June 1, 2017 (Vol. 37, No. 11)
Angelo DePalma Ph.D. Writer GEN
Increasingly, Systems Are Embracing Molecular Intersectionality
Dealing with biomolecular diversity becomes the over-arching theme for downstream bioprocessing as antibody variants, vaccines, recombinant proteins, gene therapies, nucleic acids, peptides, and various combinations thereof enter development pipelines.
Bioprocessors will therefore increasingly rely on alternative purification modes like hydrophobic interaction chromatography (HIC).
“When we developed the Poros™ HIC resins, our goal was to include design features that address current industry challenges,” says Orjana Terova, product manager for purification at Thermo Fisher Scientific. One is the interaction between resin and bound protein driven by kosmotropic or chaotropic salts.
Kosmotropic agents cause favorable interactions between water and proteins; chaotropic salts have the opposite effect. So, unsurprisingly, common high-concentration buffer ingredients ammonium sulfate, sodium citrate, and sodium sulfate may introduce product instability (plus handling and disposal issues).
Thermo Fisher’s HIC resins allow optimization of downstream purification in lower conductivity buffers while still maintaining superior resolution.
Lower salt loads also improve process flexibility. For example, Thermo Fisher’s Poros Benzyl Ultra resins in flow-through mode at low salt concentrations reduced the formation of high-molecular-weight mAb aggregates.
HIC resins operate in either flow-through or bind/elute mode, depending on the conditions. Poros Ethyl, for example, works in bind/elute mode for moderately to considerably hydrophobic molecules. Poros Benzyl operates in both bind/elute or flow-through modes depending on the molecule’s hydrophobicity. But as a general rule, highly hydrophobic resins purify less hydrophobic molecules, and less hydrophobic resins excel at purifying highly hydrophobic molecules.
“It is important to also mention that HIC is a powerful chromatography tool and it is not exclusive to any one expression system or molecule,” Terova explains.
Thermo Fisher’s goal is to provide process development scientists with a suite of Poros HIC resins offering a differentiating range in hydrophobicity suitable for bind/elute and flow-through applications at lower salt concentrations for purifying therapeutic proteins, antibody fragments, antibody-drug conjugates (ADCs), and other biomolecules. “We would like to see HIC as the go-to option in the purification toolbox for impurity removal as well.”
Tweaking the hydrophobicity of HIC resins allows for wide-ranging specificity and flexibility. Poros HIC resins are functionalized with unique ligands to span a wide range of hydrophobicity, which increases the chance of successfully purifying difficult molecules, Terova says.
“In cases where the molecule is highly hydrophobic, or contains a very hydrophobic conjugate, it can be difficult to elute even in low salt. By altering the hydrophobicity of the resin surface, the retention strength of the resin can be increased or decreased to produce a range of resins to cover a wide variety of molecule hydrophobicity.”
Protein A Evolving Formats
Mats Gruvegård, downstream product marketing manager at GE Healthcare, expands on the molecular diversity theme. “We still see strong growth in monoclonal antibodies, but more-diverse company pipelines present new challenges for purification. Established mAb purification platforms sometimes work with novel therapeutics but often must be tweaked, and in other cases new approaches must be taken to achieve desired purification.”
For example, ADCs allow standard mAb purification, more or less, because they comprise an intact antibody plus a much smaller cytotoxic drug attached through a chemical linker. The analytical challenges of ADCs are substantial depending on the conjugation mode. From a process purification perspective, the issue lies in removal of the highly toxic unconjugated drug.
“It’s a new challenge compared with mAbs, a new level of complexity,” Gruvegård says.
Since drugs attached to ADCs are significantly more toxic, final purification is often conducted within an enclosure, so single-use processing can help to minimize operator exposure. The benefits of disposable processing also become apparent with smaller ADC batch sizes relative to processes for unmodified mAbs.
By contrast, antibody fragments lack the protein A-binding Fc region and are molecularly highly diverse. These could require, in Gruvegård’s words, a “different toolbox” for purification.
“If you work with different antibody fragments you will probably rely on a number of purification technologies, depending on the molecule’s structure,” he says.
Here, as with ADCs, high potency becomes an argument for single-use processing.
Bispecific antibodies are also chemically diverse, with varying propensity for aggregation. But to the extent that one has an intact Fc region, protein A would probably be part of the purification train. GE also is promoting ion exchangers and multimodal resins to meet the challenges of bispecifics.
Gruvegård relates the experience of a GE customer that was developing a bispecific antibody drug with modified Fc domains, whose process generated three distinct molecules: one with no affinity to protein A, one with partial affinity (the target), and one with normal affinity. Tweaking chromatographic conditions did not provide sufficient resolution with protein A so the customer employed two different affinity steps, which added time and cost.
GE’s solution was a protein A resin using smaller bead sizes but with typical protein A-antibody affinity. This change resulted in satisfactory purification and elimination of a costly step.
During the course of this project, GE scientists noted that the new resin MabSelect SuReTM pcc could support continuous chromatography as well, which resulted in a new resin specifically for that purpose.
Gene therapies, vaccines, and cell therapies present even more challenges for downstream processing. These categories are so diverse, even within each, that platform purification processes are rare.
“Here, binding capacity is a challenge and many products are susceptible to shear forces,” Gruvegård notes.
