A panel of anti-gp120 human monoclonal antibodies (HuMAbs), CD4-IgG, and sera from people infected with human immunodeficiency virus type 1 (HIV-1) was tested for neutralization of nine primary HIV-1 isolates, one molecularly cloned primary strain (JR-CSF), and two strains (IIIB and MN) adapted for growth in transformed T-cell lines. All the viruses were grown in mitogen-stimulated peripheral blood mononuclear cells and were tested for their ability to infect these cells in the presence and absence of the reagents mentioned above. In general, the primary isolates were relatively resistant to neutralization by the MAbs tested, compared with the T-cell line-adapted strains. However, one HuMAb, IgG1b12, was able to neutralize most of the primary isolates at concentrations of < or = 1 microgram/ml. Usually, the inability of a HuMAb to neutralize a primary isolate was not due merely to the absence of the antibody epitope from the virus; the majority of the HuMAbs bound with high affinity to monomeric gp120 molecules derived from various strains but neutralized the viruses inefficiently. We infer therefore that the mechanism of resistance of primary isolates to most neutralizing antibodies is complex, and we suggest that it involves an inaccessibility of antibody binding sites in the context of the native glycoprotein complex on the virion. Such a mechanism would parallel that which was previously postulated for soluble CD4 resistance. We conclude that studies of HIV-1 neutralization that rely on strains adapted to growth in transformed T-cell lines yield the misleading impression that HIV-1 is readily neutralized. The more relevant primary HIV-1 isolates are relatively resistant to neutralization, although these isolates can be potently neutralized by a subset of human polyclonal or monoclonal antibodies.