Unique characteristics in Japanese dermatitis herpetiformis

DEAR EDITOR, Dermatitis herpetiformis (DH) is an intensely pruritic papulovesicular skin disease, and is considered to be a cutaneous manifestation of gluten sensitivity, that is, an extraintestinal presentation of coeliac disease (CD). The processing of the dietary gluten antigen (gliadin) requires human leucocyte antigen (HLA)-DQ2 or HLA-DQ8, and approximately 85% of white patients with DH carry HLA-DQ2; the majority of the remaining 15% carry HLA-DQ8. Tissue transglutaminase (tTG) (transglutaminase 2) is the major autoantigen in CD, whereas epidermal transglutaminase (eTG) (transglutaminase 3) is the major autoantigen in DH. We previously reported that the unique features of Japanese DH were (i) a high frequency of fibrillar deposition of IgA in the papillary dermis; (ii) the rare occurrence of gluten sensitive enteropathy (GSE); (iii) the absence of HLA-DQ2 and HLA-DQ8; and (iv) a rare association with autoimmune diseases or lymphomas. In this study, between April 2012 and December 2013, we sent questionnaires about DH to the dermatology departments of all Japanese universities and major hospitals. Finally, we obtained clinical data, skin biopsy specimens, frozen skin sections, sera and DNA samples from 21 patients newly diagnosed with DH and who were not included in our previous study. We performed a HLA study, serological tests and immunofluorescent (IF) studies for TG (detailed methodology can be found in Data S1; see Supporting Information). Approval for this study was obtained from the ethics committee of Kurume University. Because no clinical data were available for one patient, the clinical data were summarized for the remaining 20 Japanese patients with DH (Table 1). The results of the DNA-based HLA study of 16 patients and a Japanese control are shown in Tables S1 and S2 (see Supporting Information). Compared with a Japanese control population, no statistical difference was observed in any allele. DQB1*02:01 (encoding the b-chain of HLA-DQ2) was not found in any patient, whereas DQB1*03:02 (encoding the b-chain of DQ8) was found in six (37%) patients, who also carried DR4 (DRB1*04:03, DRB1*04:05 or DRB1*04:06) and DQA1*03. Four of six patients with DQB1*03:02 were found by direct IF to have granular IgA deposition; two of these patients also had anti-eTG antibodies. DRB1*09:01 was found in eight (50%) patients, who also bore DQB1*03:03 (encoding the b-chain of DQ9). The results of the serological tests for all 21 patients are shown in Table 2 and Table S3 (see Supporting Information). IgA anti-tTG and anti-eTG antibodies were positive in eight (38%) and nine (43%) patients, respectively. Indirect IF of normal human skin and 1 mol L 1 NaCl split skin was negative in all patients tested (n = 14). Skin samples from 14 patients were available for direct IF for antibodies to tTG and eTG. IgA was co-localized with eTG in the papillary dermis in nine (64%) patients with eTG deposition (Fig. S1; see Supporting Information). When the results of serological tests and direct IF were combined, 14 (67%) of 21 patients were positive for eTG reactivity (Table S4; see Supporting Information). In this study, no patient had symptoms of GSE. As no small bowel biopsy was performed, which is a big limitation of this study, some patients may have histopathological findings suggestive of CD, particularly patients with positive IgA anti-tTG or antiendomysial antibodies. However, in duodenal biopsies, no Japanese patients with inflammatory bowel disease who had elevated IgA anti-tTG antibodies showed an endoscopic appearance or histopathological changes suggestive of CD. In contrast to the invariable occurrence of GSE in white patients with DH, GSE seems unlikely to develop in Japanese patients with DH. In this study, DQB1*02:01, which is most frequently detected in white people with DH, was not found in any Japanese patients with DH. In our previous study, none of the HLA-DQ loci in 10 patients who were examined showed HLA-DQ2 (DQB1*02:01) or HLA-DQ8 (DQB1*03:02). DQB1*02:01 was not found in the Japanese control in this study. Therefore, it may be reasonable to consider that no Japanese patients with DH have DQB1*02:01, even in further study. In contrast, six (38%) patients had DQB1*03:02. However, the significance of this association is unclear as the frequency of DQB1*03:02 in Japanese patients with DH was not significantly higher than that in the Japanese control population (18%). In addition, it was unclear whether the relatively high frequency of DQA1*03 contributed to the increased frequency of DQB1*03:02 (Table S5; see Supporting Information). In our previous study, nine (64%) of 14 patients with a granular pattern detected by direct IF had HLA-DR9, and the frequency was significantly higher than that of a Japanese control population. In this study, DRB1*09:01 was found in four (44%) patients with a granular pattern, and this frequency was not significantly different from that in the Japanese