May 15, 2014 (Vol. 34, No. 10)

Angelo DePalma Ph.D. Writer GEN

Continuous processing has become a hot topic among bioprocessors. Perfusion cell culture is the most prominent example upstream. But downstream?

Aside from centrifugation, real-world implementations are rare. That is why GE Healthcare’s projects in continuous capture for monoclonal antibody purification could turn out big.

According to the company’s senior director, Günter Jagschies, most top bioprocessing organizations are examining continuous capture or, as he refers to it, periodic countercurrent chromatography (PCC). “All major bioprocessors have continuous processing on their list of potential options,” Jagschies says.

A simplified version of simulated moving bed chromatography, PCC involves the use of three or more capture columns. Once two columns are connected, the lead column is loaded to its maximum capacity. Breakthrough, which occurs relatively early in the loading phase, is captured by the second column. When the first column is saturated with product, it is removed and a third column is connected to the eluent end of the second column.

While technicians wash, elute, and regenerate the first column, the first step is repeated with columns two and three. Then, as the second column becomes saturated, it is removed, and column one is returned to the effluent end of column three.

The technique allows columns to reach full capacity, which normally does not occur in batch mode because of breakthrough. Here, breakthrough occurs intentionally and is part of the process.

Paradoxically this approach uses less media, not more, than conventional batch mode capture. At the point when breakthrough occurs during batch capture, product is distributed unevenly throughout the bed, with higher concentrations toward the top of the bed. Maximum capacity utilization is 60–70%. Capturing breakthrough product in a downstream column allows the first column to become fully saturated, and the substantial investment in protein A resin to be utilized more fully.

PCC makes most sense for affinity chromatography because protein A resins have intrinsically low capacity—one-half to one-third that of ion exchange resins.

GE Healthcare is involved in testing PCC and has successfully deployed it at approximately 20 locations. “But that doesn’t necessarily mean that companies will implement it,” Jagschies warns. The process is mechanically more complex than batch chromatography, and controls are more demanding as well. “There’s always the possibility that engineers on the manufacturing floor will simply reject the whole idea,” he adds.

Whole Greater Than Sum of Its Parts

Related to continuous processing is the view of improving processes holistically, as opposed to focusing on discrete unit operations. This approach, says Michael Phillips, Ph.D., research fellow at EMD Millipore, more efficiently promotes overall process optimization.

One such approach involves linking unit operations, or optimizing how two consecutive steps work together. Dr. Phillips calls this holistic or integrated processing.

To illustrate, he gives an example of aggregate removal in mAb downstream processing. Several technologies exist for aggregate removal, including traditional cation-exchange and hydroxyapatite chromatography, but these technologies often are operated in the bind-elute mode that necessitate large columns with sufficient capacity to bind the product and require increased solution conductivities to adequately recover the product upon elution.

“These work fairly well, but the next step, typically a flow-through anion exchange operation, requires low conductivity,” Dr. Phillips says. This requires either diluting the product or removing the salt. Both approaches add a step and consume resources, while dilution increases process volume, thereby increasing the need for hold tanks. And since bind-and-elute must capture all the product, it consumes copious quantities of buffer and resin.

EMD Millipore is working on a technology (not yet commercialized) that clears aggregates in flow-through mode at low conductivity. Product is ready for the next anion exchange step, at more or less the correct salt concentration.

“This type of strategy is beginning to dominate process development, and vendors are noticing,” Dr. Phillips asserts. “Instead of forcing customers to work through unit operations that are less than compatible, we’d prefer to develop products that allow process steps to operate harmoniously.”

Dr. Phillips could not say if the new product would debut as a resin, a membrane, or in some other format. Regardless, at the right price point it could be introduced as a single-use product.

Integrated processing is part of a larger effort toward what Dr. Phillips describes as “manufacturing of the future.” Given the reluctance of bioprocessors to go off the deep end with regard to manufacturing risk, making unit operations more intercompatible will probably be the first step. “We won’t magically go from batch processing to continuous,” Dr. Phillips declares. “It will be more like an evolutionary approach toward revolutionary change.”


Linking downstream operations more seamlessly is an accepted strategy for downstream process improvement. For example, a technology that is being developed by EMD Millipore clears aggregates in flow-through mode at low conductivity. Product is ready for the next anion exchange step at approximately the right concentration.

Smoother, but Not Quite Continuous

Despite few actual implementations of continuous processing, a good deal of research and development involves linking steps more harmoniously. Such steps may include depth filtration and concentration before protein A capture.

“Continuous processing has always been challenging,” observes Jon Petrone, vice president for global technical service at Pall Life Sciences. “You think it’s going to save time, and it may at process scale when everything is well defined. But development takes time. There’s a lot of buzz, but only time will tell if continuous processing makes its way to production plants.”

Process changes pose engineering challenges, but regulatory hurdles may be even more significant. Recent FDA data for generic drugs suggest that the overwhelming majority of post-approval CMC submissions involve facility, packaging, and controls. No more than 2% involve a significant adaptation to the chemical process.

While not directly applicable to bioprocessing, these figures confirm that even exquisitely defined chemical-pharmaceutical manufacturing processes rarely change. So, to expect bioprocessors to scrap batch processes and related capital equipment in favor of continuous unit operations is probably unreasonable.

“Unless it’s a new process—for example a biosimilar—it may not make sense,” Petrone says, adding that in the short run we are more likely to see unit operation linkage and process intensification.

Cost savings have thus far not driven major biomanufacturers to make wholesale CMC changes after the process is established, and more so post-approval. Given the reliance on platform purification technologies, especially for monoclonal antibodies, Petrone sees product integrity as the only incentive for change.

