CPSA 2011

Science and Technology Coming Together to Make a Difference

October 3 - 6, 2011
Bucks County Sheraton Hotel
Langhorne, PA

Program Abstract

Tuesday, Metabolite Identification Session

A Comparison of Metabolic Profiles Following Targeted and Non-Targeted Analysis Using QTrap Technologies

Veronica Zelesky¹, Rick Schneider², and Alfin Vaz²
Pfizer Global R&D Pharmacokinetics, Dynamics & Metabolism, Groton, CT 1 ADME Screening Group; 2 Biotransformation Group

Metabolic instability, mediated by cytochrome P450, is a common occurrence in most new chemical entities (NCE's) during the lead optimization stage in drug discovery. Although high-throughput screens for liver microsomal stability provide facile means of "rank-ordering" large numbers of NCEs, oxidative "soft spots" in metabolically labile compounds cannot be elucidated from HTS analysis and requires separate, labor intensive metabolite identification studies. Our recent efforts have focused on utilizing rapid analytical methods suited for a discovery environment and comparing the metabolic profiles using both targeted (theoretically predicted) and non-targeted (full scan m/z acquisitions) methods of metabolite ID for nefazodone and several proprietary drugs. A summary of experimental methods and resultant advantages and disadvantages for each technique is presented.

Compounds were incubated [1 μM] with appropriate cofactors in human liver microsomes and hepatocytes using discovery protocols for assessing hepatic clearance. Following precipitation and collection of the supernatant, both control (no substrate added) and fortified samples were injected [10 μL] onto a rapid gradient LC system and, for comparative purposes, data was collected in several modes of analysis using an AB Sciex 5500-QTrap platform. Non-targeted analysis was conducted by acquiring enhanced mass spectrometry (EMS) data and multiple ion monitoring (MIM's). Targeted analysis was conducted by running up to 70 predictive MRM's (pMRM's) that surveyed potential Phase I and II metabolites. Data was analyzed with LightSight (metabolite ID) and MarkerView applications (PCA/multivariate analysis) to elucidate metabolic profiles of candidate compounds.

This comparative unit of work was conducted to explore the benefits and limitations of the aforementioned approaches to metabolite ID. To maximize chromatographic resolution with rapid gradients, a fused core HALO C-18 column (Mac-Mod Analytical) was chosen and resultant analyte base peak widths were ~4-7 sec. Though ideal for metabolite separations, these narrow peaks presented a challenge when trying to acquire large quantities of Q-trap generated data (molecular ion or MRM transition, enhanced resolution, and enhanced product ion scans) across a peak. The challenges were met with good results by carefully adjusting the optimal dwell, pause, and overall cycle times for each experiment. For nefazodone, during this evaluation, the most comprehensive method for detecting and identifying metabolites was the nontargeted, EMS/multivariate analysis approach. With it, novel or unexpected biotransformations could be easily detected. Nearly half of the nefazodone metabolites represented unpredictable cleavage products derived from N-dealkylation or N-dearylation, some of which existed at relatively low concentrations. The attractiveness of the multivariate analysis approach is that no predetermined knowledge is required to assess metabolic profiles since the results are determined by considering statistically significant differences between groups of samples. In contrast, the targeted, pMRM approach yielded good results for being able to identify expected metabolic transitions. However, for those drugs subject to intramolecular rearrangements which aren't predictable, this approach lacks the inherent ability for detection. A final comparison was made using the MIM's approach. Initial results suggest that valuable data can be obtained across a m/z range, but dependent on the drug's mass and biotransformation reaction, detecting low mass metabolites can be limited by the existing background signal. Higher concentrations of metabolites were needed for detection in the lower mass ranges.

Return to program