Problems and pitfalls with measurement of antibody affinity using solid phase binding in the ELISA.

Current methods of estimation of antibody affinity constants using ELISA assume homogeneous binding of antibody to the solid phase, despite many reports in the literature that this is not true. I have derived theoretical antibody binding curves for solid phase antigen assuming homogeneous antibody binding. I have compared these curves with a set of experimental binding curves of monoclonal antibodies to the serum protein fibronectin. The results conclusively show that while some monoclonal antibodies behave as predicted by theory, others show departures from homogeneous binding which can be explained by various surface effects. I have discussed how these surface effects can cause errors in estimates of either liquid phase or solid phase affinities using the ELISA, and have demonstrated the limitations of methods of affinity ranking.