Formalin fixation, generally followed by paraffin embedding, is the standard and well-established processing method employed by pathologist. This treatment conserves and stabilizes biopsy samples for years. Analysis of FFPE tissues from biopsy libraries has been, so far, a challenge for proteomics biomarker studies. Herein, we present two methods for the direct analysis of formalin-fixed, paraffinembedded (FFPE) tissues by MALDI-MS. The first is based on the use of a reactive matrix, 2,4-dinitrophenylhydrazine, useful for FFPE tissues stored less than 1 year. The second approach is applicable for all FFPE tissues regardless of conservation time. The strategy is based on in situ enzymatic digestion of the tissue section after paraffin removal. In situ digestion can be performed on a specific area of the tissue as well as on a very small area (microdigestion). Combining automated microdigestion of a predefined tissue array with either in situ extraction prior to classical nanoLC/MS-MS analysis or automated microspotting of MALDI matrix according to the same array allows the identification of both proteins by nanoLC-nanoESI and MALDI imaging. When adjacent tissue sections are used, it is, thus, possible to correlate protein identification and molecular imaging. These combined approaches, along with FFPE tissue analysis provide access to massive amounts of archived samples in the clinical pathology setting.

Direct tissue analysis using MALDI-MS allows the generation of profiles while maintaining the integrity of the tissue, displaying cellular localizations and avoiding tedious extraction and purification steps. However, lower spectral quality can result from direct tissue analysis due to variations in section thickness, the nature of the tissue, and the limited access to peptides/proteins due to high lipid content. To improve signal sensitivity, we have developed a tissue-washing procedure using organic solvents traditionally used for lipid extraction, i.e., CHCl3, hexane, toluene, acetone, and xylene. The increased detection for peptides/proteins (m/z 5000-30,000) is close to 40% with chloroform or xylene, and 25% with hexane, while also improving sample reproducibility for each solvent used in the present study. This strategy improved matrix cocrystallization with tissue peptides/proteins and more importantly with cytoplasmic proteins without delocalization. The extracted lipids were characterized by nanoESI-QqTOF/MS/MS using the precursor ion mode, lithium adducts, or both and were identified as phospholipids including phosphatidylcholine, phosphatidylethanolamine, phosphatidylinositol, and lysophosphatidylinositol, confirming membrane lipid extraction from the tissues.