Protein Peptide Purification using the Multicolumn Countercurrent Solvent Gradient Purification (MCSGP) Process

January 1, 2009

A new method for MAb purification.

ABSTRACT

Increasing upstream yields have led to the need for downstream purification capacities. In this article, a new chromatographic process for the purification of biomolecules is presented. Its working principle is explained and possible areas of application in monoclonal antibody and polypeptide purification are discussed. For an industrial polypeptide purification problem, experimental work showed that a multicolumn countercurrent solvent gradient purification (MCSGP) process could raise the productivity 25-fold, compared to the current batch chromatographic step in production.

The majority of the therapeutic proteins undergoing late-stage development or entering the market are monoclonal antibodies (MAbs). Because high doses of these products are needed per patient, relatively large amounts of protein must be produced and purified. The past decade saw tremendous advances in cell culture technology producing higher yields. As a result, purification has now become a bottleneck, or at least a major cost driver with a high potential for cost cutting. Consequently, MAb purification has received a lot of attention in academia and industry and a variety of solutions have been proposed, including but not limited to crystallization, (i.e., ion-exchange), Protein-A mimetic ligands, liquid–liquid extraction, and non-affinity chromatography in batch or continuous operation. The non-affinity chromatography used in a continuous process such as multicolumn countercurrent solvent gradient purification (MCSGP) shows high potential to replace current standard MAb purification schemes, as seen in a recent presentation by Merck-Serono.1–8

The benefits of continuous chromatographic processes such as MCSGP have spurred interest in using their technology for other purification challenges, such as the purification of polypeptides.9 In the past years, a revival of polypeptides for therapeutic purposes has been seen, particularly for treating diabetes. Polypeptides are produced by either synthetic routes or fermentation. Because polypeptides have a higher production cost than MAbs but are generally produced in lower quantities, the major concern in the purification of these molecules is yield.

Conventionally, the purification of polypeptides is performed using reversed-phase chromatography in a batch operation. Often, two or more chromatographic steps with different mobile phases are used to achieve the specified purity. Continuous processes are ideally suited for such difficult purification steps because they can significantly improve the separation efficiency compared to conventional batch chromatography. The best known continuous chromatographic process is the simulated-moving-bed (SMB) process.10 In the pharmaceutical industry, that method has been applied successfully at commercial scale for the purification of small, chiral molecules. The challenge in such purifications is to separate two enantiomers from a racemic mixture. But the SMB technology is generally of no use in the purification of therapeutic polypeptides or proteins because it can perform only binary separations where gradient chromatography cannot be properly implemented. For very specific purification problems, however, academia has investigated the use of SMB and SMB-derivatives for the purification of biomolecules.11–15 A detailed comparison of the existing continuous chromatographic process has been published elsewhere.6

It is also worth noting that the SMB technology can only perform binary separations i.e., producing two streams, such as an early stream eluting waste and a later stream/eluent product or two product streams (product A and product B). The purification of therapeutic polypeptides and proteins in non-affinity chromatography, however, requires ternary separations, i.e., three streams containing the early eluting waste, the product, and the late eluting waste. Unlike the SMB process, the MCSGP process is suited for ternary separations using gradient chromatography and can therefore be used for the continuous purification of therapeutic polypeptides and proteins.

Recently, the MCSGP process was tested for the purification of the polypeptide Calcitonin (a growth hormone) in a joint project with Novartis Pharma AG (Basel, Switzerland), ChromaCon AG and ETH Zurich (both in Zurich, Switzerland).

In this article, the purification challenges for the reversed-phase chromatography of Calcitonin are described and the experimental results of the purification of this polypeptide with the MCSGP process are shown. A comparison is given in terms of purity, yield, and productivity between the MCSGP process and the conventional batch process for the purification problem presented.

THE MCSGP PROCESS

The working principle of the MCSGP process can be explained using a conventional batch elution chromatogram, as shown in Figure 1. The x-axis in the figure indicates operating time. The following steps are taken sequentially: 1) load the column with feed, 2) wash the column to elute nonadsorbing impurities, 3) gradient elution, 4) clean-in-place (CIP) or wash with a strong eluent, and 5) wash with a weak eluent for regeneration. The next batch elution can then be started by loading the column again.

Figure 1

The molecules eluted during the gradient operation can generally be classified into the following three categories: early eluting impurities, product, and late eluting impurities. These categories will be abbreviated as follows: W for early eluting or weakly adsorbing, P for product, and S for late eluting or strongly adsorbing. These three categories are indicated schematically in Figure 1 by three triangles: the first triangle represents W and ends at time tB, P elutes between tA and tD, and S starts eluting at tC. In general, the elution time windows of W–P and P–S overlap in the time intervals tA to tB and tC to tD, respectively, as indicated in Figure 1. In conventional batch chromatography, the P eluted in the overlapping regions would have to be discarded (or could be recycled in peptide purification) and only the P eluted between tB and tC is the collected product meeting specifications.

