A n increasing number of pipeline drugs plus higher protein titers and cell densities are among the factors driving growth in protein purification. The U.S. market for protein separation systems was valued at approximately $3 billion in 2006, according to a report by Business Communications Company (BCC).
BCC expects the market to grow at an average rate of 11.1% per year. Sales growth for chromatography equipment, estimated at $944 million in 2004, is rising somewhat faster than the overall market at 13% per year. Meanwhile, sales of membrane filtration equipment was pegged at $538 million in 2006, with estimated growth of 10.8% per year.
Given the number of filtration membrane materials that have been used from lab to production scales, it is noteworthy that one has emerged as the bioprocess standard. That distinction goes to polyethersulfone (PES), which has begun to dominate the marketplace.
Sartorius Stedim Biotech (www.sartorius-stedim.com) found PES as the sole fabrication material in 18 of 19 sterile filtration membranes introduced since 2000. The lone outlier was a PES-PVDF (polyvinylidene fluoride) product. Also, of all significant sterile filtration membranes available today, 18 of 26 use PES exclusively.
Another trend is increasing interest in disposable filtration products, particularly for multiproduct facilities seeking to minimize labor- and cost-intensive cleaning and cleaning validation. Leading filtration companies continue to innovate in response to demand for disposability and scalability.
For example, Millipore’s (www.millipore.com) Millistak+® Pod line of depth filtration products promises scalability of 5–12,500 L. The Pod filter eliminates the need for cleaning validation, since 90% of the flow path is disposable. The Millistak+ HC line of filters incorporate multiple graded-density layers, a microporous cellulose layer to protect downstream filters, and adsorptive, positively charged filter media.
One Filter No Longer Fits All
Not too long ago the leading filtration companies focused on general-function sterile filters that did a lot of things fairly well. “Industry is moving away from a one-filter-fits-all mentality to application-driven filter designs,” explains Maik Jornitz, group vp for filtration and fermentation technologies at Sartorius Stedim Biotech. “Leading filtration product vendors must follow suit with products that satisfy this need.”
Media filtration and buffer filtration have a different set of requirements, for example, that cannot be satisfied by the same device. That level of differentiation, Jornitz notes, “will be cut finer and finer as we move down the road.” Sartorius Stedim Biotech’s Sartopore 2 HF membrane product, with its high-flow design, is tailored for buffer filtration.
“Traditionally, people take one sterilizing filter and use it everywhere,” points out Jerold Martin, svp for scientific affairs at Pall Life Sciences (www.pall.com). “Now they’re looking for bioburden reduction filters or sterilizing filters for specific applications.” The point of application-specific filtration products, he says, is “the best performance at the lowest cost in each application.”
Buffers: A High-volume Application
Larger cell culture volumes and more concentrated process fluids are critical factors dictating the course of downstream purification, Martin continues. “As processes keep growing in size, manufacturers are looking to reduce costs, particularly those associated with generating buffers for chromatography and diafiltration.”
Generally, biomanufacturers treat buffers with sterilizing-grade filters to control bioburden. Since steps downstream of buffer prep like chromatography, ultrafiltration, and diafiltration are aseptic but not sterile, processors can save money by switching to higher flow, higher capacity bioburden reduction filters. One product doing well in this area is Pall’s Supor® UEAV, which features nominal 0.2-micron pore ratings and is validated like a sterilizing filter.
Instead of retaining greater than 107 Brevundimonas diminuta (a small test bacteria) per square centimeter like true sterilizing-grade filters, Supor UEAV is validated to retain 106 B. diminuta from the influent. While falling out of strict standards for sterility, process fluids treated this way are clean enough for chromatography columns and tangential flow filtration.
The significant process advantage is the Supor membrane’s flow rate of 20 L/min/100 mbard (14 L/min/psid), about twice that of a typical 0.2-micron, sterilizing-grade filter. Supor UEAV’s filter size assembly is about half the size of a sterilizing-grade membrane.
Pall still also offers Supor EKV for those continuing to use sterilizing-grade filters. This product uses a 0.65-micron, graded-pore prefilter layer over a sterilizing-grade, 0.2-micron layer. Both filters utilize Pall’s Ultipleat® technology, which makes the cartridge stronger and squeezes more membrane area into a cartridge. The sturdier pleat configuration reduces potential damage to pleats from pressure surges.
Another related trend is the greater use of cellulosic depth filters with or without preliminary centrifugation for cell harvest. These two unit operations have for the most part replaced tangential flow filtration, which is now principally relegated to downstream ultrafiltration. Introduced in August, Pall’s SupraCap™ 100 depth-filtration capsule is an example of a typical depth filter for removal of cell debris from processes of 5–100 L.
Mycoplasma Clearance as a Cottage Industry
Impurity clearance has long been a top priority of bioprocessors. “Mycoplasma contamination control has become a red-hot button in filtration,” according to Jornitz.
