Critical Evaluation of Sphingolipids Detection by MALDI-MSI

Abstract

The increasing interest in the role of sphingolipids in (patho)physiology has led to the demand for visualization of these lipids within tissue samples (both from animal models and patient specimens) using techniques such as matrix-assisted laser desorption/ionization mass spectrometry imaging (MALDI-MSI). While increasingly adopted, detection of sphingolipids with MALDI-MSI is challenging due to: i) the significant structural variations of sphingolipid molecules, ii) the potential breakdown of the more complex molecules into structurally simpler species which may confound the analysis, and iii) the great difference in levels among sphingolipid classes and subspecies, with the low-abundant ones often being close to the detection limit. In this study, we adopted a multi-pronged approach to establish a robust pipeline for the detection of sphingolipids by MALDI-MSI and to establish best practices and limitations of this technology. First, we evaluated the more commonly adopted methods [2,5-Dihydroxyacetophenon (DHA) or 2,5-Dihydroxybenzoic acid (DHB) matrix in positive ion mode and 1,5-Diaminonaphthalene (DAN) matrix in negative ion mode] using MALDI-MS on reference standards. These standards were used at ratios similar to their relative levels in biological samples to evaluate signal artifacts originating from fragmentation of more complex sphingolipids and impacting low level species. Next, by applying the most appropriate protocol for each sphingolipid class, MALDI-MSI signals were validated in cell culture by modulating specific sphingolipid species using sphingolipid enzymes and inhibitors. Finally, the optimized parameters were utilized on breast cancer tissue from the PyMT mouse model. We report the optimal signal for sphingomyelin (SM) and, for the first time, Sph in DHB positive ion mode (in cells and PyMT tissue), and the validated detection of ceramides and glycosphingolipids in DAN negative ion mode. We document the extensive fragmentation of SM into sphingosine-1-phosphate (S1P) and even more so into ceramide-1-phosphate (C1P) using DAN in negative ion mode and its effect in generating an artifactual C1P tissue signal; we also report the lack of detectable signal for S1P and C1P in biological samples (cells and tissue) using the more suitable DHB positive ion mode protocol.