Investigation of the HDL Molecular Architecture Before and After Exercise Training.

The functional characteristics of high-density lipoprotein (HDL) particles are influenced by their structure, including their molecular composition (e.g. HDL lipidome and proteome). However, the effects of exercise training on HDL molecular composition and its relationship with HDL-related traits are largely unknown. In this ancillary study, we investigated the HDL lipidome and proteome of 154 adults before and after 20 weeks of exercise training in the HERITAGE Family Study. The HDL-sized plasma fraction was isolated utilizing FPLC-SEC, followed by untargeted lipidomic and proteomic analysis using LC/MS. We identified 341 known lipid species while 409 HDL associated proteins. Clinical and functional related HDL traits were measured at baseline and post-training. Exercise response of the HDL lipidome/proteome and its associations with HDL-related traits were examined, with statistical significance set to FDR< 0.05. We employed sparse canonical correlation analysis to investigate the potential multivariate relationships between HDL-associated lipids and proteins with various HDL-related traits. Exercise training did not significantly alter the abundance of any HDL lipid class or HDL proteins. However, HDL phosphatidylethanolamine (PE) significantly increased in females only, while HDL diglycerides (DG) nominally (p< 0.05) decreased in males only. Moreover, the abundance of 16 HDL lipid species at baseline and 33 HDL lipid species at post-training were significantly different between males and females. Association analysis revealed higher abundance of HDL surface lipids was associated with larger size and cholesterol content of HDL particles before and in response to exercise training. Most associations of HDL-related traits with the HDL lipidome were similar in strength and direction across sex groups before and in response to exercise training. Canonical correlation analysis identified distinct sets of correlated HDL lipids and proteins, whose unique associations with HDL-related traits suggest biologically meaningful interactions. However, the interaction of HDL lipids and proteins, along with their associations with HDL-related traits, are observational in nature and require further investigation. In conclusion, our analysis indicates that exercise training may have a limited impact on HDL lipidome and proteome. We found sex differences in the abundance of HDL lipids. The investigation of potential mechanisms for these sex differences and the interaction of HDL lipidome and proteome in a larger cohort and/or experimental models are needed to further explore our preliminary findings.