The binding of aprindine to serum proteins with statistical considerations concerning the analysis of binding data

The binding of the potent, basic antiarrhythmic agent aprindine to serum proteins was studied in solutions of human serum albumin (HSA, lyophilized, 98% pure, KABI) and in human sera. The percentual binding in serum (90.9-96.7%) was considerably higher than in HSA solutions. The binding in serum decreased from 95-97% to 91-93% as the aprindine concentration was raised from 4 to 15 micrograms/ml. The serum binding differed significantly in sera from three normal individuals. By plotting (Formula: see text), as a function of bound drug and performing statistical analysis of the curves a striking difference between the number of primary binding sites in serum and in HSA solution was found (N1 for the HSA solution was 0.0016 and for one serum it was 0.020 "per albumin molecule".) The association constants for the primary binding sites (K1) were 4.3 X 10(6)M-1 for serum and 2.1 X 10(6)M-1 for the HSA solution. Furthermore, the analysis indicated that considerably less than one secondary binding site was present for each albumin molecule in the HSA solution. Therefore, no binding of aprindine to albumin molecules has been demonstrated and it is concluded that the entire binding of aprindine in serum as well as in 98% HSA solutions may be due to the presence of acid protein molecules. In particular, low percentual binding in HSA solutions in spite of the high association constant can be explained by a very low concentration of the binding proteins. The addition of propranolol or phenprocoumon did not cause any displacement of aprindine. In the appendix statistical problems in the analysis of the binding curves are elucidated. The experimental variance and the statistical acceptability of the binding model are discussed.