In the MCSGP process, the same elution scheme as shown in Figure 1 is maintained, but the process is operated with three small columns instead of the one large column used in batch chromatography.

The three columns used in the MCSGP process are operated either in the connected mode (upper flow path in Figure 2) or in batch mode (lower flow path in Figure 2). Each flow path is operated for a fixed time, which can be the same for the batch and the connected mode, but does not need to be. The arrows in Figure 2 show how a single column is switched through the different positions.

Figure 2

Tracking a single column, the following tasks are carried out: in the first position in the lower right corner of Figure 2, the column is loaded with the feed so that a first portion of the W is eluted from the column, but no P is eluted. This task corresponds in batch chromatography to the elution of all molecules from t = 0 to t = tA. Before any product is eluted, the column is switched to the middle position in the connected mode.

At this position, the overlapping fraction W–P (i.e., tA to tB in Figure 1) is eluted. It should be noted that this fraction that is being eluted from the middle column in the connected mode is not leaving the process, but gets mixed with the stream coming from a gradient pump before it enters the column in the upper right position, where it adsorbs.

The column is then switched to the lower middle position, where the pure P can be eluted, equivalent to the time interval from tB to tC in the batch chromatogram in Figure 1.

Before any S is eluted into the pure product outlet stream, the column is switched again to the upper left position where the overlapping region P–S is eluted, but because of the connected mode, it stays in the process. This corresponds to the elution time interval tC to tD in Figure 1. Now the remaining molecules that are still in the column all belong to the category S.

Therefore, the column is switched to the lower left position, where it is treated with a strong eluent to wash the column. If necessary, an additional, more rigorous treatment with another eluent can be considered.16

It is not in the scope of this paper to provide a detailed description of the principles or design of the MCSGP process. A typical application for MCSGP will be shown in the following section. More details about the design of the MCSGP can be found elsewhere.1–3

THE PURIFICATION OF A THERAPEUTIC POLYPEPTIDE

The purification of polypeptides is generally challenging because of the presence of many impurities that are very similar to the target molecule. An average polypeptide consists of 20–40 amino acids and many of the impurities have a variation in just one amino acid compared to the target molecule. For such purification problems, reversed-phase (RP) chromatography has emerged as the major tool and it is often applied two or three times sequentially with different eluents (different buffer species, different organic solvents).

For this work, a single RP-chromatography step of a polypeptide has been investigated. The current, optimized purification process which is operated with conventional batch chromatography, has been taken as a benchmark. The aim of the present work was to use the same eluent and stationary phase used in the batch setup and apply it to a continuous chromatographic process, i.e., the MCSGP process, to show the benefits of the latter process in terms of yield, productivity, and solvent consumption. The specifications in terms of the overall purity and the largest single impurity were equal for both processes.

EXPERIMENTAL RESULTS

For the purification of the polypeptide, a benchtop MCSGP unit was used with three columns in C18 stationary phase. Each column was 0.75 cm i.d. x 5 cm to minimize the amount of crude material needed. The operating conditions for the MCSGP-process, including flow rates, gradients, and switch times, were calculated following an explicit procedure that uses mainly information from existing batch elution chromatograms.1 After the operating conditions were set, the unit was started and samples taken at regular time intervals. As with every continuous process, the MCSGP process undergoes a transient phase until it reaches a steady state. This can be observed by the fact that the outlet concentrations, averaged over one cycle, are constant over time. Using the steady state values regarding productivity, yield, and solvent consumption, the comparison with the batch chromatographic process can be carried out. An overview of the experimental results is shown in Figure 3.

Figure 3

All experimental results of the batch and the MCSGP processes are compared in terms of yield and productivity, under the constraint that the same purity specifications are met. The productivity values are given as mass of purified target product per time and per column volume. The squares in Figure 3 show the experimental results of the MCSGP process. Some variations of the default operating points have been investigated, such as different load volumes, different switch times, etc., so the performance in terms of yield and productivity is different. The circles with the error bars in Figure 3 indicate the performance of the current batch purification process, where the left square shows the performance of a single batch run and the right circle shows the performance of several batches including one recycle run with impure side fractions. Consequently, the single batch run has a lower yield compared to the batch-plus-recycle run, but it has a slightly higher productivity. This is because the same material has to be handled twice to run the recycle chromatography, which reduces productivity.

Figure 3 shows that for the purification problem presented, the productivity of the best MCSGP operating point is about 25 times higher than for the current batch purification process. In terms of yield, all operating points of the MCSGP process showed higher values than the batch-plus-recycle purification scheme. For the best operating points, solvent consumption could be reduced by 60%.

