Table 5. IgM benchmark purification process summary
|
Table 5 summarizes process metrics. The impact of monoliths on throughput is striking throughout the process but is especially
so at the capture step. A 1,250 mL sample was loaded on an 8 mL monolith in 62.5 min. The same load would have required 375
min on a 10 mL conventional cation exchange column (11.3 x 100 mm, 200 cm/hr). This fact highlights but under-expresses the
ability of monoliths to relieve the much-discussed bottleneck in downstream processing. The ÄKTA used in these experiments
was configured for a maximum flow rate of 20 mL/min, limiting the 8 mL monoliths to a maximum flow rate of 2.5 Cv/min. Separation
conditions and capacity determinations during early development were conducted at 12 Cv/min. Had the ÄKTA been configured
to accommodate its maximum flow rate, the cation exchange sample loading time would have been 13 min.
After fundamental deficiencies in the basic process have been addressed, individual variables can be optimized. For all methods,
inline dilution factors are key determinants of overall productivity. Higher dilution factors support higher binding capacities,
which support the use of smaller columns and elute more concentrated product in a smaller volume. Lower dilution factors consume
less buffer and less process time.
With hydroxyapatite, one should consider evaluating binding and elution pH values as low as 6.5 and as high as 8.0. As in
the previous discussion concerning ion exchangers, it may be useful to bind at one pH to maximize capacity, and to elute at
another to improve contaminant removal. CHT Type II, 40 μm, is typically the best option for process development and manufacturing.
On most industrial-scale columns, 20 μm is too small for the frits, and 80 μm provides less backpressure but has lower capacity.
CHT Type I (for all particle diameters) has a substantially lower range of pore diameters and will likely support lower binding
capacity for IgMs.45 Selectivity is slightly different, however, and could be useful in a given situation. Ceramic fluorapatite (CFT) has yet
to be characterized for IgM purification.
For cation exchange, one should consider evaluating more common manufacturing buffers such as citrate and phosphate. Using
citrate or phosphate as eluting ions in place of NaCl may also improve pH control during the separation.46 Citrate and phosphate are generally considered inappropriate for anion exchangers, but they may still be effective. Many
IgMs tolerate pH values above 8.0 or below 6.0, but these should be evaluated with care. Weak anion exchange monoliths (DEAE;
EDA, or ethylenediamine) or weak cation exchange monoliths may be evaluated. Conventional ion exchangers may also be considered.
For HIC on weakly hydrophobic media, different binding salts can give remarkably different selectivities.47 A good starting point for such media is 1.2–1.5 M ammonium sulfate. Potassium phosphate supports similar average binding
strength at the same concentrations. Also of possible interest is 1.0 M sodium sulfate or sodium citrate. Sometimes 4.0 M
sodium chloride is used for IgG on phenyl columns, but it generally does not support adequate IgM binding on moderately hydrophobic
columns. Different pH values occasionally produce worthwhile selectivities.
Scale-up Issues
The low diffusion constant of IgM imposes slow flow rates on porous particle media such as hydroxyapatite, HIC, and conventional
ion exchangers. A practical maximum is 200 cm/hr; 100 cm/hr provides better performance. Slow flow rates favor scale-up on
columns with shallow bed heights, ideally no greater than 20 cm. Consistently good quality packing at industrial scale can
be obtained with 15 cm, or even 10 cm bed height, in columns that permit packing by dynamic axial compression. Such columns
are especially effective for hydroxyapatite because they can accommodate its high density and rapid settling rate. These columns
also overcome the primary cause of performance loss by hydroxyapatite: particle damage coincident with repacking. When repacking
is required, these columns resuspend the hydroxyapatite by upward flow, thereby avoiding the use of tools that might damage
the particles.
Experimental results from cycling studies indicate that hydroxyapatite can be used for at least 50 cycles without detectable
change in performance; however, these studies have mostly been performed using purified IgG as a model.48 As noted above, CCS contains components that can interact directly with hydroxyapatite and may reduce its effective lifetime.
Hydroxyapatite has been shown to withstand more than 15,000 hours of exposure to 1.0 M NaOH.45
Scale-up with monoliths is simpler because column packing is eliminated. If air is introduced into a monolith, it can be displaced
quickly and efficiently, and without loss of column performance, by restoration of buffer flow. Industrial monoliths are available
at 8, 80, 800, and 8,000 mL volumes. For the antibodies in this study, an 8 L monolith represents an IgM capacity of approximately
250–300 g per cycle. Larger scale requirements can be accommodated by plumbing multiple units in parallel, or by configuring
multiple units to create a simulated moving bed. The current generation of preparative monoliths is synthesized from polymethacrylate,
which is the same polymer found in many porous-particle ion exchangers and HIC media used for industrial purification of antibodies
and other injectable products. Likewise, monoliths withstand repeated exposure to sodium hydroxide, and they can be expected
to support reusability similar to other polymethacrylate media.
|
