Is Sub-2 Micron Worth It?
For Simon Robinson, HPLC product manager at Shimadzu Scientific Instruments, the most significant trend in HPLC is fast instrumentation that delivers rapid results at higher resolution. “Fast LC has stimulated a lot of instrument sales, and is causing users to look closely at what they’re doing, how their columns work, and which stationary or mobile phases are best for their application.”
Shimadzu’s Prominence HPLC product line includes the Prominence UFLC, an “extremely fast” system operating at standard pressures, the higher-pressure Prominence XR, and the Prominence Nano, a nanoflow system for proteomics.
Robinson delineates cutting edge HPLC instrumentation as those in the 2–3 micron particle size range and sub-2 micron. The former, he says, is optimal in terms of physics, specifically with respect to the ratio of height-equivalent theoretical plates and backpressure. Shimadzu has long played down sub-2 micron systems as too complex for the average user.
“Higher backpressures present more consequences you have to deal with, not to mention higher initial investment and maintenance.” Separating fact from fantasy on particle size, he says, is simple. “The hard part is getting people to believe it. What they learn is that running at 10,000 psi and up is an absolute nightmare on your sanity and wallet. The worst part is when the customer finds out optimizing at a lower pressure can frequently produce better results.”
Tom Jupille, an LC trainer and consultant with LC Resources, is another advocate of optimizing larger particle size systems to avoid the complexities and costs of ultrahigh pressure operation.
Firm in its convictions, Shimadzu promotes LC systems optimized for columns employing 2.2 micron particles which, says Robinson, may be run under normal-pressure conditions.
Bioprocess quality efforts have traditionally been after-the-fact activities rather than pro-active. The notion of Quality by Design, versus testing-in quality, was the impetus behind FDA’s process analytic technology (PAT) initiative. PAT projects have been occurring, particularly at large bioprocessors, but not with the frequency, fanfare, or urgency that might be expected.
For example Shimadzu has an ongoing collaboration, with a major pharmaceutical company, to interface one of its HPLC systems to a bioreactor through an auto-sampler. But this project is “still in the prototype stage,” according to Robinson.
What’s holding back HPLC-based PAT, says DeLand, is “lack of confidence” by an increasingly conservative market. “These reactions and fermentations are worth a lot of money, and a lot is invested in process development,” he says. “But thanks to initiatives from regulators, companies are beginning to open up and embrace the idea of PAT.”
Nevertheless, the tools for PAT exist. Last December, Dionex introduced a new system, the Integral™ process analyzer platform, that works with the company’s ion chromatography and rapid separation LC systems. Integral draws a sample and passes it on to the LC for analysis. The system is suitable for use in process development, pilot, and production systems.
According to George Barringer, Ph.D., CSO of Groton Biosystems, HPLC/PAT has been slow to catch on because it requires changes “in process and focus on the part of users and suppliers.” More complex analyses usually go to an external analytical lab, far from the processing area.
“Moreover practitioners lack the facilities for HPLC, and using it online will require a mindset change.” The other hurdle, Dr. Barringer says, has been instrument suppliers, who focus on analytical laboratories to the detriment of process groups.
Yet HPLC is ideally suited for feedback control, particularly for quantifying the appearance of the main product or consumption of amino acids, carbohydrates, and other nutrients. Groton has demonstrated this with its flagship products, in the ARS line of automated auto-samplers and feedback control systems that connect up to eight reactors to four analytical instruments.
The company has collaborated with Agilent (on HPLC systems), DASGIP (bioreactors), and YSI (bioprocess analyzers). Thus far Genentech, Pfizer, Novozymes, and Biogen IDEC have evaluated real-time analysis that combine Groton autosamplers and various analytical instruments.
Process optimization is the principal driver behind greater use of techniques like LC and LC/MS for peptide mapping and intact protein analysis, says Jeff Mazzeo, Ph.D., biopharmaceutical business director at Waters. This idea also ties in with Quality by Design, which seeks to control critical post-translational modifications through process parameters.
“We’re seeing that users want to use mass spec more and more for protein characterization, even during the later stages of process development,” Dr. Mazzeo adds. These customers generate LC/MS datasets easily, but spend an enormous amount of time analyzing, annotating, and reporting that data. Waters has continuously upgraded its LC and LC/MS software for automating information-creation from raw data. Most notable is the latest iteration of BiopharmaLynx software.
The company has similarly been focusing on systems for faster, higher-resolution, higher-sensitivity peptide maps and extended its branded UPLC product line to intact protein separations, including the introduction of the Acquity® UPLC BEH Glycan, a new UPLC column for glycan analysis.
Regulators have recently been emphasizing host-cell protein (HCP) analysis. Traditionally bioprocessors obtain a number quantifying HCPs generally through immunoprecipitation with polyclonal antibodies. The problem with this approach is a lack of discrimination between benign and immunogenic HCPs. Waters has applied its proteomics technologies to this problem, which allow picking out and quantifying specific HCPs.
One such technique is single-dimension LC/MS. Waters recently introduced a two-dimensional technique based on its nanoAcquity nanoscale UPLC system. Normally, 2-D LC/MS on low-abundance proteins uses ion exchange in the first dimension and reverse phase (RP) in the second.
Waters’ method employs high-pH RP followed by low-pH RP, which provides orthogonal separations, which are sought after in 2-D methods. The benefit is the higher resolving power of RP vs. ion exchange.
One does not normally consider nanoscale methods where sample is not limited, but here the advantage is to exploit instrument sensitivity to overcome the huge concentration dynamic range differences between the therapeutic protein of interest and trace HCPs. Dr. Mazzeo claims a sensitivity of 20–30 ppm. “We’ve seen tremendous interest for this system from customers,” he says.