Platform Technology for Developing Purification Processes

ABSTRACT

To shorten time to market for new therapeutic proteins, new and fast methods, such as high throughput screening, are needed to speed up downstream processing. The platform technology discussed in this article includes a structural approach that can be used as a general procedure to purify therapeutic proteins. The approach starts with ligand screening and selection-on-a-chip, with the Surface Enhanced Laser Desorption Ionization–Time of Flight (SELDI–TOF) mass spectrometer system. Next, resin screening and supplier selection are performed using robotics, followed by scouting studies under dynamic conditions to select the best resin. Finally, optimization studies of critical parameters are carried out with statistical design approaches (design of experiments). Examples are presented to explain the platform approach for purification development in more detail.

NV Organon

Downstream processing is the most costly part of biopharmaceutical production and there is considerable demand to reduce the costs involved. Also, an increase in the throughput of downstream processing instrumentation is needed to cope with upstream developments such as high throughput screening.¹ In the drug discovery area, rapid systems for one-step protein purification are already common, such as fast methods to isolate new tag proteins or to develop new constructs of monoclonal antibodies.^2,3 In light of these advances, it would be beneficial if related methods could be established to isolate therapeutic proteins in the process development area. Therefore, a platform technology for purifying new biopharmaceuticals was developed (Figure 1). This approach can be applied at the initial screening stage through to the development of a controlled, robust process. The downstream processing platform technology described here is a collection of current techniques used in a structured approach to develop purification processes. For a new purification step in a downstream process, the procedure is summarized as follows. First, ligand selection (i.e., ionic, hydrophobic, hydrophilic, or affinity) is carried out using the Surface Enhanced Laser Desorption Ionization–Time of Flight (SELDI–TOF) system and ProteinChip Array technology.⁴ Second, resins from various suppliers with the specified ligand are selected and screened using robotics. The most promising resins obtained from these studies are dynamically tested by performing scouting experiments to select the preferred chromatographic resin. Finally, the most important parameters (binding capacity, flow rate, pH, salt, etc.) for a given purification step are optimized with experimental design. Figure 2 outlines this platform technology approach when going from small to large scale.

Figure 1

Overall, this structured method for developing new purification processes can speed up overall process development, and also may provide a scalable process. It may also be particularly helpful for process development scientists who are not specialists in downstream processing. However, note that certain parameters (bead size, scalability, reuse, pressure, cleanability, price, binding capacity, etc.), may influence the screening procedure, such as whether certain resins will be tested for use in capture, intermediate, or polishing steps. For example, resins with larger beads and high binding capacity are preferred for capture to increase throughput and prevent clogging, whereas smaller beads are preferred in the polishing phase for better resolution and purity. The following is a detailed description of the various parts of the platform technology.

Figure 2

SELDI–TOF Studies

The method of choice for starting the development of downstream purification processes is the SELDI–TOF technique, using the sophisticated ProteinChip Array technology from Ciphergen Biosystems (Framingham, CA).^5,6 This concept was developed in the 1990s and commercialized by the end of the twentieth century. SELDI–TOF uses coated chips (ionic, reverse phase, hydrophobic, hydrophilic, or affinity ligand groups) for the binding of proteins, a laser focused beam to promote gaseous ions from solidstate matter, and a mass spectrometer (MS) as a detection system. This system has already proven its functionality in clinical research for elucidating new biomarkers in a diverse range of diseases.⁷ The principle of this system is illustrated in Figure 3. Special ligands covalently attached to the chip are able to bind proteins. Selection occurs by washing away nonspecific contaminants (proteins) with pH or salt buffers, whereas the target protein still binds to the chip. Mass spectrometric detection is a powerful analytical method to identify the target protein and monitor the removal of contaminants during washing. This system is ideal for testing pH and salt conditions by determining the conditions under which the target protein binds to the covalently attached ligand. For efficient protein ionization, additional matrix sensitizers (sinapinic acid or alpha cyano-4-hydroxycinnamic acid) are added. These initial screening studies can be performed in a few hours.

Figure 3

Figure 4 displays the mass spectrometric results of a monoclonal antibody (MAb) binding study. The MAb binds to a strong anionic exchange (SAX) chip under mild conditions (e.g., TrisCl pH 8.0). Elution takes place at salt concentrations from 0 to 450 mM NaCl. The salt screening results clearly show that the MAb, with a molecular mass of ~147 kD, elutes around 100–150 mM NaCl, as shown by mass detection. At higher salt concentrations, the ~147 kD signal disappears, indicating that the MAb elutes from the chip. With these preliminary results (binding or elution conditions on a chip), the second step, resin selection with the specified ligand, is performed in a more static set-up. However, some proteins do not present an ionization pattern with MS, therefore detection with the SELDI–TOF system is not always feasible during the initial screening stage.

