High sensitivity proteomics of prostate cancer tissue microarrays to discriminate between healthy and cancerous tissue.

Biopsies, in the form of tissue microarrays (TMAs) were studied to identify anomalies indicative of prostate cancer at the proteome level. TMAs offer a valuable source of well-characterized biological material. However, because of the small tissue sample size method development was essential to provide the sensitivity and reliability necessary for the analysis. Surface digestion of TMA cores was followed by peptide extraction and shotgun proteomics analysis. About 5 times better sensitivity was achieved by the optimized surface digestion compared to bulk digestion of the same TMA spot and it allowed the identification of over 500 proteins from individual prostate TMA cores. Label-free quantitation showed that biological variability among all samples was about 3 times larger than the technical reproducibility. We have identified 189 proteins which showed statistically significant changes (t-test p-value <.05) in abundance between healthy and cancerous tissue samples. The proteomic profile changed according to cancer grade, but did not show a correlation with cancer stage. Results of this pilot study were further evaluated using bioinformatics tools, identifying various protein pathways affected by prostate cancer progression indicating the usefulness of studying TMA cores to identify quantitative changes in tissue proteomics. SIGNIFICANCE: Detailed proteomics analysis of TMAs presents a good alternative for tissue analysis. Here we present a novel method, based on tissue surface digestion and nano-LC-MS measurements, which is capable of identifying and quantifying over 500 proteins from a 1.5 mm diameter tissue section. We compared healthy and cancerous prostate tissue samples, and tissues with various grades and stages of cancer. Tissue proteomics clearly distinguished healthy and cancerous samples, furthermore the results correlated well with cancer grade, but not with cancer stage. Over 100 proteins showed statistically significant abundance changes (t-test p-value <.05) between various groups. This was sufficient for a meaningful bioinformatics evaluation; showing e.g. increased abundance of proteins in cancer in the KEGG ribosome pathway, GO mRNA splicing via spliceosome, and chromatin assembly biological processes. The results highlight the feasibility of the developed method for future large-scale tissue proteomics studies using commercially available TMAs.