Bioprocessors have always tested protein products during and after manufacturing, but it has only been a few years since terms like “process proteomics” have emerged to describe these activities. Broadly speaking, process omics refers to the application of molecular analytical tools to improve the manufacture of biological products.
Suitable points at which this definition applies are also quite open-ended. While one would prefer analysis to occur during the process to provide feedback for optimizing process parameters on the fly, this will not become common practice among biotechnology companies for several more years, if ever. For now we must be satisfied with discovery-through-packaging as the relevant timeframe for applying the term process omics.
In a recent GEN article, Dan Allison, Ph.D., principal R&D scientist at SAFC Biosciences (www.sigmaaldrich.com), outlined approaches to process omics that include genomics, proteomics, and metabolomics, during cell-line engineering and process optimization, particularly the design of media and nutritional factors (see GEN, November 15, 2006, p. 40).
Dr. Allison described methods that apply both genomics and proteomics to identify cellular pathways associated with high productivity. These pathways can then be manipulated to improve protein production, for example, by over-expression or knockdown experiments, or employed as biomarkers for higher productivity during process optimization.
Similarly, proteomics may be employed to characterize protein expression productivity as a result of genomic manipulations or in response to changes in culture medium. Among proteomic approaches, Dr. Allison described traditional and emerging gel electrophoretic methods as well as more advanced mass spectrometric measurements.
One emerging omics technique, metabolomics, measures panels of small molecule biomarkers for almost any cellular process, either native or in response to process changes. Metabolomics is not as simple as measuring levels of a protein product. Metabolomic information must be correlated to the primary measure (gene, protein) and often extrapolated to be useful.
According to Lisa Bradbury, director, R&D cell therapy, Pall (www.pall.com), genomics methods are most applicable during cell-line optimization and expansion, but less useful during production. Since protein production is the objective, proteomics are the most useful.
“Genomics can tell you where DNA is located in the genome, and if it is being transcribed to RNA, but when you are talking about expressing foreign proteins the rate-limiting step is not RNA production but the ability to process that RNA into protein of interest.”
Through a process proteomics collaboration, Pall combines Ciphergen’s (www.ciphergen.com) SELDI ProteinChip® proteomics-profiling capability with Pall’s Biosepra line of separations products (acquired from Ciphergen in 2004). ProteinChip is a microarray product that interfaces directly with a SELDI mass spectrometer. SELDI mass spectrometry acts as a rapid, straightforward mass detector for proteins ideal for optimizing separation systems, says Pall.
At the time of the Biosepra acquisition, Pall announced plans to establish process proteomics service centers to help customers select and optimize sorbents and membranes for protein purification. It has since opened several centers, operating on a fee-for-service basis, where Pall scientists optimize protein separations on customer feedstock using Pall/Biosepra separations media through a variety of sorbent chemistries.
Services provided at the centers are suitable for expression optimization, purification development, and QA/QC through analysis of impurities. Pall claims that microscale methods and processes are scalable to pilot and manufacturing scale.
Invitrogen (www.invitrogen.com) has begun incorporating omics technologies into its PD-Direct™ bioprocess development services. PD-Direct operates from cell line selection through media development to scale-up for manufacturing. Invitrogen has invested “quite a lot of money and energy” into building PD-Direct into a continuation of the company’s discovery-stage proteomics and genomics customers, says Trent Carrier, Ph.D., director of business development for PD-Direct. “Our goal is to help these clients bridge from discovery through clinical stages and commercialization.”
PD-Direct builds on Invitrogen’s genomics and proteomics toolkits for discovery efforts, plus its Gibco subsidiary’s expertise in cell culture media development. “Connecting the pieces along the development path provides clients with consistent, integrated solutions,” Dr. Carrier says.
Programs like PD-Direct are causing Invitrogen to rethink how genomics and proteomics fit into the overall development scheme, particularly in cell-line development. Most process omics efforts to date, notes Dr. Carrier, have been based on cellular mechanisms. “You plate cells, see which genes are turning on and off, and manipulate or monitor those pathways relevant to cell viability.”
Invitrogen’s approach, which he calls “observation-based,” uses its Revolution™ technology to accelerate cellular genetic evolution into desirable phenotypes—for example high protein titers, faster growth, or longer productive culture life.
