Altered antigenicity of human monoclonal antibodies derived from human-mouse heterohybridomas.

Abstract

We have generated milligram quantities of human monoclonal antibodies (Hu-MAbs) in the ascites of pristane-primed nude mice injected with human-mouse heterohybridomas. After contaminating mouse immunoglobulins were removed by affinity chromatography, an enzyme immunosorbent assay (EIA) was used to measure the concentrations of human immunoglobulins. Ten different partially purified preparations were tested. The titration curves with all 5 IgG Hu-MAbs were unusual, reaching a plateau at a very low apparent maximum concentration of antibody. In contrast, the EIA yielded more usual titration curves and thus apparently more reliable estimates of the concentrations of 4 IgM and 1 IgA monoclonal antibodies. An analogous EIA for the quantitation of mouse IgG monoclonal antibodies also gave accurate estimates. To understand the nature of the discrepancy with human IgG, 5 Hu-MAbs of the 3 classes (2 IgG, 2 pentameric IgM and 1 IgA) were purified to homogeneity for a more detailed analysis. The inability to quantitate the human IgG monoclonal antibodies by EIA was not due to defective molecules, as shown by SDS polyacrylamide gel electrophoresis. The human IgG monoclonal antibodies were found to consist of intact heavy and light chains, as were the IgM and IgA antibodies. The possibility that the human IgG monoclonal antibodies differed antigenically from polyclonal IgG was explored by comparing the concentrations by EIA with the protein concentrations determined by absorbance at 280 nm. This analysis permitted a comparison of the detectability of antigenic determinants on Hu-MAbs with those on polyclonal Ig with goat antibodies to Ig or Ig subclass. The IgG monoclonal antibodies differed from polyclonal IgG in both their heavy and light chains. Goat antiserum monospecific for the gamma chain in fact underestimated the concentration by as much as one hundred-fold. IgM and IgA monoclonal antibodies were less antigenically distinct from their polyclonal counterparts even though their light chains were also underestimated, because goat monospecific antibodies were more efficient at recognizing their heavy chains. The molecular basis for the observed difference in antigenicity is not yet known. These findings have important implications for the analysis of the binding of IgG Hu-MAbs. A direct binding assay with the label directly conjugated to the Hu-MAb should be used in preference to an indirect assay with a labeled detecting antibody to maximize the sensitivity of the assay. The altered antigenicity of IgG Hu-MAbs may also imply decreased immunogenicity when they are given in vivo as carriers for radionuclides or cytotoxic antitumor materials.