The maturation and automation of histo-chemical techniques, in which antibodies are used to investigate the expression and distribution of specific protein biomarkers across many patient biopsies, has led to the establishment of databases that contain the distribution of biomarker proteins for many pathologies. The analysis of patient biopsies with antibodies for validated biomarkers are now established tools for patient diagnosis and prognosis, and remains one of the key technologies for validating new candidate biomarkers.

Mass spectrometry based analysis of proteins now allows the quantitative analysis of the protein content of tissue and body fluids, simultaneously determining the levels of thousands of proteins, including protein isoforms. Many studies have now established how changes in the total protein profile can be more effective ‘biomarkers’ because it provides a more complete representation of the biochemical system. Clinical tissue analyses would benefit from such a parallel analysis of multiple proteins, distinguishing protein isoforms and relative protein stoichiometry.

We have shown how mass-spectrometry based methods can be directly applied to tissue to measure its molecular composition and protein distributions. Imaging mass spectrometry has rapidly progressed, current technologies allow the parallel analysis of 100’s of proteins with high sensitivity and selectivity. A low resolution (250 m) imaging mass spectrometry analysis of a small tissue microarray has demonstrated how the protein signatures from each tissue biopsy can be used to classify their pathological state. In this case the ‘biomarkers’ are based on the interplay of multiple proteins.

In this work package we will develop new tools that for imaging mass spectrometry analysis of clinical tissues to address some of the cutting-edge topics in modern cancer research, namely the interaction between tumor cells and neighboring non-tumor cells (tumor interface zones). Grid based computing will enable the large datasets from these analyses (30-100 Gb) to be integrated with immunohistochemical analyses of adjacent sections, and the resulting multiplex imaging data classified according to its morphological-biochemical state. New data analysis and visualization tools will be developed that allow the multiplex imaging data to be interrogated by all members of the collaboration, thus enabling the expertise of each member of the consortium to be fully exploited. All metadata and processing steps will be recorded in the workflow based knowledge management system created in P24.2 and model datasets and the results of the analyses uploaded into the Netherlands Virtual Tissue Bank.

These tools will be used to investigate if changes in the expression patterns of large panels of proteins can provide improved diagnostic/prognostic capabilities. The data analysis capabilities provided through VLe++ are crucial to the development of multiplex imaging of clinical patient tissues, and to establish if a system- wide approach can provide diagnostic/prognostic tools for currently difficult-to-diagnose/distinguish cancers.