August 1, 2006 (Vol. 26, No. 14)
Liquid Chromatography Sees Significant Progress in Novel Processing Applications
Multidimensional chromatography and analysis was a hot topic at “HPLC 2006”. The belief, outlined in detail by Peter J. Schoenmakers, Ph.D., Van’t Hoff Institute for Molecular Sciences, University of Amsterdam (www.uva.nl), is that we have reached the end of any significant improvements in peak capacity that can be gained through traditional methods, so that multidimensional processing and analysis offers the greatest return on efforts.
The questions, he says, are “How much peak capacity can we get from a column and how far can we stretch the gradient?” Using one-dimensional methods, even when optimizing separation methods with high pressure and high temperature, peak capacities remain in the hundreds, Dr. Schoenmakers says, so “this is the end of our possibilities using conventional methods.”
Multidimensional LC extends those options, making it possible to see thousands of peaks. With 2DLC, “everything is separated twice.” Within 30 minutes, a capacity of 1,800 peaks was demonstrated, with two 2-D columns used alternately at elevated temperatures and gradient elution in both dimensions, he says, citing data from Peter Carr, Ph.D., professor at the University of Minnesota (www.umn.edu).
The technology requires only minor modifications, namely, a switching valve, so that one loop is eluted while the other loop is being stored, such that there are two fractions in the works at all times. For such systems, “detection and data analysis are bottlenecks,” Dr. Schoenmakers notes.
An Alternative to Gel Electrophoresis
PerkinElmer (www.perkinelmer.com), working with Northeastern University in Boston (www.northeastern.edu) and Merck KGaA (www.merck.de),developed a new method to map phosphopeptides that offers an alternative to gel electrophoresis.
That method combines 2-D electrically driven planar chromatography and thin-layer chromatography, followed by mass spectrometry to produce three dimensions of separation and analysis that delivers “maximum resolution of the peptides,” according to Ira S. Krull, Ph.D., professor of analytical chemistry at Northeastern University.
This work shows that phosphorylated peptides migrated toward the anode, away from the other peptides in the first dimension, based upon their charge. In the second dimension, they could be distinguished from adventitial peptides, based upon their hydrophilicity.
“We got very good results, similar to 2-D gel electrophoresis, but with a different pattern. So, see what shifts and you know what’s phosphorylated and what’s not.” Consequently, there’s no need to derivatize the sample before MALDI-TOF.
HPLC for Fractionation
Agilent Technologies (www.agilent.com) is using multidimensional HPLC (MDLC) for fractionation and analysis of intact proteins, according to Alex Apffel, Ph.D., senior research scientist, molecular technology laboratories. This method extends Agilent’s work in off-line multidimensional coupling strong anion exchange and reversed phase HPLC by directly splitting the flow from the second-dimension reversed phase separation, so it simultaneously collects fractions and ESI-TOF MS data on intact proteins.
Adding the prefractionation techniques increases dynamic-range capabilities and reduces sample complexity to create an integrated workflow to understand protein-expression patterns, Dr. Apffel explains. And, he continues, using orthogonal information, based on intact protein characteristics constrains the search for such attributes as hydrophobicity. Separations are performed in liquid phase, Dr. Apffel explains, for improved recovery, resolution, and easy automation.
“Compared to conventional 2-D gel electrophoresis, this method separates a broader range of proteins, and is fast, robust, automatable, reproducible, and easy to use.”
In terms of applications, “the MS data can be used for protein characterization,” Dr. Apffel adds and to compare and differentiate samples by deconvoluting partially separated peaks in the 2-D separation space. “Informatics and data visualization play a key role,” he says, “in identifying key features for subsequent identification and analysis at the peptide level. Statistical comparison of multiple samples allows identification of significantly differentially expressed features.
“The challenge in fractionating and analyzing intact proteins is like finding a pico-needle in a stack of needles,” because the dynamic range of protein expression can be very broad (1012).
“Proteomics extracts such vast quantities of data that sometimes we don’t remember the question. Therefore, we need to ask specific questions,” Dr. Apffel adds, to return to hypothesis-driven science. “For example, if identifying key proteins in a process is useful, that may be more important than identifying 100 proteins in a sample.”
Novel ultrahigh-pressure split-less nano HPLC systems are resulting in dramatic increases in resolution and sensitivity for complex proteomic analysis, according to Frank Yang, Ph.D., president, Micro-Tech Scientific (www.microlc.com).
The challenge, Dr. Yang says, is to move beyond atta-mole detection and increase complex biological sample resolution using small diameter nano-HPLC columns with small particles. To that end, Micro-Tech has developed an ultrahigh-pressure splitless nano-HPLC system that operates at pressures of up to 15,000 psi (1,000 bars) with split-less gradient flow rates as low as 100 microliters per minute for columns with inner diameters of 25 to 150 microns.
