Molecular imaging of human brain tumour tissue by mass spectrometry

Introduction: Mass spectrometry imaging (MSI) has become a useful tool for investigating the distributions of molecules within tissue sections, particularly lipids and small molecules. However, the analysis of proteins and peptides by MSI is more challenging due to difficulties with dynamic range, ionisation efficiency and spectral complexity that impede detection, quantification and identification. Typical liquid chromatography mass spectrometry (LC-MS) based proteomics experiments lack spatial resolution and give little to no knowledge of the underlying tissue structure. Here we apply a sensitive, spatially resolved workflow for the proteomic analysis of a tumour and identify proteins displaying spatial expression patterns in an unsupervised manner and integrate this data with MSI of lipids and small molecules.

Methodology: We used laser capture microdissection (LCM) to isolate tissue voxels from a human brain tumour obtained post-mortem in a raster pattern. Captured voxels were proteolytically digested with trypsin and peptides analysed by LC-MS. Quantified proteins were tested for spatial expression patterns using a measure of spatial autocorrelation, Moran’s I. Adjacent sections were analysed by MALDI MSI for lipids, small molecules and peptides (after tryptic digestion).

Results: We identified approximately 4000 proteins from the LCM-LC-MS data. The regions sampled for proteomics included the tumour solid core, the brain/tumour interface and blood vessels at ~300 µm resolution. Approximately 500 proteins display significant evidence by Moran’s I for differential spatial expression. These proteins include examples of blood proteins, tumour and neuronal markers and show enrichment of Gene Ontology terms related to tumour biology. Unsupervised segmentation of MALDI MSI data identified histopathological features and several lipids & small molecules colocalise with the tumour region.

Conclusion: LC-MS based mass spectrometry imaging of proteins has the potential to provide novel insights into the proteome of tissues.