Samenvatting

Over the last years, Mass Spectrometry Imaging (MSI) has proven to be a useful tool for direct in-situ analysis of thin tissue sections. This label free technique was quickly recognized and applied in different fields by demonstrating its advantages over other imaging techniques such as autoradiography, immunohistochemistry or positron emission tomography [1]. The MSI is capable of providing simultaneous distribution and identification of hundreds of different (unknown) compounds in a single imaging experiment [2-4].
In every successful MSI measurement, the sample preparation plays an essential part of the analysis. Therefore, it is necessary to optimize and adapt the procedures according to the tissue and analyte under investigation [5]. In this case, the protocol was developed for the analysis of rodent brain tissue and proteins/peptides within the tissue.
Different washing procedures, enzyme application techniques, incubation times, protease solutions, matrices and matrix application techniques were tested to define the optimal protocol. The chloroform/ethanol washing procedure was found to be the best in lowering the ion suppression effect. The optimal digestion was achieved with trypsin dissolved in ddH2O, applied with SunCollect and incubated for 18 hrs. Good extraction/ionization of peptides during the MALDI-MSI experiments was achieved with α-CHCA matrix applied with SunCollect. All acquired spectra were baseline corrected, normalized by TIC, autoscaled and peak picked. Furthermore, Principal Component Analysis (PCA) and Discriminant Analysis (DA) were applied to uncover and visualize the variability within the data.
The developed protocol for on-tissue digestion demonstrated that it is feasible to localize and identify peptides/protein directly from brain tissue. More than 30 proteins could be identified from the wild type (WT) tissue. The optimized protocol was also used to study distribution of peptides/proteins within the Alzheimer mouse model (APP/PS1) brain tissue, showing subtle differences between the two models.