Gene(s) in the MHC of the NZW strain (H-2z) up-regulate(s) systemic lupus erythematosus (SLE) in (NZB x NZW) F1 mice. So far, two plausible mechanisms have been implicated: (i) unique mixed haplotype class II molecules formed in the F1 mice act as a restriction element for self-reactive T cells and (ii) a unique polymorphism in the H-2-linked NZW tumor necrosis factor (TNF)-alpha allele which down-regulates TNF-alpha is contributory. Because of the difficulty in dissecting these alleles within the H-2 complex, it has not been determined which is indeed the case. We addressed this issue by establishing three different H-2-congenic (NZB x NZW) F1 mice bearing distinct haplotypes at class II and TNF-alpha regions, i.e. (NZB x NZW) F1 (H-2d/z:A(d/u)E(d/u)TNF(d/z)), (NZB x NZW.PL) F1 (H-2(d/u):A(d/u)E(d/u)TNF(d/d)) and (NZB x NZW.H-2d) F1 (H-2(d/d):A(d/d)E(d/d)TNF(d/d)). Among these, only (NZB x NZW) F1 produced a markedly lower level of TNF-alpha, due to the unique NZW TNF-alpha allele (TNF(z)). Studies of anti-DNA antibodies and lupus nephritis revealed that, compared to (NZB x NZW) F1, the disease of (NZB x NZW.H-2d) F1 was markedly reduced. In (NZB x NZW.PL) F1, the onset of renal disease was significantly delayed, while the extent of proteinuria and renal histopathology in individuals that had developed the disease was comparable to that seen in (NZB x NZW) F1. It seems likely that both class II and TNF-alpha gene polymorphisms are functioning as H-2-linked predisposing genetic elements, and that the TNF-alpha polymorphism acts to modulate an initial process of the renal disease.