Send to printer »

Feature Articles : Aug 1, 2010 (Vol. 30, No. 14)

Metabolite Profiling Powered by Mass Spec

MS-Based Methods Are Gold Standard for Identification of Metabolic Byproducts
  • James Netterwald, Ph.D.

Metabolites are byproducts of the body’s response to a drug. In many cases, identifying the correct metabolite could carve the right clinical development path. So, obviously, metabolite identification is a really big deal to the pharmaceutical industry. Metabolites, however, do not pertain solely to drug metabolites—they can be food metabolites as well. Either way, they must be identified by an objective, nonbiased tool that is sensitive enough to detect small quantities of them in body fluids.

In modern times, that tool is the mass spectrometer (MS). The type of MS system used depends on a number of factors including the metabolite being measured, its quantity in body fluid, and the sample type.

At this year’s American Society for Mass Spectrometry (ASMS) meeting, there were a number of new variations on mass spectrometry, novel approaches and methods for metabolite ID, and collaborations that brought together MS technology providers and both basic and pharmaceutical research.

Richard Schneider, senior principal scientist at Pfizer, compared four different metabolic profiling methods using a Triple Quad™ 5500 system (AB Sciex).

“The MS hardware currently available is incredible because it can rapidly scan and simultaneously generate high-quality MS and MS/MS data in a short amount of time, but the difficulty comes in the way that one attempts to interpret that volume of data and conducts structural elucidation studies.”

The project compared four profiling methods that revolve around targeted versus nontargeted analysis using various data-analysis software packages from AB Sciex (MarkerView® and LightSight®), which minimize the time needed for evaluating data from a metabolite ID study.

Targeted analysis is considered a biased approach in which the researcher hypothesizes, based on a drug’s chemical structure, the type and number of theoretical biotransformations (e.g., dealkylations and oxidations). “Conversely, the nontargeted approach is one in which there are no assumptions that go into data acquisition, and it is then necessary to analyze a wide mass range and not just a specific mass domain as in targeted analysis,” Schneider explained.

The model compound for these comparisons was nefazadone—a well-characterized drug often used to investigate metabolic profiling. “Metabolism of nefazadone was analyzed using multivariate analysis of full scan data, essentially scanning from 150 to 1,000 mass units and looking at the relevant metabolites formed.” 

Data was presented from the MarkerView PCA analysis, which compared a control versus a fortified microsomal incubation, and the results were comparable to literature results for the metabolism of this drug. With this approach a new metabolite representing a combination of intramolecular cleavage and oxidation was identified that was not previously reported in the literature.

“We also processed the data through LightSight, and the resultant profile yielded many of the same metabolites as did the PCA analysis. This data was derived from EMS scans, otherwise known as enhanced mass spectrometry data, which utilized the QTRAP functionality and provided enhanced mass resolution and sensitivity for that particular analysis.”

The third evaluation—the molecular ion monitoring (MIM) approach—involved conducting a nontargeted analysis in which up to 350 molecular ions, with masses both above and below that of the parent drug, were analyzed. The MIM  approach allowed Schneider and his co-workers to detect and identify many of the same metabolites as the other approaches, but the method was best suited for detecting those metabolites that existed at higher concentrations.

“The last approach, predictive-MRM [multiple reaction monitoring], is a targeted approach where you can assign up to 300 MRMs based on key product ions that represent theoretical biotransformations for the molecule. This technique proved to be the best method to identify low-level metabolites, but the technique suffers by missing those unusual metabolites that cannot be predicted.

“The take-home message is that you really need to apply a nontargeted approach (PCA, EMS, MIMs) to metabolic profiling to identify those metabolites that are not anticipated. Hardware and rapid scanning functionalities are currently available to acquire rich datasets, both MS and MS/MS, from narrow chromatographic bandwidths.”

Improving Pharmacokinetic Liability

Scott Coleman, Ph.D., director of discovery toxicology and pharmacokinetics at Cubist Pharmaceuticals, described his company’s research into microsomal stability and metabolite identification and how information from this work can help project teams move novel compounds forward by eliminating potential metabolic liability early in the development process.

“We identified a key moiety on a scaffold that was responsible for the metabolism that occurs in vitro, and we could correlate that to an in vivo finding. Modifications around that moiety can help us improve in vivo pharmacokinetics.”

Using liver microsomal preparations and an Orbitrap LC-MS instrument from Thermo Fisher Scientific, Dr. Coleman and his team were able to identify, with a high degree of confidence, the structure of the metabolite of interest and then propose alternate scaffolds and medicinal chemistry opportunities to improve pharmacokinetic liability.

Nontargeted Analysis

Waters recently collaborated with Stephen O’Shea, Ph.D., associate professor of chemistry at Roger Williams University, on a metabolic profiling project. The collaborators developed two different workflows—a targeted and a nontargeted approach.“We developed these workflows for the purpose of resolving metabolic-profiling challenges,” explained Kate Yu, principal scientist at Waters, who is a member of the metabolic profiling business development group that worked on the project.

The entire workflow is called UPLC-QTof-MSE coupled with multivariate statistical analysis for sample profiling. In this workflow, Waters’ Acquity UltraPerformance LC (UPLC) system and Waters’ Synapt mass spectrometer were employed for the metabolic profiling.

“One of the key components was  MarkerLnx software, which allows us to identify the key biomarkers from complex samples,” Yu said. When these samples are separated on UPLC you can see that they are complex, but when you dig down deeper into the chromatogram by using multivariate statistical analysis, for example, you will be able to clearly identify more key biomarkers by using their exact mass.”

Dr. O’Shea is working on developing an aquaculture system for breeding clownfish for use as domestic pets and as model species for other marine tropical fish. “Our targeted analysis focused on developing proper feeds, which have a specific fatty acid component, for these fish,” said Dr. O’Shea, who added that, traditionally, fatty acid analysis had been performed using an HPLC/quadrupole MS before this collaboration work.

With UPLC-QTof-MSE, the analysis is performed faster and with less starting material, which is why the collaboration was undertaken. Using nontargeted analysis on whole fish eggs, the team was able to establish biomarkers to determine the identity of required dietary components, which will enable them to develop better diets for breeding.

“The fish eggs have a nine-day incubation and we were able to do metabolic profiling with Waters technology over that nine-day period, i.e., to see changes in metabolite levels over that period and also pick up major biomarkers that could be investigated.”