Human primary lung cancer samples, patient derived lung cancer xenografts and lung tumors from the TetO-KRASG12D mouse model were analyzed using MALDI-MS. We determined the spatial distribution and relative abundance of lipids of interest. 

10-µm-thick tissue sections were mounted on Superfrost Microscope Slides (Fisherbrand) and stored at -80°C. They were processed by the Chemical Imaging Research Core, UT MD Anderson Cancer Center. A HTX Sprayer M5 matrix applicator (HTX Technologies, LLC., Chapel Hill, NC, USA) was used to apply 2,5-Dihydroxybenzoic acid (DHB) for positive mode, at a flow rate of 100 μL/min with temperature set to 75/30 °C for sprayer/tray. DHB was dissolved to a concentration of 15 mg/mL in 50% acetonitrile with 0.1% TFA. Before being loaded into the mass spectrometer, the slides were placed in a MALDI plate, scanned using an EPSON scanner (Epson, Suwa, Japan), and the sections of tissue were mapped into High Definition Imaging software (HDI 1.4; Waters, Milford, MA, USA). MALDI-MS imaging was performed using a Water Synapt G2 Si (Waters Corporation, Milford, MA). Data were acquired with a spot of 60 µm with 300 laser shots at 1 kHz, using a pulse energy with an average of 25 µJ. The laser intensity was adjusted to 60%. The mass range was 50-2000 m/z and the instrument was calibrated using peak signals from red phosphorus. We converted all files into imzML using the Waters High Definition Imaging (HDI) software. We performed all data image visualization and data analysis using msIQuant. Within msIQuant, Peak option signal to noise ratio (SNR) was set at 3.0 and peak Group Detection within 0.5 Da as for marker selection, Marker Mass Range was ±0.05 Da and the Maximum Intensity in range was used as intensity method. Lipid identification was performed comparing observed peaks (experimental mass) with the theoretical values reported in the Lipid MAPS (http://www.lipidmaps.org/) and Madison Metabolomics Consortium (http:mmcd.nmrfam.wisc.edu/) databases (theoretical mass), using the mass accuracy (0.5 Da) as a tolerance window. In this way, we identified a single candidate for each peak in the spectrum. Using experimental and theoretical values, mass error (ppm) was calculated.

For each sample we provide two paired data files, one with .ibd extension and one with .imzml extension that were generated by the Water Synapt G2 software (Waters Corporation, Milford, MA). The files can be visualized using the free, publicly available software msiQuant available at https://ms-imaging.org/paquan/ from DOI: 10.1021/acs.analchem.5b04603 or Datacube Explorer.

In both the software, the .imzml files can be uploaded, but both the ibd and imzml files must be in the same directory for the software to work.