Improving productivity in downstream processing

Antibody-based therapeutics are expected to continue to be a major source of new therapies for the next decade. However, monoclonal antibodies (MAbs) are among the world's most expensive drugs. Pressure from healthcare providers is driving the need to lower the costs of manufacturing MAbs by as much as an order of magnitude — from $1000 per gram to $100 per gram.¹

So far, the most significant improvement is the increasing titres from concentrations of low milligram to multigram per litre (Table 1).² This increased upstream productivity has subsequently created a demand on downstream processes to provide enhanced capacity and speed. In addition, because only a small proportion of biotherapeutics that enter clinical trials make it to market, companies tend to populate the candidate pipeline quickly, meaning more processes require development in shorter timeframes.³

To solve these multiple demands, several solutions have been developed, including the use of platform technologies and chromatography resin improvements.

Table 1 Upstream productivity increases 1994–2004 (g/L).

Platform technologies

A downstream platform technology is based on experience with purifying a class of products with similar properties. With more than 500 MAbs currently in preclinical studies, monoclonal antibodies represent the greatest number of biotechnology products to which similar purification processes can be applied. In particular, the unique selectivity of protein A has enabled the development of platform technologies for the purification of monoclonal antibodies.

Platform technologies facilitate rapid and economical process development and scale-up, potentially allowing evaluation of a larger number of product candidates, speeding market entry and even reducing validation. Familiarity with a process can also result in more robust processes and better technology transfer from development to manufacturing. Platform technologies create predictable activities and durations that result in a generic timeline.⁴ Further advantages are the use of established vendors for raw materials and standardized waste disposal procedures. Platform technologies are composed of a number of unit operations and methods, and not just purification steps. An example of a platform technology for a monoclonal antibody is shown in Figure 1.

Figure 1

Monoclonal antibodies usually differ from one another in surface charge and glycosylation. By selecting and adjusting unit operations, selectivity is achieved for each MAb. Typically, the order of anion and cation exchange steps might be reversed, hydrophobic interaction chromatography might replace cation exchange chromatography, or hydroxyapatite columns might be used. Before adding or substituting a chromatography step, it is usually advisable to first try optimizing the process by evaluating column load, wash and elution buffers, and pH. Of particular importance is pH, as some proteins tend to aggregate when exposed to low pH. Downstream platform technologies thereby simplify development across the breadth of MAb products.

Most host cell impurities will be very similar to process impurities. By extending the platform to the upstream processes of cell culture and cell clarification, even greater success rates can be achieved for the downstream process. Platform analytical approaches are also being applied to the development of monoclonal antibodies, reducing the time to toxicological and first-in-human (FIH) studies. Other important efficiency gains include fewer protocols and lower complexity of multisite operations, more lenient preliminary specifications for investigational new drugs (INDs), and more robust FIH methods and testing.⁵

Platform technologies can also reduce validation costs. The use of platform technologies, combined with the precision and sensitivity of quantitative polymer chain reaction (QPCR) has helped the evaluation of the efficiency of generic and matrix viral clearance. Anion exchange steps in the platform are used as polishing steps; that is, to remove impurities such as DNA, host cell proteins and adventitious viruses.

Because anion exchange chromatography is commonly used as part of the platform purification process for monoclonal antibodies, a great deal of knowledge has been gained in this area. Most monoclonal antibodies have basic isoelectric points, typically greater than 8, so they do not bind under neutral pH and low conductivity conditions. Experience with fast flow agarose anion exchangers combined with QPCR analysis has shown that bracketing and generic approaches demonstrate efficient removal

from the process of a nonenveloped virus, SV40.⁶

Another potential improvement in validation is minimizing resin lifetime studies for protein A and anion exchange chromatography. Studies with anion exchangers in flow through mode have demonstrated that surrogate measurements such as DNA clearance, back-pressure, and band spreading can predict when columns will no longer provide sufficient virus clearance.⁷ It has been shown that when Protein A columns are multiply cycled, antibody step yield and breakthrough are performance quality attributes that decay before a decrease in retrovirus log reduction values (LRV).⁸