Figure 4

Robotic Screenings Studies

After initial selection of the preferred ligand (ionic, reverse phase, hydrophilic, hydrophobic, or affinity) chromatographic resins from various suppliers are selected and screened in a special 96-well filter plate (microtiter plate containing a filter on the bottom), using a pipetting robot system in a dynamic procedure. This technique, used in few other biopharmaceutical companies, has only recently been developed following the commercialization of sophisticated new robotic systems (Figure 5). The sample, containing the target protein, is incubated batchwise with the resin for binding, and elution occurs again by using different pH or salt concentrations. The supernatant is collected by centrifugation or vacuum filtration in a microtiter plate, leaving the resin in the filter plate. Protein- or product-specific tests are used for analysis. Figure 6 illustrates the screening results with different salt conditions of a specific glycoprotein. The results show that most resins depict a similar binding profile—elution of the glycoprotein occurs around 100–200 mM NaCl. Based on these screening results, the best resins (with respect to binding capacity, throughput, cost price, etc.) are selected and studied in more detail in the next step, in which scouting experiments are carried out under dynamic conditions.

Figure 5

Scouting Studies

In the scouting experiments, the most suitable resins are studied in greater detail in packed bed columns. The technique became available in the second part of the nineties following the commercialization of new purification systems such as the Akta design (GE Healthcare, Uppsala, Sweden).

Figure 6

Scouting studies are carried out by running the columns in a series with different buffer conditions (salt, pH, buffer, and soon). During both the screening and scouting stages, it is also possible to fine-tune the selected ligand of the resin (e.g., selecting hydrophobic or ionic ligands). The purification results from a hydrophobic interaction step (i.e., scouting various hydrophobic ligands) for a specific glycoprotein are shown in Figure 7.

Figure 7

As observed from the elution profiles, the target protein elutes in the salt gradient depending on the structure of the ligand coupled to the resin. With some ligands, the target protein elutes partly in the regeneration phase, together with other proteins. Most host cell proteins remain bound to the column and elute mainly after finishing the salt gradient. After these scouting studies, the best chromatographic resin is selected and further optimization occurs with experimental design approaches.

Statistical Design Studies

Experimental design is a statistical approach that uses various matrices; response surface and screening designs are the most preferred methods.⁸ Screening designs are normally used to analyze large numbers of parameters to determine which are the most critical; at first, only the main effects of the parameters are tested. The selected parameters are then analyzed in more detail with a response surface design that handles fewer parameters (the most critical ones), but also determines the influence of the interactions between parameters. Overall, experimental design is needed to increase the robustness of purification processes.

Figure 8A

In the example below, the most important parameters—such as pH, the salt and protein concentrations in the loading phase, and the pH and salt in the elution phase—of a specific purification step for a protein, are analyzed with a response surface design. Figure 8 shows the effects of the main parameters (pH,conductivity, protein) in the loading and elution phases on product content. The results in Figure 8A clearly show the effects of the main parameters—pH,conductivity, salt—in the loading (elution) phase over a broad range around the standard operating ranges. The outcome of these studies can be used to set or restrain limits for certain critical parameters that may influence or become critical in this specific purification step. Moreover, optimization in terms of interaction effects, as shown in Figure 8B (interaction between pH in loading phase and conductivity in the elution phase) indicate that not only are the main parameters important, but their interactions are, too. Finally, this platform technology shows that a structural approach for setting up a downstream process (for each purification step) is feasible. Speed, understanding, and robustness may aid in the use of this platform technology as a very useful approach for every downstream process in development.

Figure 8B

This article first appeared in BioPharm Int. 2007 Mar;20(3)44-50.

MICHEL H.M. EPPINK is director of the downstream processing methodology and troubleshooting section of the API/biotech division at NV Organon, Oss, The Netherlands,+31 412 665850, michel.eppink@organon.comRICK SCHREURS is group leader of the section, and ANKE GIJSEN and KEES VERHOEVEN are research technicians.

References

1. Frost & Sullivan. Strategic analysis of world downstream processing markets in biopharmaceutical production. 2004.

2. Wang T, John S, Archuleta S, Jonsson CB. Rapid, high-throughput purification of HIV-1 integrase using microtiter plate technology, Protein Expres Purif. 2004;33:232–37.

3. Eeckhout D, De Clercq A, Van de Slijke E, Van Leene J, Sals H, Casteels P, et al. A technology platform for the fast production of monoclonal recombinant antibodies against plant proteins and peptides. J Immunol Methods. 2004;294:181–7.

4. Merchant M, Weinberger SR. Recent advancements in surface-enhanced laser desorption/ionization-time of flight-mass spectrometry. Electrophoresis. 2000;21: 1164–77.

5. Weinberger SR, Viner RI, Ho P. Tagless extractionretentate chromatography: A new global protein digestion strategy for monitoring differential protein expression. Electrophoresis. 2002;23:3182–92.

6. Seibert V, Ebert MPA, Buschmann T. Advances in clinical cancer proteomics: SELDI–T–of-mass spectrometry and biomarker discovery. Br Funct Genomic Proteomic. 2005;4:16–26.

7. Seibert V, Wiesner A, Buschmann T, Meuer J. Surface-enhanced laser desorption ionization time-of-flight mass spectrometry (SELDI TOF–MS) and ProteinChip technology in proteomics research. Pathol Res Pract. 2004;200:83–94.

8. Montgomery DC. Design and Analysis of Experiments. New York, NY: Wiley & Sons, Inc;2001.