Lipidomic profiling of red blood cells in the mitochondrial fatty acid β-oxidation disorder MCADD reveals phospholipid and sphingolipid dysregulation

The RBC lipidome from MCADD patients was compared with those from control children (CTRL) using the advanced lipidomic approaches based on C18-liquid chromatography-high resolution mass spectrometry approach. The total lipid extracts from RBC samples were analyzed using reverse-phase liquid chromatography coupled with high-resolution tandem mass spectrometry (C18-LC-MS/MS). An Ultimate 3000 Dionex system (Thermo Fisher Scientific, Bremen, Germany) equipped with an Ascentis® Express C18 column (90 Å, 2.1 × 100 mm, 2.7 μm; Sigma-Aldrich®) was paired with a Q-Exactive® hybrid quadrupole Orbitrap mass spectrometer (Thermo Fisher Scientific, Bremen, Germany) for this purpose.A volume of 5 μL of each sample mixture was injected into the HPLC column. The separation was performed at 50 °C with a flow rate of 260 μL/min. The mobile phase A was Milli-Q water:acetonitrile (40:60%) with 10 mM ammonium formate and 0.1% formic acid, and mobile phase B was isopropanol:acetonitrile (90:10%) with the same additives. The gradient started at 32% B and increased to 45% B at 1.5 min, 52% B at 4 min, 58% B at 5 min, 66% B at 8 min, 70% B at 11 min, 85% B at 14 min, and 97% B at 18 min, which was maintained until 25 min. The system returned to 32% B at 25.01 min and underwent an 8-minute re-equilibration before the next injection. Mass spectrometry was performed in both positive (3.0 kV electrospray voltage) and negative (2.7 kV electrospray voltage) ionization modes. The capillary temperature was set to 320 °C, the probe temperature to 300 °C, and gas flows were set at 35 U and 3 U for sheath and auxiliary gases, respectively. The S-lens RF level was 50 U. Full-scan data were acquired at a resolution of 70,000, with an m/z range of 200–1600, an automatic gain control (AGC) target of 3 × 106, a maximum injection time (IT) of 100 ms, and 2 micro scans. Tandem mass spectrometry (MS/MS) spectra were obtained at a resolution of 17,500 with an AGC target of 1 × 105, 1 micro scan, and the same maximum IT. Data were acquired in a cycle of one full scan followed by 10 data-dependent MS/MS scans, with dynamic exclusion set to 30 seconds and an intensity threshold of 8 × 104. The normalized collision energy ranged from 20 to 28 eV in negative mode and from 25 to 30 eV in positive mode. Data acquisition was performed using the Xcalibur data system (V3.3, Thermo Fisher 6 Scientific, Bremen, Germany). The LC-MS data were identified using Lipostar2 software (Molecular Discovery Ltd., version 2.1.5). The software was used to process raw data, detect peaks, align chemical features, identify lipid species, and measure peak areas. Lipid identification was carried out in both positive and negative ionization modes by comparing raw MS/MS data files to a refence spectral library generated with the DB Manager module of Lipostar by importing the LIPID MAPS structure database (accessed on May 14, 2024) that was further fragmented accordingly to the fragmentation rules encoded in the Lipostar software. The raw data were imported and aligned with the following parameters: MS signal filtering thresholds of 10,000 for positive mode and 5,000 for negative mode; MS/MS signal filtering threshold of 100. Peaks were smoothed using the Savitzky-Golay method (window size: 7, degree: 2 and multi-pass iterations: 1), and detected with an m/z tolerance of 5 ppm. Isotope clustering was performed with a 10 ppm tolerance and a retention time (RT) tolerance of 0.2 minutes. Sample alignment settings included a 10 ppm tolerance and an RT tolerance of 0.5 minutes. The MS/MS filter was employed to selectively preserve features with MS/MS spectra for identification, with precursor ion and product ion mass tolerances set to 5 ppm and 10 ppm, respectively. Lipid identifications were rated with a quality score of 3 to 4 stars. Finally, peak areas from the extracted ion chromatograms (XIC) were exported, and normalization was performed using the area of a selected lipid internal standard.Multivariate and univariate statistical analyses were performed using R version 4.3.1 in RStudio version 2024.04.2. The areas of lipid species (excluding neutral lipids such as TG, DG, and CE) were first normalized to selected lipid internal standards, scaled by the total area sum then log₁₀-transformed and subsequently normalized using EigenMS. Lipid class ratio data were similarly normalized to the total area sum, log₁₀-transformed, and adjusted using EigenMS. The R packages FactoMineR and factoextra were used to perform principal component analysis (PCA). Heatmaps were created from autoscaled data using the R package pheatmap , using “Euclidean” as the clustering distance, and “ward.D” as the clustering method. Data normality and variance homogeneity were assessed using Shapiro-Wilk and Levene’s tests, respectively. Based on these tests, data were analyzed using Welch t-test if assumptions were met, or the Mann–Whitney test if not. P-values were corrected for multiple testing using the Benjamin−Hochberg method for the false discovery rate (FDR, q-values). All univariate analyses were performed using the r package rstatix with a significance threshold of p < 0.05. All graphics were created using the R package ggplot2.