Resin improvements

Today, some chromatography resins can achieve binding capacities greater than 250 g/L of protein. Fast mass transfer rates enable the use of more than 80% of the capacity after only 1 min, and flow velocities can now exceed 1000 cm/h. For ion exchangers, the most commonly used type of resin, these improvements have been achieved through a greater understanding of resin structure and surface chemistry that enables bead design with appropriate mechanical properties, porosity and distribution of charged functional groups.

Binding capacity can influence process economics in several ways. High capacity can reduce column size requirements, which in turn can reduce space requirements and water consumption for buffers and cleaning solutions. Binding kinetics also influence raw material costs per gram of product, though this area is rather complex. In the example shown in Figure 2, resin A has a high total capacity, slow kinetics and low maximum flow velocity compared with resin B, which has low total capacity, fast kinetics and a high maximum flow velocity. At very short residence times, resin B gave higher productivity than resin A; but when residence time was optimized for resin A, it provided a more economical solution. The example also points out the importance of process development for each individual biotherapeutic.

Figure 2

Increased velocity

Improvements in resin design can be illustrated by comparing a classic fast flow, agarose-based resin with a recently introduced resin designed to improve productivity. The increased flow velocity that can be achieved with this second-generation anion exchanger is demonstrated in an industrial column with a 20 cm bed height and a 1 m internal diameter (i.d. [Figure 3]). A constraint of the industrial production environment is the need to work with moderate pressures to achieve adequate flow rates. Realistically, very large columns for capture applications should not need to handle more than 1–2 bar to avoid weight and handling issues. Designing an ion exchanger for capture means striking a balance between bead properties that give high capacity and high resolution, and properties that allow high flow rates at low back-pressures. The packed bed of the new resin enables a linear flow rate of more than 700 cm/h at a back-pressure of less than 3 bar. High flow rates can improve turnaround time in downstream processing by decreasing washing, cleaning, and re-equilibration times. In most cases, sample loading and elution times can also be reduced, though this reduction depends on the properties of the target molecule and impurities. These time savings can result in the production of more batches to meet the quantity demands for high-dose biopharmaceuticals.

Figure 3

Increased capacity

High flow velocity alone does not make a productive process. Binding capacity at high speeds, however, is essential for capture steps. The second-generation resin provides significantly higher dynamic binding capacities compared with classical fast flow agarose anion exchangers (Figure 4).

Figure 4

A high binding capacity of 80–120 g/L immunoglobulin G (IgG [at residence times of 2–6 min and typical pH and conductivities]) could be obtained with a novel, recently introduced cation exchanger belonging to the same high flow agarose generation.⁹

Overall productivity

By combining speed and binding capacity, overall productivity is enhanced, as seen in Figure 5, which compares high-scale productivity for the capture of a target protein from an E. coli homogenate on three different resins. This study demonstrated that it is feasible to purify more than 100 kg of target protein product in 24 h.

Figure 5

Selectivity and purity

Selectivity plays a major role in platform technologies because it reduces the number of purification steps, thereby shortening overall production times and increasing recovery. Specific binding of the target molecule to a ligand-substituted resin enhances productivity by rapidly removing the bulk of the impurities from the feedstream and concentrating the product in one step. This point is supported by the successful use of immobilized protein A as the capture step in the production of most approved therapeutic monoclonal antibodies as well as hundreds in development. Protein A binds human IgGs at the Fc region, which facilitates the isolation of monoclonal antibodies from cell culture harvest. Purity levels are 98% or greater after this one unit operation, greatly simplifying the challenges further downstream and allowing robust, economical purification. Successful therapeutic applications of MAbs have put great demands on production quantities, where expression levels of more than 1 g/L are commonly reported. A new generation of high flow agarose-based protein A resins has been designed to handle large volumes of feed with high titres (Table 2).