“Maybe it will happen when a manufacturer has an otherwise intractable problem, that is, one for which continuous processing is the only solution,” Petrone says. “For example, continuous processing could be implemented if it were to reduce processing time for a highly labile product or significantly improve quality. Or maybe it could be used to enable high-volume production that wouldn’t fit into a platform process. But it won’t be implemented to save money.”

Working with Smaller Volumes

Downstream bioprocessing has generally done well to keep pace with higher upstream product titers and cell densities. Thus, the “capacity mismatches” and purification bottlenecks predicted by some experts have for the most part not materialized, or have been solved. To some degree, this is due to upstream efficiencies, which have pushed process volumes downward. While 20,000 L bioreactors are by no means extinct, they are less common in new processes.

“Two thousand liters is a nice capacity to work with, and is ideally suited to single-use technologies,” says Francis Bach, director of purification technologies at Sartorius Stedim Biotech.

Buffer conservation continues downstream, where membrane chromatography for endotoxin removal significantly reduces process volumes and buffer consumption compared with a 60 L anion exchange column, while eliminating the column packing step.

Membrane chromatography will never replace column chromatography for bind-and-elute operations such as capture, and many ion exchange columns will remain. Membrane chromatography, however, fit nicely into platform antibody processes, immediately after the protein A column, and can replace polishing anion exchange columns as well.

Membrane chromatography is also accepted by regulators as a validatable virus removal step. “Column chromatography cannot make the same claim, at least not as easily,” Bach asserts. Part of the problem is variability in column packing, which complicates validation. Through a combination of ion exchange and mechanical-filtration effects, membranes can achieve two- to four-log virus reduction.

Further upstream, Sartorius Stedim Biotech is developing products that minimize reliance on centrifugation, a process Bach describes as “expensive, high-maintenance, and labor intensive,” and perhaps not suitable for handling very high cell densities. Toward this end, the company is modernizing an older clarification technology, diatomaceous earth, which is already prevalent in vaccine manufacturing.

Bach concedes that companies that have made capital investments in centrifugation are unlikely to switch to depth filtration, regardless of the platform. Nor has centrifugation stood still. kSep Systems, a division of KBI Biopharma, has developed the kSep® single-use centrifugation system, which features disposable product contact surfaces. While not enough to cause bioprocessors to switch in mid-development or post-approval, kSep is an attractive option for companies designing new processes.

Here Come the Fragments

The age of the monoclonal antibody is not over by a long shot, but numerous derivative antibody-like molecules and variants are squeezing into biopharmaceutical development pipelines. Manufacturers of chromatography resins have taken notice, and they are commercializing a host of new affinity resins to serve emerging molecule purification markets.

“As the spectrum of molecules broadens, so do challenges for purification tools,” says Paul Lynch, senior product manager at Thermo Fisher Scientific.

Platform processes based on protein A for antibodies will not work as effective as for antibody-like molecules, particularly if these derivatives lack the heavy-chain Fc region. That is why Thermo Fisher acquired CaptureSelect™ affinity products and services from BAC in 2013. Thus far, CaptureSelect products exist for purifying antibodies and antibody fragments, fusion proteins (Fc and albumin), biosimilar hormones, and even viral particles.

CaptureSelect affinity ligands are based on single-domain VHH antibody fragments and can be developed against type of biopharmaceutical. For antibodies and antibody-derived therapeutics, it has been possible to make a number of purification products capable of binding to epitopes that are not recognized by protein A or protein L. Examples range from CH3 for Fc-fusion proteins to CH1 for Fab fragments, irrespective of the type of light chain.

According to Lynch, affinities are comparable to those of conventional antibodies and protein A. “They provide the benefits of affinity purification in a single step, which is what people have become used to.”

GE Healthcare, having similarly recognized this opportunity, sources ligands from Thermo through its Life Technologies business unit.

The emergence of new classes of recombinant proteins guarantees employment for process development scientists for years to come. “We’re probably going to need a different affinity resin for each molecular type [depending on affinity], and there is no guarantee that resins will work universally, even within molecular classes,” Lynch admits. “Eventually, there may be need for two, three, or more resins.”

Choices may extend beyond capture as well, to ion exchange, hydrophobic interaction, and mixed-mode ligands. “We are working in those areas as well,” Lynch adds.

However the market sorts out resins, platforms, and molecules, the goals are constant: robust, reusable, high-capacity, highly resolving resins that get the job done.

Less Pain, More Gain—Faster Setup, Purer Proteins

You can’t eliminate trial-and-error from protein purification, but now there’s a way to speed it up with a new type of chromatography system.
Proteins with high purity are essential for applications such as biochemical characterization and antibody production. However, researchers often settle for lower purity protein since optimizing the purification process can be time consuming and labor intensive.
Why is identifying an ideal purification workflow often so impractical? Because it requires optimization of multiple variables—column resin type, pH, and buffer gradient (%B)—at each step (Figure). This means that a researcher must try a multitude of runs at each step to identify the workflow that produces protein of the desired purity. Traditionally, this process requires individually set-up and programmed runs for each tested column, pH, and %B condition.
Equipped with column switching valve, sample pump, buffer blending valve, and the ChromLab software scouting feature, Bio-Rad’s NGC medium-pressure chromatography system can allow researchers to automate the process of column, pH, and %B optimization. Users can program ChromLab to execute several runs sequentially. The software controls different components of the NGC system, such as:

•  the sample pump that directly injects sample onto the column.
•  the buffer blending valve that allows scouting of buffer pH and %B.
•  the column switching valve that can automatically switch between up to five different columns.

This gives users the ability to scout five columns in the same setup and instrument programming time as that required for one traditional run.


The Bio-Rad NGC system gives users the ability to scout five columns in the same setup and instrument programming time as that required for one traditional run.

Previous articleU.K. Research Organizations Commit to Openness on Animal Testing
Next articlePatentability of 3D-Printed Organs