Advanced Mass Spectrometric Strategies for The Selective Detection of Oxidized Lipids

Abstract

Oxidized lipids are involved in inflammation, immune responses, and disease progression, making their detection important for many fields of biomedical research. However, analyzing the oxidized lipids is difficult because of their low abundance, chemical instability, and suppression by major lipids like phosphatidylcholine (PC) in matrix-assisted laser desorption/ionization time-of-flight (MALDI-TOF) mass spectrometry. This dissertation introduces analytical strategies based on selective derivatization and digital filtering to improve detection. Two complementary methods were developed. The first method, Mass-Difference Digital Filter (MDDF), uses stearic acid hydrazide to derivatize aldehyde-containing oxidized phospholipids. It monitors paired signals formed between native lipids and their hydrazones. While MDDF works well under ideal conditions, it becomes less effective when isobaric lipids interfere or when the oxidized lipids are very low in abundance. To solve these problems, the Automated Bell-Curve Selectivity Algorithm (ABSA) was developed. It uses fixed-charge Girard’s reagents (GRT, GRP, and synthesized GRB) to derivatize oxidized lipids. ABSA identifies true hydrazones by detecting a bell-shaped signal pattern across reagent concentrations, where the signal increases during optimal derivatization and decreases due to ion suppression at higher reagent levels. This method successfully detected low-abundance oxidized lipids, even when background phospholipids were present in sixteen-fold excess. In addition, peak shape analysis was performed to improve mass accuracy and signal consistency. Twelve statistical models were tested for fitting MALDI-TOF lipid peaks. The Johnson SU distribution showed the best fit for asymmetric peaks, followed by the Extended Skew Normal and Exponentially Modified Gaussian distributions. Standard symmetric models like the normal distribution did not perform well, confirming that asymmetric models are more suitable for lipid peak shapes. Overall, combining chemical derivatization with digital pattern recognition offers a practical and effective solution for detecting oxidized lipids. These methods reduce the need for complex sample preparation or chromatography and make it possible to selectively identify and quantify oxidized lipid species in complex biological samples.