Large-scale clinical platelet lipidomics in coronary artery disease patients by untargeted data independent liquid chromatography hyphenated with high-resolution mass spectrometry

Abstract

Cardiovascular diseases (CVD) are one of the main causes of morbidity and mortality worldwide and are highly prevalent in modern societies. Thus, research in this field is expedited and here, lipidomics emerged to a valuable tool in assessing the physiological and pathophysiological lipid profile in human blood. Lipids, being essential metabolites e.g. for cellular structure or inter- and intracellular signaling, are critically involved in disease development, progression, and prognosis. While CVD research is typically based human serum or plasma, the lipid profile of platelets, as the cellular component in haemostasis and thrombosis, is highly interesting but rarely addressed. In this thesis, large-scale lipidomics was conducted on platelets from clinical cohorts of patients suffering coronary artery disease, to explore the lipid profile in dependence on various clinical conditions. The platelet lipids from two cohorts were investigated by reversed phase ultra-high-performance liquid chromatography (RP UHPLC) coupled with high resolution mass spectrometry on a hybrid quadrupole-time-of-flight (QTOF) instrument in an untargeted data independent acquisition approach using SWATH for MS and MS/MS data collection. Due to the complexity of lipid profiling data, i.e. data from thousands of metabolites in hundreds of samples acquired in multiple batches, the challenge was the development and establishment of a data processing workflow for large-scale lipidomics. A combined untargeted-targeted data extraction approach was proposed with i. batch-wise untargeted data extraction of selected batches, ii. reference target list preparation by inter-batch feature alignment of untargeted processed batches, and iii. batch-wise targeted data extraction of the entire cohort based on the reference target list. For inter-batch feature alignment, a VBA-based tool was developed to merge metabolite data from multiple batches based on accurate mass and retention time similarity. Further, several steps of curating annotated lipids as well as unknown features were established to improve the quality of the lipidomics datamatrices. By this strategy, representative lipid profiling data of platelets from the entire cohorts were obtained. Due to the untargeted data structure, the clinical platelet lipidomics data were used for hypothesis generation on different clinical conditions or for cardiovascular risk assessment. In one exploration, the platelet lipidome was found to be significantly altered in dependence on disease severity, most prominently with an upregulation of medium chain fatty acyl phosphatidylcholines. The upregulation might lead to a hyperreactive state of the platelets under more severe disease conditions, potentially originating from a fatty acid oxidation disorder. Secondly, the effect of lipid-lowering therapy with statins was investigated, indicating a pleiotropic effect by beneficially modulating the platelet lipid profile. Furthermore, the association of the platelet lipid profile with the risk of adverse cardiovascular events was shown, emphasizing the potential of platelet lipids in cardiovascular risk assessment.

Die Dissertation ist gesperrt bis zum 18. Juli 2025 !