OR WAIT 15 SECS
Choosing the right tools to enhance the process.
The number of biotechnology-based human therapeutic products in the late-stage pipeline along with the average cost to commercialize a biotech product has been steadily increasing over time. In addition, the biotech industry is facing unprecedented challenges in the form of a sagging global economy and rising regulatory expectations. Companies have to continue to evolve their approaches to be more efficient with respect to time, resources, and cost. This article describes some of the technologies that can help optimize time and cost of biopharmaecutical manufacturing.
Creating processes that are not only optimal and robust but also economical in a time and resource efficient manner is the Holy Grail for the biotech industry today. Regulatory initiatives such as Quality by Design (QbD) and process analytical technology (PAT) require an improved process understanding earlier in the product lifecycle.1–4
Recent advances in fermentation and cell culture processes have increased cell densities and protein titers in process fluids, requiring reliable upstream technologies to handle growing production demands.5,6 Moreover, within the bioprocessing industry there is a strong trend toward the application of single-use technologies, eliminating the expense of cleaning and cleaning validation, a major cost factor involved in operating reusable systems. One of the commonly used technologies for primary separation after fermentation is tangential flow filtration (TFF).7 TFF processes have traditionally used cassettes or hollow fiber formats. Typically, TFF cassettes are installed in stainless steel housings that require post-use cleaning and validation of cleaning. The installation and preconditioning steps before process use are time consuming and require a reasonable level of skill to avoid potential malfunction. Hollow fiber filter systems provide an alternative primary separation technology. However, this open channel format tends to require relatively high pumping rates for optimal performance. Achieving these high rates with fully disposable equipment can be difficult.
Membrane chromatography has already proven to be a powerful alternative to traditional packed-bed chromatography in flow-through operations, such as polishing for the removal of viruses and contaminants in biologics manufacturing.8 Case studies have been described for the purification of monoclonal antibodies (MAbs) with a high isoelectric point and have demonstrated the popularity of their implementation. As flow-through utilization has expanded, membrane chromatography applications have also included the capturing of large molecules.9 Such bind-and-elute applications imply the demand for higher capacity and larger surface membrane area compared with flow-through applications.
Manufacturers of biopharmaceuticals are required to characterize the ability of key process steps to clear viruses. Typically, the entire manufacturing process is qualified to attain a cumulative virus reduction factor, which significantly contributes to the documentation of virus safety. In many of these manufacturing processes, a virus retentive filter is used to achieve a robust and effective virus clearance step. The key to implementing virus filtration is the assurance of virus retention and device integrity while maintaining high productivity and ease of use.
Anurag S. Rathore
While developments in chromatography media (resins) have brought the technique much closer to practical and theoretical limits of binding capacity, future advances are likely to focus on new surface chemistries and a more efficient use of chromatography as a unit operation. One way to gain productivity improvements is to use smaller equipment or the similar size equipment in a more time-efficient manner.
This article is the fifteenth in the Elements of Biopharmaceutical Production series and will present some of the most promising technologies that are being introduced in biotech manufacturing by some of the major biopharmaceutical vendors.
The Kleenpak tangential flow filtration microfiltration (MF) capsule from Pall Life Sciences (East Hills, NY) uses an Ultipleat pleated filter cartridge and Kleenpak Nova capsule with their operation based on TFF principles. The capsule is configured similarly to hollow fibers or cassette formats with feed, retentate, and permeate ports. As Figure 1 shows, the pleated construction is used inside the capsule to create multiple parallel flow channels while increasing the effective membrane area in a reduced footprint. It is simple to install and requires limited preconditioning before exposure to process fluid.
This technology has been tested with representative industrial strains such as Chinese hamster ovary (CHO) cells, bacteria such as E. coli, and yeast. Feed volumes of 50–100 L have been processed per single full-scale capsule (0.5 m2), depending on cell type and characteristics. The capsules can be manifolded together in series or in parallel to accommodate larger process requirements. For example, a six capsule assembly can process up to 1,000 L. Stable performance has been observed for monoclonal antibody (MAb) clarification from CHO cell feed with starting cell counts up to 1x107 cells/mL and step yields of greater than 98% have been obtained. For clarification of hybridoma cells expressing monoclonal antibodies, filtrate flux rates of 70 L/m2/h (LMH) to a volumetric throughput (VT) of 180 L/m2 have been obtained with a protein transmission of 100%. Protein expression for these cells was approximately 0.1 mg/mL.
