Autoimmunity mediated by cross-reactive antibodies, elicited by HIV-1 envelope glycoproteins gp120/gp160, has been postulated to contribute to the pathogenesis of AIDS. Partial amino acid sequence homology between gp120/gp160 and several human host proteins, including MHC antigens and immunoglobulins, has been perceived as the basis for immunological cross-reactivity. Binding of antibodies from sera of HIV-1-infected individuals to selected host proteins and/or to synthetic peptides derived from them and the inhibitory activity of such sera in assays measuring the functional activity of T cells provided apparent support for the autoimmunity hypothesis, which is also relevant to the issue of safety of anti-HIV-1 vaccines. Considering the possibility that the detected autoantibodies may arise for reasons other than antibody responses to gp120/gp160, the immunological cross-reactivity between gp120/gp160 and the relevant host proteins was investigated using hyperimmune rabbit anti-gp120/gp160 and monoclonal antibodies. As determined from dilution end-point comparisons for polyclonal anti-gp120, the cross-reactivity of anti-gp120 with CD4 was undetectable (< 10(-5)%). The cross-reactivity of anti-gp120/gp160 with HLA-I and HLA-II antigens was also undetectable (< 4 x 10(-4)%) and that with other human proteins reported to have partial sequence homology with gp120/gp41 was < or = 0.013%. Anti-gp120/gp160 did not have detectable inhibitory effects in functional assays measuring proliferative T cell responses. Therefore, immunization with gp120/gp160 is unlikely to elicit harmful autoimmune responses.