The Quality Control Session

Analytical Instrumentation: The Quality Control Session

IMAGINE if the global analytical instrumentation industry was a person who walked among us.

Would they be a man or a woman, and what kind of person would they be?

Perhaps she's an A-list movie star, basking in the glory of better-than-projected global market figures? Or maybe he's an athlete, an MVP time and time again, who capitalizes on automation, integration, multiple-capability technology platforms and market consolidation.

Both scenarios might conceivably be phantasmagorical, but circumstances surrounding analytical instrumentation?that's AI to friends?are in the spotlight, just like an A-list actor or an all-star athlete is.

PHARMA'S PAPARAZZI

While life science industries will undoubtedly continue to face ups and downs, the diagnosis for the global life science analytical instrumentation market is solid growth, continual influence on quality control (QC) and chronic cross-pollinating of other industries and technologies.

Figures from "The Laboratory Life Science & Analytical Instrumentation Industry," a report published by Strategic Direction International (Los Angeles, Calif.) last year, have since been revised.

The projected compounded rate of 7.1 percent of annual growth for non-discovery, analytical scale, pharma, biotech and contract lab instrumentation markets-the areas most associated with drug formulation and quality-is actually closer to 9 percent.

The 2005 global analytical instrumentation market is valued at $28 billion, which is about 6 percent more than 2004's $26 billion assessment and approximately 9 percent more than 2003's $24 billion valuation, according to Glenn Cudiamat, vice president of research services for Strategic Directions International.

For the sector close to drug formulation and quality, that means a continued command of approximately 9 percent of the global market.

AI, however, is back on the proverbial psychiatrist's couch, continually working through the looming demands from lab techs and chemists alike for greater productivity in high-performance liquid chromatography (HPLC) applications, while grappling with the challenges of managing data, recalls, withdrawals and a trickle of new products in the drug pipeline.

"We're on pace with what we projected in 2003, but we have revised some of the figures," Cudiamat says, adding that the 2006 global market has a projected 6 percent growth for $30 billion.

If that estimation is realized, AI's pharma formulation and quality side will command about $2.7 billion of the global market, and that's just $4 million or so shy of the $3.1 billion figure that was projected by Strategic Directions for 2008.

Moreover, Cudiamat says there's a lot of talk about process analytical technology (PAT) finding its way into the manufacturing area.

That rapidly growing market is incorporating numerous technologies from the petrochemical area, which are finding applications in the pharmaceutical arena.

"When you're manufacturing a pill, if a parameter is out of spec or not within the threshold, it might impact the overall QC. With PAT, the FDA could end up requiring less testing because you'd have online QC at the manufacturing level," Cudiamat says.

PAT, he says, could augment QC testing and, if implemented, cause other systems to be benchmarked against it, ultimately boosting the QC market.

At a time when 21 CFR Part 11 is now a big piece of life science landscape, the boost from PAT will also have an impact on the electronic record-keeping and compliance-management software market. (See "Conquering Compliance Management," p. 62).

BACK ON THE COUCH

Just like a celebrity, AI is sometimes unable to dodge the niche-snapping life science paparazzi and is left to search for an identity that ultimately needs to be all things to all life sciences.

The challenges facing AI were not only spelled out in "Analyze This: What's driving the Analytical Instrumentation Sector, PFQ, February/March 2005, p. 24, but the article also covered knowing customer demands, market consolidation and integration and trend-tracking.

AI, however, is back on the proverbial psychiatrist's couch, continually working through the looming demands from lab techs and chemists alike for greater productivity in high-performance liquid chromatography (HPLC) applications, while grappling with the challenges of managing data, recalls, withdrawals and a trickle of new products in the drug pipeline.

And it doesn't stop there.

If AI was a person walking among us, he or she would most certainly have the weight of the world on their shoulders, for demands for its penchant for HPLC techniques from the pharmaceutical and other life sciences sectors keep coming, and there are other industries standing in line awaiting turns to pick the brains of the innovative superstar and get that valuable John Hancock of versatility.

Brian Smith, senior director for pharmaceutical market development for Waters Corp., the Milford, Mass., maker of the ACQUITY Ultra Performance LC System, is still seeing a strong uptake in pharma method development, discovery and manufacturing sectors.

While many in the pharmaceutical industry are using its ACQUITY in search of new biomarkers, trace impurities and to support the clinical trials component of drug discovery, other sectors, particular the food and environmental markets, are looking at ways to use LC platforms to complete the QC picture.

"There is also a tremendous uptake with the technology in food safety," Smith says. "This is a market that is growing very rapidly, certainly because of the demand for the quality of food and product recalls. The environmental market is particularly strong right now for testing water for pesticides."

In fact, Waters has collaborated with the Chemical Surveil-lances Lab of Veterinary Sciences Division for the Northern Ireland Department of Agriculture and Rural Development to address food safety and environmental issues using its trademarked UPLC.

According to case study penned by Waters, a UK-based analytical laboratory deployed the UPLC technology in an effort to crack down on the practice among some farmers and processors, who attempt to export poultry and livestock that contain banned substances or legal drugs at concentrations that exceed maximum allowable European Union levels.

