Interaction of antibodies with human glomerular epithelial cells.

We tested the hypothesis that the pathogenesis of human idiopathic membranous glomerulonephritis is similar to that of Heymann glomerulonephritis, a model of membranous glomerulonephritis induced in rats by immunization with renal brush border preparations; the characteristic subepithelial deposits result from interaction of antibodies with a brush border antigen (gp330) expressed on the plasma membrane of glomerular visceral epithelial cells (GEC), followed by redistribution and shedding of gp330 immune complexes. The experiments were performed in cultured glomerular visceral epithelial cells, in living monkeys and rats, and in isolated perfused human, monkey, and rat kidneys. Antigens from plasma membranes of human renal brush border vesicles (HBBV) and GEC vesicles (HGECV) and their corresponding polyclonal and monoclonal antibodies reactive with human and monkey GEC were prepared. First, polyclonal antibodies to HGECV bound diffusely to cultured GEC; monoclonal antibody 8G5, recognizing a 60-kDa protein, mainly bound to the coated pits and apical invaginations; both polyclonal HGECV and 8G5 monoclonal antibodies induced antigen redistribution (capping) at 37 degrees C. Second, monkeys were actively or passively immunized, and isolated human and monkey kidneys were perfused with the antibodies. Active immunization with HBBV induced tubular immune deposits, whereas active immunization with HGECV did not provoke renal lesions. After passive immunization HBBV and HGECV antibodies bound diffusely to glomerular cells, and subepithelial deposits were observed during the autologous phase; in contrast, 8G5 induced early (day 3) granular deposits. Third, fine granular deposits developed in glomeruli of human and monkey kidneys perfused for 4 hours at 37 degrees C with 8G5; these deposits were more difficult to detect by electron microscopy than those occurring in kidneys of Lewis rats perfused with sheep antiHBBV. The results show that some antibodies redistribute antigens at the surface of human and monkey GEC in vitro, in vivo, and ex vivo and induce formation of granular deposits in human glomerular capillary walls. Failure to induce more severe lesions in human and monkey kidneys may be ascribed to lack of GEC antigens comparable to rat gp330, insufficient cross linking by monoclonal antibody, lack or insufficient concentration of epitope-specific antibodies, insufficient time of kidney perfusion, or a combination of these factors.