W18 9,W5 9, 10, 12, W18 12, W29 W15 12 9 W18
4a
4b
+ + + + + + + — + + — + + + + + + + — + + + + -
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
Histocompatibility antigens in dermatitis herpetiformis
135
twenty-two (58%) of the thirty-eight DH patients and in twenty-seven (75%) ofthe thirty-six ACD patients. The incidence of HL-A i was significantly greater than in the controls in both the DH (P
A comparison of histocompatibility antigens in dermatitis herpetiformis and adult coeliac disease P.P.SEAH, LIONEL FRY, J.W.KEARNEY, ELIZABETH CAMPBELL, J.F.MOWBRAY, J.S.STEWART AND A.V.HOEEBRAND Departments of Dermatology and Experimental Pathology, St Mary's Hospital, London, W.2., West Middlesex Hospital, Isleworth, Middlesex, and Department of Haematology, Royal Free Hospital, London, N.W.3 Accepted for publication 29 May 1975
SUMMARY
The incidence of histocompatibility antigens HL-A, 4a and 4b was studied in thirty-eight patients with dermatitis herpetiformis (DH) and thirty-six patients with adult coeliac disease (ACD). 3]he 4b antigen was found in all the DH and ACD patients. HL-A 8 was found in 89% ofpatients with ACD ^^^milar to the mcidence reported in previous studies—and in 79% of patients with DH, a higher incidence than in previous studies which may be due to stricter criteria being used here to diagnose DH. There was no significant difference in the incidence of HL-A 8 between those patients with DH whose small intestinal biopsies appeared macroscopically abnormal and those with a normal macroscopic appearance. These findings suggestthat patients with DH form a single disease group and do not support the concept previously postulated that there are two groups of patients with DH, one w i t h ^ increased incidence of HL-A 8 antigen similar to that in ACD who have a gluten sensitive enteropathy (GSE), and anothgr with a normal incidence of HL-A 8 antigen and without enteropathy.
The association between dermatitis herpetiformis (DH) and an enteropathy was first clearly demonstrated by Marks, Shuster & Watson (1966) and the enteropathy was soon shown to be gluten sensitive (Fry et al, 1967, 1968). However, the incidence and significance of gluten sensitive enteropathy (GSE) in DH is still disputed. Gebhard et al (1973) adhere to the earlier view that two-thirds of patients with DH have GSE and suggest that only in these patients is the incidence of HL-A 8 raised, as it is in adult coeliac disease (ACD) (Falchuk, Rogentine & Strober, 1972; Stokes et al, 1972). However, F£y & Seah-(i974) now believe that all patients with DH have GSE, and that earher reports of only two-thirds of patients having GSE were based on inadequate criteria for diagnosing DH and GSE. There are also two different views held on the relationship between the skin and intestine in DH. First, there are those who believe that the relationship is direct and that both the skin and gut lesions are gluten dependent (Fry et al, 1968, 1969, 1973; Barnetson, Heading & White, 1973). I jSecondly, there is the view that the relationship is indirect and the skin lesions are not gluten depeniment (Shuster, Watson & Marks, 1968; Weinstein et al, 1971; Gebhard et al, 1973). Reprint requests to: Dr L. Fry, Department of Dermatology, St Mary's Hospital, London W2 iNY. 131
132
P.P.Seah et al.