A comparison of the Kleenpak TFF technology and two different types of cassettes for concentrating E. coli with all three devices incorporating membranes rated at 0.2 µm (nominal) is presented in Figure 2. All three devices were operated at a crossflow flux rate (CFF) of 1.5 L/min/ft2. A five-fold volumetric concentration followed by a three-fold diafiltration buffer exchange was performed. The data shows that all three devices can be operated at a flux rate of 30 LMH throughout both concentration and diafiltration stages. Although the start and finish transmembrane pressure (TMP) values are different for the Kleenpak TFF filter compared to the cassette formats, the TMP gradient is similar for all three technologies and suggests stable operation. The flux and throughput performance of all three technologies was similar during these trials. Figure 3 presents a comparison of a Kleenpak TFF MF capsule with a Pall Microza hollow fiber module for clarification of IgG from a CHO cell culture with a cell density of approximately 2x106/mL. Each format was operated at a pumping capacity or crossflow flux rate (CFF) of 20 L/min/m2. It is seen that the Kleenpak TFF outperforms the hollow fiber module by processing more than three times the filtrate volume for a similar terminal TMP.
Experiments were conducted with a recombinant E. coli solution at a cell density of approximately 4% w/w for production of a plasmid used for gene therapy treatment. The primary objective of the application was to concentrate the E. coli feed by a minimum concentration factor of 10-fold using a single-use technology. Initial trials were conducted with the small-scale device and the filtrate flux rate for concentration was selected to be 22.5 LMH. It was estimated that six full-scale modules (membrane area of 3 m2) will be required to process 500 L of feed at a feed flow rate of approximately 21 L/min (CFF of 7 L/min/m2). To demonstrate scalability to a series and parallel configuration, a trial was conducted to compare the performance of a single full-scale module against the performance of three full-scale modules operated in series. The filtrate flux rate was maintained at 22.5 LMH for each module throughout the trial. Pressure gauges were installed throughout the system to enable accurate TMP determination across each module. The results shown in Figure 4 indicate that each of the three capsules manifolded together in series provided a VT versus TMP curve very similar to a single module operated in isolation.
In summary, the Kleenpak TFF MF capsules' single-use TFF technology offers major benefits for cell harvesting and clarification. The self-contained format eliminates the need to purchase and install holders or housings and also reduces exposure to biologically active solutions. This is a critical consideration for manufacturers of vaccines and other products that present potential workplace hazards. The optimized feed channel screen provides good performance at low flows and the technology is scalable and reproducible.
Sartobind Phenyl membrane adsorber from Sartorius Stedim Biotech (Goettingen, Germany) uses hydrophobic interaction chromatography (HIC) principles and assembled into a 30-layer radial flow process capsule (Sartobind). It is based on hydrophilic regenerated stabilized cellulose with the hydrophobic phenyl groups covalently attached to the cellulose matrix. The novel macroporous membrane structure with 1–3 µm pores has been designed for high flow rates and high binding capacities. As seen in Table 1, the binding capacity for a MAb on HIC membrane is found to be at par with conventional HIC resins even at significantly higher flow rates. Two MAbs were loaded onto a membrane device and on a column at specified flow rates. The bed height for the column was 30 mm and 8 mm for the membrane. The dynamic binding capacity as illustrated in Figure 5A was similar for the two formats. Combining the advantages of membrane chromatography with virtually no diffusion limitation, shorter processing time because of convective flow, and a ready-to-use disposable capsule format, the Sartobind HIC membrane adsorber represents a new membrane-based tool applicable for flow-through as well as bind-and-elute applications. It provides a robust alternative capturing large and unstable molecules that need to be processed rapidly.
At least two subsequent polishing steps are typically implemented in MAb downstream processes. The Sartobind anion exchange membrane chromatography in combination with a hydrophobic interaction membrane chromatography offers different selectivities and could reduce host cell proteins, high molecular weight aggregates, DNA, and leached protein A to acceptably low levels that assure safety of the product. This may enable the downstream process to be just a one-column-step purification process.
Figure 5B shows the effect of concentration of ammonium sulphate in 50 mM potassium phosphate buffer pH 7.0 on the static protein binding capacity of the phenyl membrane compared to the base cellulose membrane (without phenyl functional groups). Data shown on binding of human polyclonal antibody to the membrane indicates negligible nonspecific interaction from the cellulose base matrix. Scalability of the hydrophobic membrane adsorbers is further shown in Figure 6. The study involved loading of human polyclonal IgG in 0.9 M (NH4)2SO4, 50 mM potassium phosphate, pH 7.0 on Sartobind phenyl membrane adsorber at 10 mL/min (Nano, 120 cm2/3 mL) and 500 mL/min (5-inch capsule, 5,000 cm2/125 mL), respectively. The amount of polyclonal IgG was normalized to membrane area. The normalized breakthrough curves of Sartobind phenyl Nano and 5" capsules in Figure 6 represent a successful 42-fold scale up.
