Myositis has been associated with HLA-B8 and DR3, especially in white patients with polymyositis and serum anti-Jo-1 antibodies. Twenty-eight patients with myositis and serum translation-related autoantibodies anti-Jo-1, anti-PL-7, anti-PL-12, anti-KJ, and anti-SRP were studied for HLA class I1 specificities by Southern blotting with HLA-DRP, DQP, and D Q a probes. The association of HLA-DR3 (DRwl7) with anti-Jo-1 antibodies in white myositis patients was confirmed (P = 0.003, relative risk 8.9). However, HLA-DRw52 haplotypes, regardless of subtype, were present in all of the white and black patients with serum anti-Jo-1 and other translation-related autoantibodies. Moreover, one anti-Jo-1 positive patient had HLADRw8, an HLA-DRw52 haplotype on which the DRp3 gene has been partially deleted. No HLA-DQ specificity or allele was common to all patients. The HLA-DR3, DRS, DRw6, and DRw8 haplotypes, which bear the HLA-DRw52 specificity, share the most homology in the DRPl first hypervariable region at amino acid positions 9-13. Thus, this DRPl region appears to be the most likely candidate "epitope" for translation-related autoimmune responses in inflammatory myositis.Inflammatory myositis, particularly polymyositis in adults and dermatomyositis in children, has been associated with HLA-B8 and DR3 in white individuals (1,2). In an earlier study using serologic HLA typing (3), the specific subset of myositis patients with serum anti-histidyl-transfer RNA (tRNA) synthetase (antiJo-1) autoantibodies (4) had an even higher frequency of HLA-DR3 than myositis patients overall; all 11 patients with serum anti-Jo-1 antibodies had either HLA-DR3 and/or HLA-DRw6. It was hypothesized that this autoimmune response in myositis may be