Mucous membrane pemphigoid (MMP) is an autoimmune mucosal scarring disease having severe ocular morbidity [1]. Disease susceptibility is associated with increased frequencies of human leukocyte antigen (HLA)-DQB1*03:01, HLA-DRB1*11, and HLA-DRB1*04 and decreased frequencies of DQB1*02 [2]. To explore correlations between clinical involvement and HLA-class-II alleles, we prospectively phenotyped a cohort of 55 British MMP patients, and 41 age/sex-matched controls (ethics approval reference 09/H0721/54).

Inclusion criteria have been described [3] and were pre-agreed clinical criteria with or without a positive direct immunofluorescence (DIF) study. HLA-typing used allele-specific sequencing protocols (Protrans S3 HLA-DRB1* and HLA-DQB1* Cyclerstrips, Protrans, Hockenheim, Germany). Statistical analysis used Fisher’s exact test (p), with Benjamini–Hochberg correction (pc) defined as significant when pc < 0.05.

The study dataset, with clinical and laboratory results, is provided in Table and Legend S1. Fifteen patients, 14 of them with ocular involvement, were DIF-negative of whom seven had antibodies to basement membrane zone epitopes. MMP-affected sites varied: eight were ocular only, ten oral only, 15 oral and ocular only and 22 multisite. Exons 2 and 3 in HLA-DQB1 were fully analysed in 54/55 patients and 39/41 controls (Table 1a). The frequency of HLA-DQB1*03:01 was increased (pc = < 0.01) in 36/54 (67%) of MMP patients (13 homozygous and 23 heterozygous) compared to 13/39 (33%) controls (two homozygous and 11 heterozygous, Table 1a). Exon 2 of HLA-DRB1 was also fully analysed in 54/55 patients and 40/41 controls; HLA-DRB1*03:01 was decreased (p = < 0.01, pc  =  0.045, Table 1b). Additionally, we compared the frequency of HLA-DQB1*0301 with controls for different sites of involvement showing that HLA-DQB1*0301 was increased in all subgroups except for ocular only MMP. Compared to controls, DIF-positive MMP had significantly increased HLA-DQB1*0301 (p = 0.00009) but this difference was not significant for either DIF-negative MMP (p = 0.113) or for DIF-positive ocular only MMP.

Table 1a Distribution of HLA-DQB1 among patients and controls
Full size table
Table 1b Distribution of HLA-DRB1 among patients and controls
Full size table

Some studies have described a correlation of HLA-DQB1*03:01 with ocular and oral MMP or in ocular only MMP, whereas others have shown HLA-DQB1*03:01 to be associated with multisite MMP [2]. In this study, the association of HLA-DQB1*0301 with MMP was lower than in the largest reported study [2] (pc < 0.00000028), possibly due to our inclusion of eight patients without detectable tissue-bound or serum antibodies, who were excluded from the latter study [2]. In ocular only MMP 50% are DIF-negative but have a phenotype that both progresses and responds to therapy in the same way as DIF-positive cases [3, 4]. DIF-negative ocular MMP may result from inadequate test sensitivity, because of the small volumes of tissue involved, to dominance of an autoreactive T-cell mediated over a autoantibody-mediated disease [4] or because this is a different disease subset. We included our DIF-negative cases because to leave these out of this analysis disregards a group of cases which do not fit criteria for any other disease.

In our prospectively characterized cohort the association with HLA-DQB1*0301, HLA-DQB1*02, and HLA-DRB1*11 was corroborated, whereas HLA-DRB1*0301 was identified as a potentially protective allele, which requires confirmation in a larger study.