To summarize, the newly designed membrane structure in combination with appropriate hydrophobic surface functionalization provides a new scalable and disposable tool for the large-scale purification and separation of biomolecules as well as for polishing applications in the biopharmaceutical industry. The unique membrane-based technology addresses the requirements for high throughput production while reducing process time and complexity (no packing and packing testing) and adding flexibility.
Parvovirus filters are commonly used as means of effective removal of viruses from biological products by manufacturers as required by applicable industry regulations.10,11 The Viresolve Pro Solution from Millipore (Billerica, MA) is based on a dual layer polyethersulfone (PES) membrane and is designed to simultaneously deliver high parvovirus retention, capacity, and flux. It is also supported by an extensive safety assurance package. The formats are fully disposable, shipped gamma irradiated, and are caustic stable. The binary gas test (BGT) has been developed as a manufacturing release test and provides an increased sensitivity over air-water diffusion in its ability to detect small defects that could negatively impact virus retention. The expectation is that high capacity, flux, and fast set-up, combined with easy integrity testing, would translate into productivity gains from process development to full-scale production, thus reducing the cost of ownership for the virus filtration unit operation.
Table 2 shows the results of a mice minute virus (MMV) retention study where grab samples were collected from two MAb solutions at various flow decay points. These samples were collected to determine if virus passage would occur through the Viresolve Pro Micro device as the membrane became increasingly fouled. Challenging the membrane at high flow decay tests the robustness. All feed and filtrate samples were assayed using validated TCID50 procedures and large volume assays were performed for filtrate samples with a total volume of 4 mL assayed for each sample. The log reduction value (LRV) was determined as the log10 of the ratio of input virus load and the output virus load as shown in the calculation below:
in which v1 and t1 are input volume (mL) and virus titer (TCID50/mL) respectively and v2 and t2 are output volume and virus titer (TCID50/mL) respectively.
Three Viresolve Pro Micro devices were tested for each MAb. The MAb 1 solution was approximately 6.0 g/L protein concentration. The MAb 2 solution was approximately 4.5 g/L protein concentration. After adding a 1% v/v MMV virus spike, feeds were filtered through a 0.22 µm filter. As can be seen from the results obtained with MAb 2, even at 90% flow decay, no passage of MMV was observed. With MAb 1, no passage of MMV was observed even after 13 kg/m2 of the protein was challenged. It was not possible to run to such high flow decays with MAb 1, because adequate feed material was not available. The results presented in Table 2 demonstrate robust virus clearance performance across a range of volumetric loadings, mass loadings, and flow decay points.
Figure 7 shows capacity performance of Viresolve Pro Micro devices on 16 proteins tested. Most feed streams were tested at 30 psig (2.1 bars) using constant pressure operation. No prefiltration was used except the industry standard practice of a sterilizing grade filter. The achieved protein mass capacity averaged 6,900 g/m2 when limited to 4 h processing time and 75% flow decay. The results demonstrate achievement of high mass capacities and short processing times with Viresolve Pro devices.
Further, the BGT has been developed as a new, proprietary release test that uses a two- component gas mixture, in which there is a large difference in permeability between the two gases across a wetted membrane and concentration is measured on the downstream side. A deviation from the expected concentration is an indication of the presence of a defect that could negatively affect the virus-retention capabilities of the device. Data presented in Table 3 demonstrates the superior sensitivity of this test in comparison to the traditional air-water flux test. Three devices, constructed from the same membrane lot, were tested with air-water diffusion, BGT, and virus retention testing. Although the BGT could discriminate among devices with small differences in retention, the air-water diffusion test could not.
In summary, the Viresolve Pro Solution can provide high capacities and high viral retention under a range of processing conditions. This performance allows for increased productivity in manufacturing operations, often reduces the system size required for manufacturing. The flexibility of the fully disposable devices allows for ease of use and speed in process development and manufacturing. Enhanced release testing delivers a higher degree of virus-retention assurance.