To exploit such a performance improvement in the chromatography step in production, various scenarios may be possible. In the most likely scenario, the same amount of material should be produced in the same amount of time. With an increase in productivity of 25 times using the MCSGP process, the column volume could be reduced 25 times, e.g., instead of 50-L column in batch chromatography, the MCSGP process would need only three columns, each with a volume of 0.7-L. This would significantly reduce the footprint of the chromatography unit and the surrounding tanks. Together with higher yields, the cost savings this setup would offer are very promising and indicate payback times for the MCSGP unit of less than one year.

CONCLUSION AND OUTLOOK

The principle of the MCSGP process and its characteristics with respect to existing technologies, such as SMB, has been explained and promising areas of application have been discussed. For polypeptide purification, a batch chromatography step has been compared to the experimental results of the MCSGP unit. It has been shown that strong performance improvement in yield and productivity are possible in the MCSGP setup. The best MCSGP operating point showed a 25-fold improvement in productivity and yields that were 5–7% higher than the batch purification.

Because of the significant potential of the MCSGP process to reduce the cost of chromatographic steps for challenging biomolecule purifications, further applications are foreseeable.

Lars Aumann, PhD, is chief technology officer, Guido Stroehlein, PhD, is chief executive officer, and Thomas Mueller-Spaeth, PhD, is chief scientific officer, all at ChromaCon AG, Zürich, Switzerland, +41 44 633 7748, lars.aumann@chromacon.chMassimo Morbidelli, PhD, is professor for chemical and bioengineering at ETH Zürich, Zürich, Switzerland. Berthold Schenkel, PhD, is head of technology group 1, CHBS, at Novartis Pharma AG, Basel, Switzerland.

REFERENCES

1. Aumann L, Morbidelli M. A continuous multicolumn countercurrent solvent gradient purification (MCSGP) process. Biotechnol Bioeng. 2007;98(5):1043–1055.

2. Aumann L, Stroehlein G, Morbidelli M. Parametric study of a 6-column countercurrent solvent gradient purification (MCSGP) unit. Biotechnol Bioeng. 2007;98(5):1029–1042.

3. Aumann L, Morbidelli M. A semicontinuous 3-column countercurrent solvent gradient purification (MCSGP) process. Biotechnol Bioeng. 2008;99(3):728–733.

4. Stroehlein G, Aumann L, Mueller-Spaeth T, Tarafder A, Morbidelli M. Continuous processing—the multicolumn countercurrent solvent gradient purification process. Supplement to BioPharm Int. 2007 Feb.

5. Aumann L, Morbidelli M, inventors; ChromaCon AG, assignee. European patent, EP 05405327.7, EP 05405421.8, 2005.

6. Stroehlein G, Aumann L, Mueller-Spaeth T, Morbidelli M. A continuous, countercurrent multicolumn chromatographic process incorporating modifier gradients for ternary separations. J Chromatogr A, 2006; 1126(1–2),338–346.

7. Mueller-Spaeth T, Aumann L, Melter L, Stroehlein G, Morbidelli M. Chromatographic separation of three monoclonal antibody variants using multicolumn countercurrent solvent gradient puri?cation (MCSGP), Biotechnol Bioeng. 2008;100(6):1166–1177.

8. Kornmann H. Purification of MAbs: Performance comparison of batch and continuous chromatography (MCSGP); BioPharm Scale Up 2008, Geneva.

9. Tarafder A, Aumann L, Mueller-Spaeth T, Morbidelli M. Improvement of an overloaded, multi-component, solvent gradient bioseparation through multi-objective optimization. J Chromatogr A, 1167(1):42-53, 2007.

10. Miller L, Grill C, Yan T, Dapremont O, Huthmann E, Juza M. Batch and simulated moving bed chromatographic resolution of a pharmaceutical racemate. J Chromatogr A. 2003;1006(1–2),267–280.

11. Paredes G, Stadler J, Makart S, Morbidelli M, Mazzotti M. SMB operation for three-fraction separations: Puri?cation of plasmid DNA. Adsorption. 2005;11:841–845.

12. Li P, Xiu G, Rodrigues AE. Proteins separation and puri?cation by salt gradient ion-exchange SMB, AIChE J. 2007;53(9):2419.

13. Antos D, Seidel-Morgenstern A. Application of gradients in the simulated moving bed process. Chem Eng Sci. 2001; 56:6667–6682.

14. Wankat PC. Simulated moving bed cascades for ternary separations. Industrial Eng Chem Res. 2001;40:6185–6193.

15. Kessler LC, Gueorguieva L, Rinas U, Seidel-Morgenstern A. Step gradients in 3-zone simulated moving bed chromatography application to the purification of antibodies and bone morphogenetic protein-2. J Chromatogr A. 2007;1176:69–78.

16. Mueller-Spaeth T, Aumann L, Stroehlein G, Morbidelli M. Role of cleaning in place (CIP) in the purification of MAb supernatants, accepted for publication in Separation Sci Tech.