In 2005, the lab acquired a Waters sub 2-micron (?m) particle liquid chromatography system that was installed on the front end of a tandem quadruple mass spectrometer. The lab director said it took him about three hours to transfer the 24-minute nitrofuran HPLC separation to a 4.5-minute separation.

The technology's speed combined with the tandem quadruple mass spectrometer meant fast scanning and rapid switching between ionization modes, allowing the lab to complete runs in less time and the opportunity to expand the scope of its multi-residue testing by being able to detect a much larger number of drugs per analytical run. For example, they can now measure 53 MS/MS transitions in a single 4.5-minute run.

The technology has allowed the lab to maximize its surveil-lance testing, according to the case study.

When the lab suspects that a farmer or processor is using veterinary drugs improperly, it has the authority to detain a shipment until the tests have been completed.

Using prior methods, results could take more than five days. If the results come back negative-and the product spoils-the farmer or processor can sue the laboratory for damages, according to the lab. If the lab knows it won't meet the deadline, it has no recourse but to release the shipment and allow potential health risks into the food supply.

"Because of all the protein work and finding of these different biological markers, they are trying to make it into a rapid screen, instead of results in 6 to 8 hours or even days, they may be able to get test results within a few minutes."

"Nothing is as infuriating as to be sitting on a bunch of samples coming up to their deadline, and knowing that you're not going to be able to get them analyzed in time," an unidentified lab manager says.

Curtis Campbell, HPLC product manager for Prominence HPLC system maker Shimadzu Scientific Instruments (Columbia, Md.), says he, too, is seeing a similar phenomenon when in comes to new applications of LC technology.

"We still have the research and development groups that are doing everything they can to find that next blockbuster drug or develop a new method for some environmental screening," Campbell says. "We still have all the users in QA and QC, too."

Lately, however, a new set of end users are applying the more advance techniques of HPLC in a more simple screening fashion. For example, with high-end mass spectrometers, a clinical lab in a hospital setting can, in some cases, replace the swab-and-grow method when determining biological aspects.

"You're taking say a blood or urine sample, doing some type of pretreatment to it and the ideal thing would be to do it all on the instrument," Campbell says.

To do that, he says, such a system needs a reagent system at the front of process to eliminate some of the proteins and biological markers that are not needed for the test. Then, the system would need an online filter with a pre-column, which would then back-flush onto an analytical column to do the separation. The results could then be examined on a mass spectrometer.

"Because of all the protein work and finding of these different biological markers, they are trying to make it into a rapid screen," Campbell says. "Instead of results in 6 to 8 hours or even days, they may be able to get test results within a few minutes."

In fact, Shimadzu, he says, is presently developing methods for hospitals with neo-natal disciplines.

"Because of the critical nature of that discipline, you need results now instead of in two days," Campbell says, adding that time could be critical for the life of a baby. "What this will tell you in a clinical environment is now you'll know which drug to use. You are certainly looking at biological markers, where some [bodily] process is creating this protein that has been linked with disease or a kidney defect or lung function, and because of the high sensitivity with mass spectrometers, it can be done a lot easier."

HPLC OR UPLC

In PFQ's discussions with officials at Waters and Shimadzu, there is certainly a silent debate over the merits of HPLC and the trade marked UPLC, but the two companies share a common diagnosis for the larger-than-life AI, and his or her inclination for liquid chromatography.

While both stick to the marketing regiments handed down by senior management, Shimadzu's Campbell and Smith from Waters are optimistic that aspects of AI's personality are works-in-progress with seemingly endless possibilities.

Shimadzu's Campbell believes that HPLC is sometimes taken for granted and that it has become a commodity technique.

"The thought is that anybody's HPLC will do everything and in some environments that true," he says. "When you start looking at the research and development person, that's where everything changes, and that's where the fun starts for the manufacturers of instrumentation; to try and find that new technique, that new niche."

The drive for higher resolution and greater throughput have many, Campbell says, reinvestigating the basic principles of chromatography and how to optimize the ideology, just as a hospital did for its neo-natal ward.

"It's kind of exciting seeing your instrumentation applied to things like that," he adds.

Smith shares that same sentiment.

"This is an exciting time," he says. "The whole area of analytical chemistry has been on a relatively modest kind of growth pattern, and the tools that are used have seen a tremendous amount of innovation from the analytical instrument vendors. There is a lot of activity and people are really looking at how they can drive their separation systems and take full advantage of these innovations in the industry." ?PFQ

Searching for Biomarkers

An interactive LC/MS Exact Mass Database Streamlines the Process and Saves Time

BY ROB PLUMB, PH.D., CHRIS STUMP AND PAUL RAINVILLE

THE POST-GENOMICS ERA offers the opportunity to understand drug toxicity and efficacy, validate disease models and move to truly personalized medicine. One of the major "omic" sciences facilitating this potential is metabonomics.

Metabonomics is the study of the changes in the endogenous metabolite profile in mammalian systems following a toxic insult or disease progression. As a result of these studies putative biomarkers are highlighted. One of the biggest challenges in metabonomics is to identify these potential biomarkers and derive knowledge of effected metabolic pathways and hence modes of toxicity/efficacy.

