BackgroundThat Epstein–Barr virus (EBV) infection is associated with systemic lupus erythematosus (SLE) is established. The challenge is to explain mechanistic roles EBV has in SLE pathogenesis. Previous studies identify four examples of autoantibody cross-reactions between SLE autoantigens and Epstein–Barr nuclear antigen 1 (EBNA1). For two of these examples, the earliest detected autoantibody specifically cross-reacts with EBNA1; thereby, defined EBNA1 epitopes induce a robust autoantibody response in animals. These results suggest that the autoantibodies initiating the process leading to SLE may emerge from the anti-EBNA1 heteroimmune response. If this hypothesis is true, then anti-EBNA1 responses would be more frequent in EBV-infected SLE patients than in EBV-infected controls. We tested this prediction.MethodsWe evaluated published East Asian data by selecting those with a positive anti-viral capsid antigen (VCA) antibody immunoglobulin G (IgG) test and determining whether anti-EBNA1 was more common among the EBV-infected SLE cases than among matched EBV-infected controls with conditional logistic regression analysis.ResultsAll the qualifying SLE patients (100%) in this dataset were EBV-infected compared to age- and sex-matched controls (92.2%) [odds ratio (OR) = 28.6, 95% CI 6.4–∞, p = 8.83 × 10-8], confirming the known close association of EBV infection with SLE. Furthermore, virtually all the SLE cases have both anti-VCA IgG and anti-EBNA1 IgG antibodies [124 of 125 (99.2%)], which are more frequently present than in age- and sex-matched EBV-infected controls [232 of 250 (93.2%)] (OR = 9.7, 95% CI 1.5–414, p = 0.0078) for an 89.7% SLE attributable risk from anti-EBNA1, which is in addition to the 100% SLE risk attributable to EBV infection in these data.ConclusionsThe association of EBV infection with SLE is reconfirmed. The prediction that anti-EBNA1 is more frequent in these SLE cases than in EBV-infected controls is true, consistent with the hypothesis that anti-EBNA1 contributes to SLE. This second EBV-dependent risk factor is consistent with a molecular mimicry model for the generation of SLE, starting with EBV infection, progressing to anti-EBNA1 response; then molecular mimicry leads to anti-EBNA1 antibodies cross-reacting with an SLE autoantigen, causing autoantibody epitope spreading, and culminating in clinical SLE. These results support the anti-EBNA1 heteroimmune response being a foundation from which pathogenic SLE autoimmunity emerges.