Application Notes

6 Table 1. T m values from the five replicate iFormulate™ formulations Replicate T m T m Δ 1 76.3 76.3 0.0 2 70.9 71.0 0.1 3 70.5 70.5 0.0 4 74.6 75.0 0.4 5 78.1 77.8 0.3 DOE analysis of iFormulate™ data. By designing a response-surface quadratic design, the linear, interaction, and quadratic terms were evaluated in a multivariate fashion (i.e. all variables simultaneously). Hence, all three interactions between the four formulation variables (pH, trehalose concentration, salt concentration, and buffer concentration) were evaluated to optimize lysozyme's T m . The output of the DOE analysis is shown in Figure 5. In the Pareto analysis, the terms are sorted in decreasing absolute effect order, so that the effect of greatest magnitude appears at the top of the graph (Figure 5A). This analysis clearly shows that pH is the most important variable in the lysozyme formulation, followed by trehalose and NaCl. The negative sign for pH indicates that pH is negatively correlated with T m (i.e. low pH conditions results in high T m values). In contrast, trehalose concentration has a positive correlation, which suggests that increasing trehalose concentration will increase the T m . This effect is illustrated by the 3-D response surface plot for pH and trehalose concentration effects on T m (Figure 5C). Note that the slope for the pH dependency is much steeper than for the trehalose dependency, validating that pH affects the T m significantly more than trehalose concentration. The effect of NaCl concentration and pH (Figure 5B) suggests that the ionic strength (concentration of NaCl) only has a minor effect on T m and is mostly independent of pH values. This dataset demonstrates that the DOE approach allows for rapidly determining optimal formulation conditions with a minimal number of experiments (Figure 5D) The analysis defines an optimal design space for lysozyme formulations for high T m 's of > 76 °C at low pH values around pH 5 and trehalose concentrations between 8-10 wt %, with 50 mM NaCl and 30 mM buffer. Figure 5D shows the results of this formulation and validates the predicted T m with the actual data. Conclusion In conclusion, using iFormulate™ and the Prometheus NT.48 nanoDSF technology, it is possible to formulate proteins rapidly within < 1 hour, with minimal material requirements (<< 10 mg), and with exquisitely high precision and reproducibility of experimental data. The four critical formulation variables of pH, ionic strength, stabilizer concentration, and buffer concentration can be evaluated by a multivariate approach with high statistical power. This approach can result in identification of stable protein formulations in rapid and precise fashion. We have extended such an approach to therapeutic proteins such as antibodies and other biopharmaceuticals that will be available in other application notes or publications. Figure 5. (next page) Panels showing various DOE analysis. (A) Pareto analysis showing the ranking of the variables. (B) 3-D response surface curve of pH and NaCl. (C) 3-D response surface curve of pH and trehalose. (D) is a 2-D contour surface curve of pH and trehalose showing the formulation design space that is optimized for the highest T m (shaded area). The red cross indicates subsequently tested parameters to confirm the DEO-output. (E) Triplicate thermal unfolding curves of lysozyme in 20 mM acetate pH 5.0, 50 mM NaCl, and 8 % trehalose. The obtained T m value (76.7 °C +/- 0.1 °C) matches the prediction shown in 5D.

• Cover