• primary sclerosing cholangitis
• HLA-DRB1*13
• tumour necrosis factor α
• genetic polymorphism

Susceptibility to primary sclerosing cholangitis (PSC) is linked to HLA-A1-B8-DRB1*0301-DQB1*0201 and HLA-DRB1*1301-DQB1* 0603 haplotypes in different populations of Northern European origin and also to HLA-DRB1*1501-DQB1*0602 in the UK.^1–,⁴

Mitchell et al have reported an association between tumour necrosis factor alpha promoter gene (TNFA) polymorphism at position −308 and PSC (Gut 2001;49:288–94). In this respect, increased distribution of the TNF*2 allele, in strong linkage disequilibrium with the HLA-A1/B8/DRB1*0301 haplotype, was observed in PSC patients from Norway but not from the UK. However, analysis of the combined data confirmed a significant association of TNFA*2 with PSC. This overrepresentation of TNFA*2 was seen only in subjects with HLA-A1-B8-DRB1*0301, indicating that the observed association of PSC with TNFA*2 might in fact be secondary to linkage disequilibrium within this haplotype.

Bernal and colleagues⁵ have previously reported an increased frequency of TNFA*2 in another cohort of British patients with PSC. This association was dependent on the presence of HLA-B8 and DRB3*0101 but not of HLA-DRB1*0301. Based on these results, the authors proposed that the associations with TNFA*2 and HLA-B8 were stronger than those observed with HLA-DRB1 and DRB3.

We have investigated the frequencies of HLA-B, DRB1, DQB1, and TNFA alleles in 63 Brazilian patients with PSC and 83 healthy controls from the metropolitan area of São Paulo, Brazil, using polymerase chain reaction based techniques, as previously described.^6–,⁸ This population is of highly admixed origin with different percentages of Caucasoid, African, and Amerindian ancestries. The diagnosis of PSC was based on the findings of typical clinical, laboratory, cholangiographic, and histological features.⁹ None of the patients had evidence of concurrent hepatitis B or C or hepatic schistosomiasis. Twenty seven patients (18 males; mean age 15 (±7) years) were less than 16 years at disease onset and were considered children, and 36 subjects were adults (23 males, mean age 34 (±11) years). Forty one patients had inflammatory bowel disease (IBD). None of the subjects, including all children, had any evidence of laboratory or histological features of overlapping syndromes of PSC and autoimmune hepatitis (AIH).

No increase in the frequency of HLA-B, DRB3, DRB4, or DRB5 alleles was observed in PSC patients compared with healthy controls. Likewise, the distribution of TNFA alleles was similar in patients and controls. The frequency of HLA-DRB1*1301 (52% v 20% of controls; p=0.00009, RR=4.3) and HLA-DQB1*06 (59% v 41% of controls; p=0.04, RR=2.1) was significantly increased in PSC patients (table 1⇓). However, one third of HLA-DRB1*13 positive patients carried other HLA-DQB1 alleles (data not shown). This overrepresentation of HLA-DRB1*13 was seen in both paediatric (44% v 20% of controls; p=0.02, RR=3.1) and adult patients (58% v 20% of controls; p=0.00009, RR=5.4). However, this association was seen only in patients with IBD (61% of patients with IBD v 20% of controls (p=0.00001, RR=6.1) and 36% of patients without IBD v 20% of controls (NS)).

In summary, our data indicate that predisposition to PSC in Brazil is primarily linked to HLA-DRB1*13 and suggest that the association with TNFA*2 previously observed in Norwegian and British patients with PSC could be due to linkage with HLA-DRB1*0301. The association of HLA-DRB1*13 with PSC was observed in both children and adults with the disease but was restricted to patients with concurrent IBD, as previously described by Donaldson and colleagues.¹⁰

Interestingly, AIH type 1 was also associated with HLA-DRB1*13 but not with the TNFA*2 allele in Brazil.^11,12 Of note, shared HLA antigens have also been associated with AIH type 1 and PSC in other populations.¹³ These findings suggest that the same HLA-DRB1 alleles confer susceptibility to distinct autoimmune diseases of the liver such as AIH type 1 and PSC and point to the presence of similar immune mechanisms leading to different clinical outcomes.