There is now increasing evidence that immunological processes may be involved in the pathogenesis of both DH and ACD, and that the^irSiunological disorder may be reflected"by a EglTtrequency of certain Kistocompatibility antigens. Using more precise criteria for the diagnosis of DH and GSE, we have studied the incidence of histocompatibility antigens, including for the first time, the complex of antigens 4a and 4b of van Rood in patients with DH and ACD, with particular reference to the presence and severity of the intestinal lesion in DH. The results of this study clearly demonstrate that DH patients fall into a single group and cannot be divided according to the presence of a macroscopic gut lesion and increased incidence of HL-A 8 on the one hand, and absence of both these factors on the other. PATIENTS AND METHODS Patients and controls
Thirty-eight patients with DH and thirty-six patients with ACD were investigated. The diagnosis of DH was made on clinical and histological features, the response of the skin lesions to dapsone and reappearance of the rash on withdrawal of the drug, and the presence of IgA in the uninvoived skin (Fry & Seah, 1974). The diagnosis of ACD was made in patients with a flat, or flat-with-mosaic proximal small bowel mucosal appearance (Holmes, Hourihane & Booth, 1961), and the characteristic histological features, particularly decreased cell height and increased lymphocytic infiltration of the small intestinal epithelium. Controls were provided from the Tissue Typing Laboratory at St Mary's Hospital. Data from 180 individuals, mainly pregnant women, blood donors and laboratory staflF were used, representing a cross-section of a London population, from which the patients were derived. For the 4a and 4b antigen, data from 923 similar controls were used. These controls were typed with the same anti 4a and anti 4b sera as the disease population. Histocompatibility
typing
Lymphocytes from defibrinated blood were used in a modified two-stage Kissmeyer-Neilsen lymphocyte micro-toxicity test. Antisera from multiple pregnancy and sensitized kidney transplant patients were used: a total of 171 antisera were employed, and the range of antisera to each of the histocompatibihty antigens varied from 1:17 different sera. The specificities tested in the first and second segregant series were HL-A i, 2, 3, 5,7, 8,9,10,11,12,13,14,17,27,28 and W5, Wio, W15, W21, W22, W29, W32, Da25. The sera used for typing were a mixture of locally obtained sera and those obtained from a number of typing centres. The specificities have been validated by parallel typing with the National Tissue Typing Reference Laboratory and London Hospital plates. Patients were also tested for the histocompatibility antigens 4a and 4b of van Rood whose relationship to the HL-A arjigprif; remains, at the present time, uncertain, it was decided to test for these antigens in view of the~ known frequent association of 4b with the HL-A i, 8 chromosome. The 4a and 4b antisera were in part a gift from Professor J.J.van Rood and some local sera which consistently gave similar reactions to these. Small intestinal biopsies
Small intestinal biopsies were taken before treatment with a gluten-free diet. The biopsies were taken from the duodeno-jejunal fiexure under radiographic control, using a Crosby capsule and polythene guide cuff as described by Evans et al (1970). Specimens were examined under a dissecting microscope for their macroscopic appearances and then processed routinely for histology. Quantification of intra-epithelial lymphocytic infiltration was also performed in patients with DH according to the method of Fry et al. (1972).
Histocompatibility antigens in dermatitis herpetiformis in
p
o •a
a
O
M
a
I 3
CO
•g & m
•a
•s s 8 .S
-S
o\
Ov 00
6
o\ 00 00
a
00
l^ M r»
M M M
ov t^ M
•* ov
m c»
p w^
M N
tvb
p m
N t^ •* t-
Ov 60 m Ov
p
o 00
00
^1^ '5 "^ II
133
134
P.P.Seah et al. RESULTS
Histocompatibility antigens The frequency of detection of the HL-A antigens in control subjects and those with ACD and DH are shown in Table i. In the controls HL-A i was present in 32%, HL-A 8 in 26% and HL-A i and 8 existing together in 19%. In the patients, HL-A i was detected in twenty-three (61%) of the thirty-eight DH patients and in twenty-seven (75%) of the thirty-six ACD patients. HL-A 8 was detected in thirty (79%) of the DH patients, and in thirty-two (89%) of the ACD patients. HL-A i and 8 were found together in TABLE 2. Tissue typing results in thirty-eight patients with dermatitis herpetiformis HL-A I 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38
+ — + + + + + + ++ + + + + + + + + + + + + + + — +
HL-A 8
Others
+ 3,12 + 2, 12 + 2 , 12, W28 + W18 2, 5, 12, W29 + 12, W29 3, 5,W5 + 2, 3, 7 + 1 2 , W29 + 11, W22 2, 12, W29, W18 + 2, W5 + + + + + + + + + + + + + + + + + + + +
2 , 2 ,
3 , 3 , 2 ,
2 ,
2, 7 ,
WIO
W27 2, W18, W27 3, W I O 7 3, 12, W22, W29 II, W5 12, W29 7 3, 7, 9, W18 7 2, 9, 12 9, W21 12, W29 9 7 9, W32, W18 I I , W18 9,W5 9, 10, 12, W18 12, W29 W15 12 9 W18
4a
4b
+ + + + + + + — + + — + + + + + + + — + + + + -
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
Histocompatibility antigens in dermatitis herpetiformis
135
twenty-two (58%) of the thirty-eight DH patients and in twenty-seven (75%) ofthe thirty-six ACD patients. The incidence of HL-A i was significantly greater than in the controls in both the DH (P