GE has introduced Capto Core 700 for intermediate purification and polishing of viruses and large biomolecules. The resin’s multimodal, octylamine ligand provides dual purification based on size exclusion and binding in flow-through mode. The company also sells AVB Sepharose High Performance, an affinity medium for purifying adeno-associated virus in one step.
Heavier Load for Membranes
Membrane adsorbers, also referred to as membrane chromatography devices, have traditionally been used in flow-through mode to replace polishing chromatography steps, particularly in anion-exchange mode. The advantages of membrane adsorbers derive from their single-use deployment (no cleaning and cleaning validation), high throughput, process intensification, low cost relative to ion-exchange resins, and the potential for buffer savings up to 95%.
Sartorius Stedim Biotech has long championed adsorbers and offers them in Q (anion exchange), STIC PA (salt-tolerant anion exchanger), S (cation exchange), and hydrophobic interaction. For example, Sartorius’s Sartobind® Phenyl hydrophobic interaction membrane adsorber enables large-scale polishing at up to ten times faster than conventional HIC resins, according to the company.
Sartorius also promotes membrane adsorbers for capture of viruses and other large species in bind/elute mode.
“Compared with resins membranes provide flow rates that are twenty-five to a hundred-fold faster based on volume,” says Stefan Fischer-Früehholz, Ph.D., senior product manager for membrane chromatography at Sartorius Stedim Biotech.
“Smaller molecules typically bind at lower binding capacity to membranes, with higher binding capacity to resins. But this effect reverses at around 300 kDa, where size exclusion effects within resins begin to dominate, reducing the binding capability. That is the main advantage of membranes for purifying viruses, virus-like particles, large or conjugated proteins.”
For example, a Q membrane adsorber shows tenfold higher dynamic binding capacity for adenovirus, a common gene-therapy vehicle, compared with resins. “Bind and elute processes are driven by binding capacity, which drives demand for these,” Dr. Fischer-Früehholz notes. A major drawback, historically, has been an upper size cap of five liters, a serious negative for vaccine manufacture.
In early 2017, Sartorius launched Sartobind Cassettes, which have the same fluidic design as Sartobind capsules, to facilitate development and scaleup. Users can combine multiple cassettes to provide up to 100 L of capacity. Dr. Fischer-Früehholz reiterates an important point: “This volume expansion was not driven by polishing, but by large-scale virus vaccine manufacture.”
Sartorius Stedim is collaborating with the Max Planck Institute, Magdeburg on a new sulfated cellulose membrane, which exhibits heparin-like affinity for influenza virus—another vaccine application. With approximately tenfold higher binding capacity for viruses than chromatography resins, the new product—scheduled for release later this year—has the potential to enable the rapid manufacture of virus vaccine production during pandemic outbreaks.
“Membrane adsorbers will continue evolving in those directions in coming years,” Fischer-Früehholz tells GEN.
Void volumes become an issue for membrane adsorbers, and a potential battleground for acceptance among the major manufacturers. The theory is simple: process fluids are expensive, and after many operations the cost of wasting product or even “empty” buffers adds up. Early membrane adsorbers routinely held back nine liters of fluid, a volume that Fischer-Früehholz says has been cut in half.
“This has also had a positive effect on improving the breakthrough curve, which while more significant in bind and elute operations is also important for polishing.”
With emphasis on large, open columns for manufacturing-scale bioprocessing, and HPLC (plus variants) for analysis, preparative-scale chromatography is often relegated to stepchild status. The NGCTM chromatography system from Bio-Rad Laboratories fills this particular gap.
The most obvious distinguishing features of the medium pressure NGC system in that regard are pressures up to 3,650 psi and flow up to 200 mL/min. The system serves for preparative work and scouting for optimal separations.
“Several parameters can be screened for optimization, such as the sample, column, pH, buffer composition, and flow rate,” says Candice Cox, global product manager for protein purification at Bio-Rad. “All parameters can be evaluated in an automated, orderly fashion via the scouting wizard in ChromLabTM software, the accompanying chromatography control and data analysis software.”
Bio-Rad also supplies resins for the NGC system that are suitable for lab scale through manufacturing. In addition to process development, the NGC system may be used in scale-down models to troubleshoot pilot- or production-scale purifications.
Cox explains that NGC mass throughput depends on the resin’s dynamic binding capacity, the chemistry used, and column size.
“The NGC can run any column that fits within the system’s pressure and flow-rate specs. The amount of protein varies depending on what the researcher wants to do with the protein downstream of purification.”
Quantities vary from the low-milligram range for small ion-exchange columns to about 300 mg of protein for a larger affinity column.
LC-like preparative chromatography is also used as an intermediate between analytical and production/pilot scale for small molecules. Late last year, Waters introduced four new columns in its Torus™ line of preparative supercritical fluid chromatography (SFC). The columns are designed for scaling up the purification of pharmaceuticals, natural products, and synthetic chemicals.
Torus columns have been under evaluation for commercial purification of small molecule investigational drugs. Gerard Rosse, Ph.D., associate director for structure guided chemistry at Dart Neuroscience, noted that the columns eliminated retention losses and provide “excellent selectivity and peak shape for basic, neutral, and acidic drug-like molecules under conditions with methanol and 0.2% ammonium hydroxide.”
He described the Waters’ 2-PIC (picolamine) column as “a promising candidate in the search for a universal stationary phase for SFC.”