Mycoplasma, the smallest known free-standing organisms, have emerged as a consequence of companies switching to animal-derived, component-free (ADCF) media. The vegetable protein hydrosylate (peptone) constituents of ADCF media carry over from the source’s agricultural environment.
Mycoplasma are frequently pathogenic to humans, difficult to detect, and are not easily removed from cell cultures. Bioprocessors who discover Mycoplasma DNA in their cultures generally scrap the batch.
The Parenteral Drug Association has formed a task force charged with identifying methods for detecting and clearing Mycoplasma from bioprocesses.
As more biomanufacturers adopt ADCF media, a cottage industry will likely emerge for Mycoplasma clearance as it did for viruses, says Jornitz. Bioprocessors currently lack robust tools for detecting, inactivating, or removing these bacteria, as well as related validation methods.
Size alone doesn’t guarantee that a filter will protect you from a contaminant. The lack of a retention standard for 0.1-micron membranes thus confounds the solution. While pore size may suggest a safe cutoff for mycoplasma, in practice, the organisms may break through 0.1-micron pores that are believed to retain the bacteria. “Clearance must be validated under real-life process conditions, preferably under actual conditions,” Jornitz points out.
Charged Membranes: Adding Another Dimension
Chemical modification of polymer filter materials can provide a measure of affinity or repulsion between the membrane and biomolecules, potentially enhancing performance.
Charged ultrafiltration and charged membrane chromatography, both specialties of Mark Etzel, Ph.D., associate professor of www.wisc.edu.com), may one day lead to improved impurity clearance and compact even capture media. These membranes separate species on charge as well as size thereby improving selectivity and flux. They reject, that is retain, molecules with the same polarity through electrostatic repulsion.
In theory, processors can use larger-pore membranes, which should better retain the target molecule while permitting more noncharged and oppositely charged species to flow through. “In other words, you can use less membrane area to do the same job,” Dr. Etzel says.
Another advantage of charged ultrafiltration membranes is the ability to separate similarly sized proteins based on different charges, which is not possible using uncharged membranes. For example, a positively charged membrane will retain a basic protein and allow an acidic protein to pass through. Similarly, positively charged mAbs will be retained and concentrated by a positively charged membrane. “This greatly increases a membrane’s separation power, which is a big step forward in ultrafiltration,” Dr. Etzel adds.
Much of the recent ground-breaking work on charged membranes was done by Robert van Reis at Genentech (www.genentech.com), who holds chemistry patents for attaching charges to cellulose membranes. The idea of charged ultrafiltration membranes is not new; van Reis took an under-exploited idea and made it into something that turned out quite useful for bioprocessors.
Charged membrane chromatography, a related idea, has come into and gone out of fashion, according to Dr. Etzel, who has worked in this field for the last 15 years. “It was hot during the early 1990s, then faded away, then came back during the last two or three years.”
Charged membranes’ waxing and waning fortunes are, in part, due to an unreasonable expectation for their use in protein capture to support or replace column chromatography. “That was clearly the wrong job for them,” states Dr. Etzel.
Membranes have to compete favorably with packed-bed chromatography, which they do for clearance of viruses, DNA, endotoxins, and other contaminants that are too large to diffuse into beads in a packed bed.
Dr. Etzek admits that his position is somewhat controversial. “Many people are trying to use membrane chromatography for protein purification vs. contaminant clearance, and that may happen someday for large-scale processes, because membrane capacities are increasing all the time.”
For now, though, Dr. Etzek argues, membrane adsorbers are best confined to small-scale capture and polishing. “Column chromatography still works pretty well and is entrenched at biomanufacturing facilities. It’s hard to unseat an established technology.”
Something Completely Different: Nanobrushes
eMembrane’s (www.emembrane.com) claim to fame is a platform technology that grafts nanobrushes onto the membrane surface. The brushes impart chemical affinity or electrostatic attraction/repulsion to membranes. In removal mode, brushes can selectively remove metals, soluble proteins, viruses, or cells while performing predictably as a molecular-weight cutoff membrane. Nanobrushes are compatible with both capture and impurity removal, according to William Lee, Ph.D., president.
Dr. Lee describes the development of a modified membrane for specific protein separations as “combinatorial.” Multiple test brushes, each with different functional groups, are deposited onto a single membrane to determine which work best.
Further combinations are possible within a functional group by changing substituents on tertiary amines or varying the pH or ionic strength, for example. Other variables include choice of base material and filtration format, for example, film, hollow fiber, nonwoven cloth, plus length and spacing of brush residues.
eMembrane’s standard base material is polyethylene, which is particularly compatible with modification chemistries and is temperature stable.
Nanobrush-modified membranes are provided as a strategic service, which Dr. Lee hopes will eventually evolve into a supplier agreement for process-scale membranes. “Customers come to us not only for R&D or feasibility studies,” Dr. Lee says. “They’re usually interested in scale-up as well, from microgram to gram levels.”