The LC/MS information generated on these "biomarkers" is normally defined by m/z and retention time values. Exact mass can be used to generate elemental compositions and reduce the number of potential compounds; however, identification can still be a long and tedious process. The use of an indexed database searchable by exact mass would be of great benefit to the metabonomics scientist.

This application note shows how a new biomarker data processing, interpretation, mining and archiving software tool has been applied to a model metabonomics experiment. The steps in a typical metabonomics study are shown as follows (Figure 1).

Fig. 1: The metabonomics approach using a unique biomarker software tool.

EXPERIMENTAL

Samples

The example data set consisted of four separate urine sample groups spiked with known metabolites at varying concentrations. The set was created by spiking 10 known metabolites into the urine matrix at varying concentrations: (Group 1) 2000 ng/mL, (Group 2) 3000 ng/mL, (Group 3) 5 at 2000 ng/mL and 5 at 3000 ng/mL, and (Group 4) spiking five components into the urine matrix at a concentration of 3000 ng/mL.

The endogenous metabolites used for spiking are shown in Table 1. Prior to analysis, all samples were centrifuged at 13,000 rpm and an aliquot of the supernatant was diluted 1:4 with deionized water.

This unique biomarker software tool processes LC/MS data to extract retention time, m/z, and ion intensity information. The resulting retention time, m/z, and peak intensity information is then assembled into a data matrix. This data matrix, once subjected to multivariate statistical principal component analysis (PCA), highlights ions which contribute significantly to the observed variance in the data. These are displayed in the scores and loadings plot and listed in the "markers" table. The identity of these ions can then be explored by querying the biomarker software's integrated database. The hits generated are displayed in the database viewer and results are automatically returned to the makers table in the biomarker software browser for review and reporting, with a dynamic link back to the raw data (Figure 1 and 2). Results can easily be exported to other statistical packages for further investigation.

Fig. 2: Searching for endogenous metabolites with the biomarker software.

Table 1: List of known metabolites spiked into the same sets.

At the heart of this software is an Oracle relational database, on an enterprise platform. The advantage of this type of database organization is that many types of spectral data can be associated with one searchable database record (e.g., MS and MS/MS spectra, chromatograms, UV data, etc.). Other types of information that can be stored include: chemical structures, exact mass, chemical name, and CAS number. Since the biomarker software's processing and database searching functions are contained in one integrated package, ease-of use and productivity are greatly enhanced (no need to switch between multiple independent platforms).

NAME	M+H (EXACT MASS)
Biotin	245.0960
Corticosterone	347.2222
11-Deoxycortisol	347.2222
Genistein	271.0606
Pantothenic Acid	220.1185
Pantethine	555.2522
Progesterone	315.2324
Phenylalanine	166.0868
17-a-hydroxylprogesterone	331.2273
Tryptophan	205.0977

RESULTS

The sample set was designed to produce four distinct groups after statistical analysis, as shown in Figure 3. The PCA loadings plot indicates the ions have the largest influence on the observed variance in the data. From this plot, we can also see that the biomarker software has correctly identified (in terms of their m/z value) the ions spiked into the sample sets (Table 1). These ions were then subjected to a database search in order to identify their chemical identity (Figure 4 and 5). The results of this search are shown in Figure 6, with the loadings plot and the markers table updated with the matches from the database.

Fig. 3: Principal Component Analysis of the LC/MS analysis. The scores plot shows four distinct groups and the loadings plot highlights the ions responsible for this group clustering. The trend plot displays the ion intensity across all samples for an ion highlighted in the markers table.

Fig. 4: The integrate database browser illustrates the variety of data stored in the biomarker software database.

Fig. 5: Alternative view of the integrated database browser using the tab function to show chemical structure and chromatographic information for the search hit.

Fig. 6: After searching the database, the markers table and the loadings plot are updated with the identity of the ions of interest.

Conclusions

In this study, we have illustrated how this new and comprehensive biomarker software tool can be used to process LC/MS data, highlight the ions responsible for the observed variance in the data, and subsequently identify these ions using an integrated database. Its integrated database can contain a range of spectroscopic, structural and text-based information that can be used to describe the analytes. This comprehensive metabonomics multivariate data processing package improves productivity in several ways: (1) exact mass de-convolution of LC/MS data is more rapid than manual inspection of chromatograms, (2) an integrated, dynamic interface links the multivariate statistical data and the retention time, intensity, and exact mass data for rapid review and interpretation, (3) unlike DDA, the database identification of known metabolites ensures that an analyst only performs de novo structural elucidation on truly unknown analytes, thus saving time. ?

Rob Plumb, Ph.D., a pharmaceutical applications lab manager at Waters Corp. can be reached at rob_plumb@waters.com or 508-482-3610. Chris Stump, a pharmaceutical informatics application specialist at Waters Corp., can be reached at 508-482-2133 or chris_stump@waters.com. Paul Rainville, an applications chemistat Waters Corp. can be reached at 508-482-3539 or paul_rainville@waters.co