“There are two ways to further improve resolution in any binary gradient system of a multidimensional processincrease column lengths or reduce particle size,” he adds. In lab tests, he found that increasing column length from 15 cm to 100 cm and using 3-micron particles (rather than 5-micron particles) achieved a greater than two-fold increase in the number of peptides and proteins identified by increasing the operating pressure under the same gradient conditions.
For example, running a 15-cm column at 1,000 psi with a flow rate of 0.4 microliters/minute and a 120-minute analysis time identified 865 peptides and 191 proteins. Increasing the column length to 100 cm and the pressure to 8,000 psi increased the number of peptides identified to 1,217 and the number of proteins to 276. Then, increasing the analysis time to 356 minutes and reducing the acetonitrile gradient percent change from 60 to 35% resulted in the identification of 2,351 peptides and 465 proteins.
“The increased resolution of complex biological samples (mouse brain homogenate, P2 fraction) was demonstrated by using a 100-cm x 75-µm ID capillary column packing with 3-µm particles operated under 8,000-psi ultrahigh pressure,” Dr. Yang elaborates.
Chip-based HPLC for pharmaceutical R&D work holds real promise, according to Yining Zhao, Ph.D., senior principal scientist, analytical R&D, at Pfizer global R&D (www.pfizer.com). Dr. Zhao predicts that chip-based HPLC will become important in discovery and development analytical arenas in the near future.
Before that happens, several misconceptions must be overcome. Namely, that it is only for large molecules, not for routine use. The mindset among manufacturers also must change, he says, from “get the chip out and then decide what to do with it,” to an application-oriented design, based on unmet analytical needs and a determination of whether such a chip can meet those needs.
As advantages, Dr. Zhao says that chip-based HPLC offers higher quality data than column LC, higher levels of integration, multiplexing for parallel analysis, green technology, and portability. In practice, it produced better MS sensitivity and performance, higher productivity, and cost savings.
“In theory, it should be able to replace column-based micro or nanoLC, offering comparable efficiency, same or better spread and capacity, with ease-of-use comparable to MS, and less use of solvent.” In practice, peak shape was very good, he states, and the results were comparable to liquid chromatography. “But, we wanted something better than conventional HPLC,” Dr. Zhao says.
By adjusting the buffers and running subsequent overnight analysis he determined that reproducibility, retention, and separation performance were excellent for small molecules and that the system could be validated.
Another type of chip-LC system featured a UV detector for greater applicability and simplicity for routine analytical environments. A Van Deemter curve showed that results were comparable with column HPLC packing performance.
“For the future, it is possible to integrate everything onto the chip for small molecule ADME bioanalysis,” explains Dr. Zhao, “which has been demonstrated by the design from a group at Caltech.” Dr. Zhao wants to optimize the chip for parallel analysis, multidimensional analysis, and PAT-enabled in-line monitoring as he continues to develop UV-detection capabilities.
Relying upon a single type of analysis, however, is just as ineffective as relying upon a single channel for processing. Instead, biologic therapeutics require an orthogonal analytic approach to help ensure comprehensive analysis, according to Wassim Nashabeh, Ph.D., director, quality control clinical development at Genentech (www.gene.com).
The broader issue, he says, is that “the absolute purity of biologics is difficult to determine. We need to determine the range of each predicted attribute.” Genentech, he says, therefore links purity to in-house reference material, because the function can’t be determined by the principal structures alone.
Dr. Nashabeh advocates quality by design. Likening combinatorics to an iceberg, he says specifications are only the tip, and should focus on molecular and biological attributes for product safety and efficacy, while characterizationthe view just below the water level provides a deeper understanding. Seeing the entire iceberg, however, is only possible when there is sufficient understanding to link process to product qualities.
In the quality-by-design model, Dr. Nashabeh focuses on orthogonal analytical methods throughout all phases of clinical development. For example, he selects attributes by physiochemical and biological characteristics, then develops degradation profiles, and determines heterogeneity to identify relevant types of modifications, and to classify them as either product-related substances or as impurities.
The key, Dr. Nashabeh says, is to “rely on a combination of different profiles to ensure 100 percent confirmation,” rather than relying, for example, on a Lys-C map, a Tryptic map, or an Asp-N map alone. By using such an approach on one example, “when you zoom into the acidic peaks, you see increased complexity,” that otherwise may have elicited little interest, but that can affect results.
Dr. Nashabeh uses a similar process for the biological and immunochemical characterization, describing relevant therapeutic activities, potency, risks, and variations, as well as the rational for selecting specific release potency assays. In Fc glycosylation in recombinant Mabs, for example, “there is variability that affects cytotoxicity, binding, and other features.”
The question, he says, is not which assay to run but which assays to run using reliable GMPs in all phases of clinical development.