INTRODUCTION

Column chromatography is a well-established method for characterization, purification and manufacture of a wide range of products, from foodstuffs to life-saving medications. In biopharmaceutical manufacture, it is critical that the purification process is robust and reproducible lot-to-lot. As such, it is vital that chromatography columns are efficiently packed, and able to be qualified within a reasonable time frame. 

A well-packed column means that a homogeneous packed bed is produced each time you perform a separation. This results in maximum efficiency, yield, and purity. Whilst process-scale columns are robust, with commercially available resins usually supplied with well-designed packing methods, at the lab-scale process development stage, variances in packing approach, and even the design of your chosen chromatography column, can significantly impact the results of screening evaluations. Resin performance cannot be truly evaluated if the column is not packed efficiently.

As scaling-up from lab to process-scale is a critical manufacturing activity for biopharmaceutical manufacturers, achieving column-packing consistency at the process development stage is a key consideration. 

To determine the effect of column design upon resin performance, an analysis was conducted between two different agarose-based chromatography resins, packed (as per manufacturer’s instructions) into two commonly-used columns at smaller scales – a TricornTM column from GE Healthcare, and Omnifit® column manufactured by Kinesis.


EXPERIMENT

MabSelect SuRe™ LX was packed into a Tricorn column, as per manufacturer’s instructions. Praesto® AP was packed into a Tricorn column using the method described in Appendix A.

MabSelect SuRe™ LX and Praesto® AP were also packed into Omnifit columns using the method described in Appendix B.

Dynamic binding capacities on the laboratory scale columns were performed using polyclonal immunoglobin G and the following protocol (Table 1).



Table 1: Dynamic binding capacity determination protocol for Protein A resins

RESULTS - MABSELECT™ SURE LX

The 10% breakthrough capacity using Tricorn columns packed using the method described in Appendix A, showed poor reproducibility and a high level of variability between different packings and operators.

The 10% breakthrough capacity using Omnifit columns packed using the method described in Appendix B, showed excellent reproducibility and low levels of variability between different packings and operators.

Each result presented in Figure 1 is an average of three experiments per analyst.



Figure 1. Dynamic binding capacities at 6 mins residence time -                                                    Table 2. Dynamic binding capacities - 10% breakthrough
10% breakthrough hlgG for MabSelect SuRe™ LX packed in                                                           (mg/ml) hIgG for MabSelect SuRe™ LX at 6 mins
Tricorn and Omnifit columns                                                                                                                residence time.

RESULTS - PRAESTO® AP

The use of Omnifit columns, in conjunction with the packing method outlined in Appendix A, provides reproducible capacity results that closely mirror data generated for the in-house reference standard, a Repligen Minichrom 8 x 20 column.

The reference column is machine packed, producing identical packing performance on each column of a batch. For this reason, it is used as a quality check.

The results in Figure 2 and Table 3 show that up to a 25% drop in capacity can be seen by a poorly packed Tricorn column. This could have a significant effect on resin screening evaluations, whereas, the design of the Omnifit column along with an efficient packing method from the resin vendors allow the user to gain maximum capacity at laboratory scale.



Figure 2. Dynamic binding capacities - 10% breakthrough hlgG for                                                   Table 3. Dynamic binding capacities - 10% breakthrough
Praesto
® AP packed in Tricorn and Omnifit columns                                                                            hlgG for Praesto® AP packed in Tricorn and Omnifit columns        

CONCLUSIONS

Tricorn and Omnifit laboratory-scale columns are both versatile small scale columns that give the user great flexibility when packing multiple different resin types.

However, the adaptor design of the Tricorn column (Appendix C) does not distribute flow evenly over the circumference of the top of the column bed. As a result, the column will appear efficiently packed or even over-packed when it is in fact underpacked. This under-packing explains the loss in overall capacity.

Kinesis Omnifit columns are similar in design apart from the adaptor and plunger mechanism. The flow is evenly distributed over the circumference of the adaptor before the frit (Appendix C), translating to an evenly distributed flow during packing and experimental runs. The same would apply for XK / HiScaleTM columns from GE Healthcare and YMC columns, which each have designs allowing for a more even flow distribution from the adaptor.

Breaking new ground for mAb purification...

.