Clogging model analysis for aggregate-spiked plasma IgG
Evaluation of the filtration behavior for the reference, control, modified CEX1 output and normal AEX output for aggregate-spiked plasma IgG solutions in Figure 3 with each of the four clogging models is shown in Figure 6. The reference solution (no aggregate spike) and modified CEX output (complete removal of larger aggregate species) showed good agreement between experimental and calculated results for all four clogging models, with all four calculated plot lines being almost identical and overlapping with the experimental results (Figure 6a and 6c). In contrast, control (spiked with aggregate) and normal AEX output (no aggregate removal) had different plots for each clogging model, and in both of these cases, the complete blocking model showed the smallest differences between experimental and calculated values (Figure 6b and 6d).
Figure 7 shows the plots of k and mean differences between experimental and calculated values for filtration behavior shown in Figure 3 for plasma IgG. From Figure 7a, the reference (no aggregate spike) and output from modified CEX1 and modified CEX2 (complete removal of larger aggregate species) was 0.01 or smaller for all of the clogging models, and modified CEX1 had the smallest k for all models. For output from modified CEX2 and reference, k was smallest for the complete blocking model, but along with output from modified CEX1, these three solutions had almost the same k irrespective of clogging model. On the other hand, the control and output from normal AEX (both with relatively high proportions of larger aggregate content) had markedly higher k than column outputs with relatively low proportion of aggregate.
The differences between experimental and calculated values for each solution and all four clogging models in Figure 7b show that for solutions with relatively high proportions of larger aggregate content (output from normal AEX) and no aggregate content (output from mixed-mode AEX1), both show the best fit with the complete blocking model. The differences between experimental and calculated values were very small for the three solutions with small k, namely the reference, output from modified CEX1 and output from modified CEX2, and the cake filtration model shows a slightly smaller difference between experimental values and calculated values than the other clogging models for reference and output from modified CEX2. A deeply interesting point is that the output for mixed-mode AEX1, which removed the larger aggregates through column chromatography and had lower dimer content than the outputs for modified CEX1 and modified CEX2 (5.8% vs. 7.3% and 7.7%, respectively), had a markedly higher k than the output for both modified CEX resins and the reference, which had 0.3% larger aggregate content. These results along with the SEC data for mAb solutions in Figure 2 suggest that there are components besides aggregates detected by SEC that impede the filterability of plasma IgG on a virus filter, and these components formed in the process of producing aggregates are reduced by processing by modified CEX but not by modified AEX1.