Invitrogen is working on a number of contracts using PD-Direct Revolution to improve existing production cell lines for products entering early clinical studies. Revolution is for proteins that are difficult to express, are not secreted, or are inhibitory to cells. “We’ve used Revolution more than a dozen times for these types of proteins, and improved productivity from 5 pg/cell/day to over 20 pg/cell/day,” says Dr. Carrier.
Incorporating omics technologies during development and manufacturing is an ongoing, evolving effort at Invitrogen, Dr. Carrier admits. “To be honest, we’re not exactly where we want to be in terms of having omics tools in place for process development. We’re still working on that synergy with the rest of Invitrogen.”
As process omics evolves, vendors will increasingly seek to leverage their expertise in laboratory and discovery-stage instrumentation to process settings. Beckman Coulter (www.beckman.com) unveiled its ProteomeLab™ initiative, including two instruments, for protein discovery in 2003.
Since then, the company expanded ProteomeLab further by drawing on its expertise in chromatography and capillary electrophoresis to provide proteomic analytical systems suitable for discovery through production. Among these systems is the PA-800 for automated, high-resolution protein characterization and the PF 2D automated 2-D protein fractionation system.
Primarily a discovery tool, PF 2D employs chromatofocusing in the first separation dimension and reverse phase in the second. According to Mark Lies, Ph.D., strategic marketing manager, the PF 2D can produce a high-resolution comparative protein map for a given cell line grown in different culture media by analyzing products according to isoelectric point followed by hydrophobisicity.
“Scientists can then make decisions on a process based on protein differences,” says Dr. Lies. Protein identities are then confirmed through mass spectrometry.
The ProteomeLab PA 800 Protein Characterization System, by contrast, operates on a capillary electrophoresis platform. The system, which has been widely adopted by the biotech industry as a QC tool, is packaged with off-the-shelf assays, chemistries, and protocols. Included chemistries include CE-SDS (a replacement for SDS-polyacrylamide gel electrophoresis) for molecular weight separation, capillary isoelectric focusing (separation based on isoelectric point and charge heterogeneity), and carbohydrate profiling, essentially a microheterogeneity assay.
The 800 system looks at protein purity, heterogeneity, and ultimately, identity. “When you take all this information and put it together, you can construct a fingerprint or identity for that particular molecule and confirm its identity by interfacing directly to a MS,” says Dr. Lies.
Focus on Early Stages
Process omics has its most profound impact not on processing or in-process analytics, but during cell-line selection, clonal expansion, and the selection and development of cell culture media.
Many cell culture experts, most notably Florian Wurm, Ph.D., at the École Polytechnique, believe that advances in media have been primarily responsible for the 100-fold rise in protein titers since the birth of biotechnology. Dr. Wurm has developed high-throughput technology to optimize and compress development times for media through a more-or-less traditional approach: altering one or more variable(s), measuring protein output, and analyzing results statistically.
One high-throughput method for measuring protein output is with CellSpot™ technology from Trellis Bioscience (www.trellisbio.com). CellSpot provides direct, simultaneous readouts on individual cells for multiple characteristics, for example protein secretion, surface proteins, and intracellular signal transduction pathways. “CellSpot helps you find the needle in a haystack,” says Trellis president Nolan Sigal, M.D., Ph.D.
Traditionally, cell-line development occurs through screening hundreds to thousands of individual cell culture supernatants. By taking the pooled average for all the cells in the culture, this method misses the individuality of cells within heterogeneous populations. CellSpot assays the productivity of up to millions of cells simultaneously, quantitatively, and individually after minimal cell growth periods. Supernatant monitoring requires much longer growth.
These advantages lead to significant reduction in cell-line development times, by up to half, for both library screening and sub-cloning. For the latter, CellSpot sensitivity is adjusted to account for smaller productivity differences among cells.
CellSpot combines nanoparticle-based detection, automated digital microscopy, and software to make sense out of protein-secretion data. The polystyrene nanoparticles, measuring 100–300 nm in diameter, are coated with an antibody to the target protein and a dye to trigger the light emission used for detection, similar to a sandwich ELISA assay.
CellSpot works for all mammalian cells, adherent or nonadherent, and expression systems, according to Dr. Sigal. Cell libraries are plated on a surface modified to capture the antibody or recombinant protein secreted by individual cells located within virtual wells approximately 100 microns across. Immobilized proteins are then labeled using nanoparticles, and protein quantified through the microscopy/ software combination.