Over the years, bioprocessors could, at various times, debate whether upstream or downstream operations represented the slow step, relatively speaking, of therapeutic protein production. Now, with protein titers rising inexorably, purification is likely to remain rate-limiting, for a given facility layout, for a very long time.
Cost considerations and ever-rising titers pose tremendous challenges to downstream operations. As a result, processors are much more open to new technologies, says Christian Manzke, Ph.D., director of purification technologies for Sartorius Stedim Biotech (www.sartorius.com). In the past, downstream specialists simply scaled up laboratory purification methods with minimal improvements and not much regard for cost. “Those days are over,” observes Dr. Manzke. Today, he says, any purification technology is potentially fair game, including precipitation and liquid-liquid extraction.
With downstream processing increasingly viewed as a bottleneck, particularly for monoclonal antibodies, bioprocessors are beginning to think not in terms of process volumes, but of the number of downstream processing suites and their capacities.
A side effect of higher titers is higher concentrations of impurities, which stress purification techniques above and beyond simple mass-throughput considerations.
Membrane chromatography is not new in terms of when it was introduced, but its utilization is not as high as it could be. Meanwhile, membrane adsorbers now boast almost as many classes of separation chemistry as conventional chromatography resins.
For example, Sartorius sells a protein A adsorber, Sartobind Protein A, which is suitable for capturing small quantities of protein. The company’s oldest commercial-scale process using a membrane adsorber is now 10 years old. But as few as three years ago, Manzke says, he was still explaining to customers what a membrane adsorber was. “Now, even companies not known for flexibility are using membrane chromatography,” Manzke adds. Sartorius is moving ahead with new chemistries for its membrane adsorber product line, which, Dr. Manzke says, will “open up many more applications” to the technology.
Other notable Sartorius downstream products include SartoClear® P depth filter modules, Virosart® CPV parvovirus-retentive filter, and the SartoClear P multilayer lenticular filter.
Traditional Separations Methods Rule
Not everyone believes that bioprocessors are eager to adopt nontraditional protein-purification methods at large scale. Chen Wang, Ph.D., development consulting engineer, Millipore (www.millipore.com), believes these techniques are not ready for prime time. “My personal feeling is they lack robustness and scalability, and will not become true platform technologies in the near future.”
According to Dr. Wang, chromatography resins are one of the most logical sources of downstream efficiencies as upstream productivity continues to place capacity demands on purification. “You need to be able to crank more mass through a facility, but there are already a number of constraints downstream such as column sizes, tank sizes, and scheduling,” he says. Higher-capacity resins provide a means to retain existing infrastructure, while continuing to meet speed and productivity demands.
An example is Millipore’s ProSep® Ultra Plus, a new protein A resin launched early in 2008. The company claims the highest innate monoclonal antibody binding capacity and flow-rate capability for the resin. ProSep Ultra Plus is based on porous glass particles, which provide linearity between back pressure and flow rate, and predictability during scaleup. “The media allows for a reduction in column size and buffer consumption, and therefore lowers the cost of operation for high-throughput monoclonal antibody purification,” says Dr. Wang.
Manufacturers are increasingly focusing on improving overall downstream purification economics, according to Dave Lescinski, vp, chromatography at Pall (www.pall.com). “Bioprocessors are paying attention to process chromatography measures like capacity, throughput, and scalability, and looking at overall operational efficiencies and ways to reduce costs of environmentally challenging buffers and cleaning solutions while utilizing existing facility design and equipment.”
One way to achieve this is through adoption of higher-efficiency chromatography sorbents and membranes that reduce consumption, disposal, and recycling of costly salts, as well as the use of cleaning agents and water, while effectively removing host cell proteins, DNA, viruses, and other impurities.
Pall addresses some of these needs through its sorbents for mixed-mode chromatography: MEP HyperCel™, HEA HyperCel, and PPA HyperCel. MEP HyperCel sorbent operates through a mixed-mode chromatography known as hydrophobic charge induction chromatography (HCIC). The MEP ligand is immunoglobulin-selective, making it suitable as a protein A substitute in antibody purification. For all three sorbents, desorption is mediated by electrostatic charge repulsion controlled by eluent pH.
HyperCel bind proteins at substantially lower salt concentrations than does hydrophobic interaction chromatography (HIC), which greatly reduces the cost of salts and the disposal of high-salt solutions. Such processes therefore operate with reduced environmental impact and take up less floor area than those employing HIC.
Mixed-mode HyperCel sorbents also enhance process economics by eliminating the diafiltration step usually required following conventional HIC. In this form of mixed-mode chromatography, the target fraction is collected in dilute buffer. The target fraction may also be collected in buffer containing a residual concentration of binding-promoting salt that is significantly lower than what is normally associated with HIC.
In one published study, process engineers at Medarex (www.medarex.com) described a two-step process employing HCIC with MEP HyperCel and cation exchange chromatography, where either technique served as the capture step. Affinity chromatography on protein A sorbent was not employed. So, HCIC on MEP HyperCel can be of value in purification schemes that retain affinity chromatography and those that do not.
Pall sorbents for mixed-mode and HCIC are somewhat more costly than conventional HIC sorbents, which might impact overall process costs, but the savings in salts and salt disposal, plus elimination of a diafiltration step, should more than make up the difference. Another benefit, with respect to conventional HIC, is predictability in chromatographic behavior. Elution behavior is fundamentally related to isoelectric point of the protein, which is not true for conventional HIC.
The trend toward disposable downstream processing makes membrane chromatography quite attractive for companies of all sizes, particularly those with early-stage molecules who seek to minimize resin and buffer costs, in addition to cleaning and cleaning validation.
Throughput and flow rate are critical issues in polishing, and more so at large scale. Typically, process columns are oversized for this step in the process, as the focus is on achieving high linear velocities and shortening process times. Alternatively, the high convective flow rates offered by membrane adsorbers make them an attractive option for polishing, as the desired throughput and flow rate can be achieved with a much smaller footprint.
Membrane adsorbers require a fraction of the equilibration and wetting of conventional sorbents, and because they are disposable, no washing or cleaning is necessary. “The savings becomes obvious when you realize that a five liter device is doing the job of a 150 liter column,” says Lescinski. Considering the column volumes needed for equilibration, washing, cleaning, and regeneration of a polishing column (3-5CVs each step), and a conservative cost of $5 per liter of buffer, processors can spend <$10,000 on buffer for a single run. The 5 liter membrane adsorber equilibrates with about 25 liters of buffer, is used, and then disposed of. Even in the event that customers re-use membrane adsorbers, the savings are still considerable