Mass spectrometry Imaging (MSI) by MALDI (Matrix Assisted Laser Desorption-Ionisation) allows to study biological tissues without previous extraction, purification and separation of the biomolecules. This in situ analysis preserves essential information concerning the location of the compounds within their tissues of origin. By direct analysis of tissue sections with a defined scan of an incident laser beam, and the recording of spectra at different points with their recording coordinates on the section, it is then possible, using signal processing software, to reconstitute maps of the molecular distribution of the compounds within the section. This technique has the advantage of delivering as many molecular images as there are ions present on the spectrum and therefore of detecting many molecular species simultaneously. These are called ion density maps.

©PIXANIM@INRAE

Using MSI, it is possible to track lipid or even peptide/protein abundance profiles directly within tissues and to overlay ion density maps with microscopic observation of the tissue. By superimposing a few ion density maps it is possible to quickly recognise major tissue structures (e.g. molecular images of frontal slices of whole sheep brains).
Using computer processing based on hierarchical clustering, it is then possible to create a segmentation map that allows for a much finer molecular histology than can be obtained with conventional H & E staining (e.g. segmentation maps of frontal and sagittal sections of whole sheep brains).

We talk about 2D imaging for the visualisation of species within a tissue section. Imaging sequences can be performed on large tissue sections (e.g. 5 cm X 7.5 cm) with different levels of spatial resolution.
Opposite: segmentation maps of ovary slices of different species (sow, ewe, cow) at 50 µm spatial resolution.
For regions of interest, the spatial resolution can reach up to 10-20 µm (e.g.: pig follicle allows the characterisation of different follicular cells including the oocyte (single cell of 100µm diameter).Imaging sequences can be applied to serial sections to switch from 2D to 3D imaging. Thanks to advances in bioinformatics in image reconstruction, it is now possible to envisage 3D images with the analysis and alignment of serial sections of a complete organ.
Below is an example of the qualitative analysis of a sheep ovary to create a 3D molecular atlas. This then enables the distribution of species to be visualised or the molecular structure of a complete organ to be observed.

©PIXANIM@INRAE

In addition, the PIXANIM platform is engaged in methodological developments concerning multimodal correlative in and ex vivo imaging by combining Magnetic Resonance Imaging (MRI) and Mass Spectrometry Imaging (MSI) in order to correlate anatomical and molecular information for a better understanding of the functioning mechanisms of organs. 

©PIXANIM@INRAE

The molecular image (e.g. below: segmentation map of a sagittal section of a sheep brain) can be co-recorded with the scan (optical imaging) thanks to an image alignment requiring linear algorithms. However, the alignment of molecular images with MRI images obtained in vivo requires elastic algorithms. This is due to organ deformation and tissue section processing.
Nevertheless, these bioinformatics treatments have allowed us to propose a 2.5 D representation of the sheep brain which combines three imaging modalities: in vivo 3D MRI, ex vivo 2D MSI of a whole sagittal brain slice and histological optical imaging.

©PIXANIM@INRAE

Our first work on the 2.5 D representation of the sheep brain was presented at the SMAP 2015 congress in Ajaccio (Mass Spectrometry and Proteomic Analysis).