A major complication of in vivo monoclonal antibody therapy in patients with cancer is the host's immune response to the administered xenogeneic immunoglobulin. We have performed parallel clinical and experimental studies to investigate the possibility that deaggregation of the therapeutic monoclonal antibody might render it non-immunogenic, or even tolerogenic, as has been suggested in several animal studies. Deaggregation of xenogeneic immunoglobulin has been shown by others to induce non-responsiveness in some ('susceptible') but not in other ('resistant') strains of mice. We have used an improved deaggregation method of size exclusion chromatography connected to FPLC and have developed a sensitive ELISA detection system to determine whether highly purified human immunoglobulin G (hIgG) monomers could be tolerogenic even to 'resistant' mice. However, our data show that all preparations of hIgG are immunogenic to 'resistant' mice, and that although deaggregation does significantly reduce the anti-hIgG response to 'susceptible' strains, tolerance is not induced. Concomitant administration of cyclosporin A and deaggregated hIgG had a additive effect in reducing the murine anti-hIgG secondary response. In clinical studies of patients with ovarian cancer who received in vivo immunotherapy with either iodine-131 (not aggregated) or yttrium-90 (aggregated) HMFG1 mouse monoclonal antibody, no significant difference was found between the immune responses to aggregated and non-aggregated murine immunoglobulin G. Our data suggest that deaggregation alone is unlikely to be useful in controlling the human anti-murine immunoglobulin G response in our outbred patient population, although in combination with an immunosuppressant it may be more effective.