Table 2 Challenges of downstream processing to produce large amounts of MAbs, addressed by second-generation, high flow agarose protein A resins.

Cleaning and sanitization

The potential for carryover of product and impurities from one run to another must be addressed for packed columns, and the cleaning protocol must be shown to remove residual materials. Many cleaning and sanitization protocols use NaOH. Another factor is the stability of resins, columns and associated equipment in sufficiently harsh conditions for cleaning and sanitization. In addition, column packing at large scale is time-consuming and costly; maintaining a packed column is essential for economical production of consistent product. Repacking columns requires large quantities of water and buffers column qualification exercises, and time. If reserve columns are not available, the failure of a column because of bacterial contamination, for example, could result in production downtime that alters in-process holding times, complicates scheduling in the plant and creates an overall decrease in productivity. Furthermore, effective cleaning and sanitization, with preserved performance, are essential to reduce the raw material costs of resins. Protein A-based resins used for the capture of MAbs are more expensive than ion exchangers, but their use becomes economically favourable if they can be used for many production cycles. An example of how increased resin stability can extend the number of cleaning cycles that can be performed in situ is shown in Figure 6. Although the dynamic binding capacity of a conventional protein A resin was below 90% of the original capacity for a polyclonal antibody after approximately 20 cycles, the novel protein A resin, designed to tolerate NaOH, was observed to maintain a capacity of 90% up to almost 200 cycles when 0.1 M NaOH was used, and up to about 60 cycles when cleaned with 0.5 M NaOH for 15 min. The resin uses a genetically engineered protein A ligand designed for improved stability under alkali conditions. Other properties were also optimized for MAb capture.

Figure 6

In a case study, a clarified Chinese hamster ovary (CHO) cell culture feedstream containing IgG₁ was purified on a column packed with the alkali-stabilized protein A media for 150 cycles, with cleaning-in-place every cycle for 15 min with 0.1 M NaOH. The yield was consistently 95–100% and purity levels were consistent, as measured by removal of host cell proteins. There was no detectable carryover and no signs of discolourations or deposits on the resin. The dynamic binding capacity remained at greater than 85% of the initial value.

Conclusions

Improvements in feedstream titres translate into the need for improvements in downstream processing productivity. A multiprong approach has enabled downstream process scientists to achieve the necessary productivity. The proven strategies include the use of platform technologies consisting of second-generation resins with increased capacity and velocity, and enhanced stability to harsh cleaning and sanitization conditions.

Key Points

Acknowledgments

The PCR is covered by patents owned by Roche Molecular Systems and F Hoffmann-La Roche. A licence to use the PCR process for certain research and development activities accompanies the purchase of certain reagents from licensed suppliers.

Gail Sofer is director of regulatory compliance

Laura C. Chirica is product manager, bioprocess, both for the life sciences business unit of GE Healthcare (USA).

References

1. S. Farid, Adv. Biochem. Biotechnol., 101, 1–4 (2006).

2. F.M. Wurm, Nature Biotechnol., 22(11) 1393–1398 (2004).

3. www.fda.gov/CDER/guidance/7086fnl.htm

4. G. Slaff, "Application of technology platforms to the purification of monoclonal antibodies," BioProcess International Conference (Berlin, Germany, 2005).

5. J. Phillips and J. Liu, "Platform analytical approaches to monoclonal antibody development," BioProcess International Conference (Berlin, Germany, 2005).

6. S. Curtis et al., Biotechnology and Bioengineering, 84(6), 715–722 (2003).

7. L. Norling et al., J. Chromatogr., 1069(1), 79–89 (2005).

8. K. Brorson et al., 989(1), 155 –163 (2003).

9. A. Grönberg et al., "A strategy for development of a MAb purification platform," BioProcess International European Conference (Prague, Czech Republic, 2006).