There are two types, A and B, of Epstein-Barr virus (EBV) and B95-8 represents the common type A laboratory strain. Herein, we show in a family study that paternal EBV-specific cytotoxic T lymphocytes (CTL) generated in short-term cultures following stimulation with the autologous B95-8-transformed lymphoblastoid cell line (LCL) or B cells freshly infected with the B95-8 isolate did not lyse haploidentical B95-8 LCL expressing the HLA-A1, -B8, -DR3 paternal haplotype. In contrast, the haploidentical B95-8 LCL expressing the HLA-A11, -B51, -DR7 paternal haplotype was strongly lysed. Moreover, paternal CTL generated in response to stimulation with the B95-8 LCL expressing the haploidentical HLA-A1, -B8, -DR3 paternal haplotype included an allogeneic response against the maternal haplotype but no EBV-specific response as shown by the poor lysis of the autologous LCL target cells. However, stimulation with the haploidentical HLA-A11, -B51, -DR7 paternal haplotype resulted in the generation of both an allogeneic and an EBV-specific response. CTL clones were generated from two HLA-B8+ donors in response to stimulation with the autologous type A LCL transformed with wildtype EBV. The clones were cross-reactive for an immunodominant B95-8-associated peptide epitope that interacted with the HLA-B8 allele but failed to lyse B95-8-transformed LCL targets unless the targets were pre-coated with the exogenous peptide. A CTL clone that was initially stimulated with the autologous BL74 LCL lysed the spontaneous autologous LCL and spontaneous LCL from an HLA-B8+ donor, but failed to lyse the B95-8 LCL from that donor. The observed haplotype preference can be explained in terms of sequence variation between the B95-8 and the corresponding wildtype epitope. Our findings may help to clarify the role of EBV in the pathogenesis of primary Sjögren's syndrome which is closely associated with HLA-B8.