Effects of ionic strength and sugars on the aggregation propensity of monoclonal antibodies: influence of colloidal and conformational stabilities.

PURPOSE: To develop a general strategy for optimizing monoclonal antibody (MAb) formulations. METHODS: Colloidal stabilities of four representative MAbs solutions were assessed based on the second virial coefficient (B 2) at 20°C and 40°C, and net charges at different NaCl concentrations, and/or in the presence of sugars. Conformational stabilities were evaluated from the unfolding temperatures. The aggregation propensities were determined at 40°C and after freeze-thawing. The electrostatic potential of antibody surfaces was simulated for the development of rational formulations. RESULTS: Similar B 2 values were obtained at 20°C and 40°C, implying little dependence on temperature. B 2 correlated quantitatively with aggregation propensities at 40°C. The net charge partly correlated with colloidal stability. Salts stabilized or destabilized MAbs, depending on repulsive or attractive interactions. Sugars improved the aggregation propensity under freeze-thaw stress through improved conformational stability. Uneven and even distributions of potential surfaces were attributed to attractive and strong repulsive electrostatic interactions. CONCLUSIONS: Assessment of colloidal stability at the lowest ionic strength is particularly effective for the development of formulations. If necessary, salts are added to enhance the colloidal stability. Sugars further improved aggregation propensities by enhancing conformational stability. These behaviors are rationally predictable according to the surface potentials of MAbs.

Ternary isocratic mobile phase optimization utilizing resolution Design Space based on retention time and peak width modeling.

An optimization procedure of ternary isocratic mobile phase composition in the HPLC method using a statistical prediction model and visualization technique is described. In this report, two prediction models were first evaluated to obtain reliable prediction results. The retention time prediction model was constructed by modification from past respectable knowledge of retention modeling against ternary solvent strength changes. An excellent correlation between observed and predicted retention time was given in various kinds of pharmaceutical compounds by the multiple regression modeling of solvent strength parameters. The peak width of half height prediction model employed polynomial fitting of the retention time, because a linear relationship between the peak width of half height and the retention time was not obtained even after taking into account the contribution of the extra-column effect based on a moment method. Accurate prediction results were able to be obtained by such model, showing mostly over 0.99 value of correlation coefficient between observed and predicted peak width of half height. Then, a procedure to visualize a resolution Design Space was tried as the secondary challenge. An artificial neural network method was performed to link directly between ternary solvent strength parameters and predicted resolution, which were determined by accurate prediction results of retention time and a peak width of half height, and to visualize appropriate ternary mobile phase compositions as a range of resolution over 1.5 on the contour profile. By using mixtures of similar pharmaceutical compounds in case studies, we verified a possibility of prediction to find the optimal range of condition. Observed chromatographic results on the optimal condition mostly matched with the prediction and the average of difference between observed and predicted resolution were approximately 0.3. This means that enough accuracy for prediction could be achieved by the proposed procedure. Consequently, the procedure to search the optimal range of ternary solvent strength achieving an appropriate separation is provided by using the resolution Design Space based on accurate prediction.