As discussed above, single-use technologies offer higher productivity in terms of process time (cleaning is not required) as well as floor space (no cleaning equipment). However, to maximize the benefit, the technology should provide the whole unit operation. For example, in the case of chromatography step, using a single-use column alone yields only partial benefit as the chromatography skid still requires cleaning post-use. The ReadyToProcess program at GE Healthcare Bio-Sciences AB (Uppsala, Sweden) aims to provide a more integrated solution that includes the prepacked, presanitized, and prequalified columns and a chromatography system with a disposable flow path. The flow path is also presanitized and contains an air-trap, UV and conductivity meters, and a flow meter. The flow path can be assembled on the system in less than 10 minutes. As seen in Figure 8, implementation of this technology could lead to reduction in changeover time by up to three days, from seven to four days. This reduction would give a 20% to 30% increase in the number of batches produced per year for a 49-week production schedule based on seven or 14-day production campaigns, respectively. Furthermore, if changeover time is reduced to no more than one day per production line, without disrupting any of the operation ongoing in the production suite (the parallel production line concept), then the increase in the number of extra batches made per year could be between 40 to 160%. The exact efficiency depends on the changeover time and the length of a single campaign. While a similar increase in number of batches could be achieved by using duplicate columns and skids such that the cleaning is performed when the column and skid are offline, this will require a significantly higher capital expenditure and floor space for the duplicate columns and skids. Another possibility is to use a single system for all chromatography steps during manufacturing of a single batch. This will result in reduction of the cost of equipment by 66% as well as lower floor space requirement.
Figure 9 presents a comparison of break-even plot for conventional and ReadyToProcess based operations. The analysis assumes that: 1) each batch produced generates the same level of revenues, either current or future; 2) the fixed cost for the conventional operation is three times higher than for the ReadyToProcess one; and, 3) the operating cost per batch for the conventional technology is 50% lower. The two arrows in the figure show the level of net income after tax for the reference (conventional) case and the case when 40% more batches can be produced per year, respectively. The increase in the number of batches is related to the decrease in changeover time. It is seen that the higher operating cost associated with the use of single-use technologies is outweighed by significantly lower capital cost contribution. Thus, an introduction of such technologies results in a higher net income as long as number of batches produced using these ready-to-use technologies is larger than the number of batches that could be produced using conventional techniques. A real revenue curve for an average biologic is expected to be steeper than the revenue curve shown in Figure 9.
Technologies enabling quick changeover times, and at the same time providing high manufacturing flexibility allowing quick responses to market demands, will become more popular because they offer higher productivity as well as better economics.
Xuemin Liu is marketing manager and Mike Collins is senior R&D engineer, applications, both at Pall Life Sciences, East Hills, NY; Nathalie Fraud is senior scientist, purification process development at Sartorius Stedim North America, Edgewood, NY; Jennifer Campbell is field marketing manager and Inese Lowenstein is director, virus safety solutions, both at Millipore Corporation, Billerica, MA; Karol M. Lacki is R&D staff scientist at GE Healthcare Bio-Sciences AB, Uppsala, Sweden; and Anurag S. Rathore is the director of process development at Amgen, Inc.,
Thousand Oaks, CA, 805.447.4491, email@example.com. Rathore is also a member of BioPharm International's editorial advisory board.
1. Rathore AS, Winkle H. Quality by Design for pharmaceuticals: I. Regulatory perspective and approach. Nature Biotechnol. 2009;27:26–34.
2. Kozlowski S, Swann P. Current and future issues in the manufacturing and development of monoclonal antibodies. Advanced Drug Delivery Rev. 2006;58:707–722.
3. Yu L. Pharmaceutical quality by design: product and process development, understanding and control. Pharm Research. 2008;25:781–791.
4. Rathore AS, Saleki-Gerhardt A, Montgomery SH, Tyler SM. Quality by Design for pharmaceuticals: industrial case studies on defining and implementing design space for the process. BioPharm Int. Part 1: 2008;21(12):37–41 and Part 2: 2009;22(1)–44.
5. Shukla AA, Rao JR Harvest and recovery of monoclonal antibodies from large-scale mammalian cell culture. BioPharm Int. 2008;21:34–45.
6. Roush DJ, Lu Y. Advances in primary recovery: centrifugation and membrane technology. Biotechnol Progress. 2008;24:488–495.
7. Wang A, Lewus R, Rathore AS. Comparison of different options for harvest of a therapeutic protein product for high cell density fermentation broth. Biotech Bioengg. 2005:94;91–104.
8. Fraud N. Membrane chromatography: an alternative to polishing column chromatography. BioProcessing J. 2008;7:34–38.
9. Zeng X, Ruckenstein E. Membrane chromatography: preparation and applications to protein separation. Biotechnol Prog. 1999;15:1003–1019.
10. International Conference on Harmonization. Harmonized Tripartate Guideline: Q5A Viral Safety of Biotechnology Products Derived from Cell Lines of Human or Animal Origin. Fed. Reg. 63 (185) 24. 1998 Sep: 51074. Available from www.ich.org/LOB/media/MEDIA425.pdf.
11. US Food and Drug Administration. Points to consider in the manufacture and testing of monoclonal antibody products for human use. CBER. Rockville, MD; 1997. Available from: www.fda.gov/cber/gdlns/ptc_mab.pdf.