High tolerance to instrument drifts by differential chemical isotope labeling LC-MS: A case study of the effect of LC leak in long-term sample runs on quantitative metabolome analysis

J Mass Spectrom. 2020 Jun 11;56(4):e4589. doi: 10.1002/jms.4589. Online ahead of print.


Metabolomics study of a biological system often involves the analysis of many comparative samples over a period of several days or weeks. This process of long-term sample runs can encounter unexpected instrument drifts such as small leaks in liquid chromatography-mass spectrometry (LC-MS), degradation of column performance, and MS signal intensity change. A robust analytical method should ideally tolerate these instrumental drifts as much as possible. In this work, we report a case study to demonstrate the high tolerance of differential chemical isotope labeling (CIL) LC-MS method for quantitative metabolome analysis. In a study of using a rat model to examine the metabolome changes during rheumatoid arthritis (RA) disease development and treatment, over 468 samples were analyzed over a period of 15 days in three batches. During the sample runs, a small leak in LC was discovered after a batch of analyses was completed. Reanalysis of these samples was not an option as sample amounts were limited. To overcome the problem caused by the small leak, we applied a method of retention time correction to the LC-MS data to align peak pairs from different runs with different degrees of leak, followed by peak ratio calculation and analysis. Herein, we illustrate that using 12 C-/13 C-peak pair intensity values in CIL LC-MS as a measurement of concentration changes in different samples could tolerate the signal drifts, while using the absolute intensity values (ie, 12 C-peak as in conventional LC-MS) was not as reliable. We hope that the case study illustrated and the method of overcoming the small-leak-caused signal drifts can be helpful to others who may encounter this kind of situation in long-term sample runs.

Keywords: LC-MS; chemical isotope labeling; metabolomics; peak intensity normalization; retention time correction.