Computational protocol: Altered metabolic landscape in IDH‐mutant gliomas affects phospholipid, energy, and oxidative stress pathways

Similar protocols

Protocol publication

[…] Ten‐micrometer‐thick cryosections of glioma xenografts were prepared on a cryostat (Microme HM560), set at −21°C for specimen and blade parameters, and mounted on conductive slides (Indium Tin Oxide, ITO) for MSI and cryodesiccated for 30 min. Slides were then transferred into a desiccator under vacuum for 30 min to complete drying and scanned for next step (teaching with imaging software). Matrix deposit for MSI : 9‐AA (9‐Amino Acridine) matrix was prepared at 5 mg/ml in 100% MeOH (LC‐MS grade) and sprayed onto the slide with the Suncollect sprayer (Sunchrom, GmbH Germany). Analyses were performed on a Matrix Assisted Laser Desorption Ionisation–Fourier Transform Ion Cyclotron Resonance instrument (MALDI‐FTICR) 7.0 T (Bruker Daltonics GmbH, Bremen, Germany) in full scan negative mode within 70–1,000 Da mass range at 100‐μm spatial resolution. The horizontal axis in a mass spectrum is expressed in units of m/z, which represents the mass over charge (number of ions) ratio. For the untargeted analysis, 50 pixels of analysis from the imaging data (i.e., 50 mass spectra) per region of interest (ROI) were selected to run a statistical test on ClinProTools 3.0 software (Bruker, Germany) () using a threshold of signal to noise of 5 to compare the peaks between conditions. Recalibration, average peak list calculation, and peak calculation from ClinProTools allowed obtaining statistical results with metabolites using t‐test. Relevant metabolites were considered for P value < 0.01 (from t‐test) and fold change > 3 between conditions ( and ). MSI data were operator‐verified to check the imaging data and suppress false positives. Imabiotech metabolite database as well as Metlin and the Human Metabolome DataBase (HMDB) were interrogated for identification of the compounds found with ClinProTools. Mass accuracy was set at 10 ppm, and lists of metabolites were established with their structures. When several hits matched with the m/z of interest, one representative compound was selected for presentation in and . For the targeted approach, measurement regions included identical large ROIs selected in biological sections and were analyzed by Fleximaging 4.0 software (Bruker Daltonics GmbH, Germany). Initially for each condition (PDX or control brain without implantation), three sections from three animals were analyzed for quantification. In a second step, sections were analyzed from six different PDX models. To ensure the correct assignment of specific metabolites, we performed fragmentation analysis to confirm their identity and were able to validate nine of 16 compounds including glutamate (Glu), glutamine (Gln), N‐acetyl aspartate (NAA), N‐acetyl aspartyl glutamate (NAAG), cystathionine, reduced glutathione (GSH), ascorbic acid, citric acid, and cytidine (). Metabolites were validated if at least one fragment was found to be relevant.LESA‐nESI‐FTICR was applied for specific metabolites that were undetectable by MALDI as described previously (Navis et al, ). A Nanomate Triversa (Advion) in “Liquid Extraction Surface Analysis” (LESA) mode coupled to a nano‐ElectroSpray Ionisation (nano‐ESI) source on the FTICR was used to extract analytes from biological sections with an extraction solution. Sample plates were cooled down to 12°C during analyses to limit degradation or secondary reactions. Fragmentation was performed in positive and negative mode within the collision cell of the SolariX after isolation of the parent compound in the quadrupole to confirm the identity of the metabolites. Spray parameters were set as follows: voltage to apply 1.30 kV and gas pressure 0.40 psi. Extraction solvent consisted of 65:15:20 MeOH:IPA: water + 5 mM ammonium acetate using LC‐MS quality solvents. Isolation of the ion in the quadrupole within a 2–5 mDa window and fragmentation with appropriate collision energies in the collision cell (qCID) (±15 V). After MALDI analysis, the 9‐AA matrix was removed from the slide with MeOH and hematoxylin–eosin (H&E) staining was performed to visualize the tumor area. This image was later implemented into Quantinetix™ software for data analysis by overlaying the optical image with molecular images. An intensity‐dependent color code shows the relative amount of a specific compound (m/z value) throughout the tissue section, which was also used for quantification (log intensity values). [...] The signal obtained by MALDI on tissue sections can be modulated by the chemical composition of the tissue and thus can differ from one tissue to another. This can be controlled for by determining the tissue extinction coefficient (TEC) for each metabolite. Since the metabolites are endogenous molecules in the brain, labeled forms had to be used in this approach. Therefore, for the TEC determination, isotopically labeled compounds were mixed with the 9‐AA matrix (5 mg/ml in MeOH) to achieve 20 μM final concentration, and the final mixture was sprayed on the entire slide for analysis. For TEC analysis, an additional region was selected outside tissue sections as a control area (no tissue effect in this area) for calculations of TEC values, defined as the ratio of compound signal in the tissue section to the compound signal outside the tissue section. The following compounds were analyzed: NAA‐d3, NAAG‐d3 (Spaglumic acid‐d3), L‐cystathionine‐d4, GSH 13C, 15N and GSSG 13C,15N. NAA‐d3 was purchased from CDN isotopes, all other itotopically labeled compounds from Toronto Research Chemical, Canada. LC‐MS grade solvents were used for sample preparation. Overall, the TEC values were comparable for the different regions (mutant or wild‐type tumor, contralateral brain, and control brain), indicating that for the majority of compounds, the quantification is not likely to be affected by strong tissue effects (). Except for GSSG, which was excluded from the present data, since all isotopic compounds were simultaneously present in the mix, oxidation reactions cannot be excluded, which may explain differences in GSSG values. Data were treated with Fleximaging 4.0 (Bruker, Germany), proprietary softwares Quantinetix™ 1.7, and Multimaging™ 1.0.30 (ImaBiotech, Loos, France). […]

Pipeline specifications

Software tools flexImaging, Quantinetix, Multimaging
Application Mass spectrometry imaging
Organisms Homo sapiens, Homo sapiens/Mus musculus xenograft
Diseases Glioma, Neoplasms, Glucose Intolerance
Chemicals Carbon, Cystathionine, Glucose, Glutathione, NADP