Cytokeratin profiles identify diagnostic signatures in colorectal cancer using multiplex analysis of tissue microarrays

Background and aims: Recent cDNA expression profiling analyses indicate that within specific organ cancers Cytokeratins (CKs) dysregulation may identify subgroups with distinct biological phenotypes. Our objectives in this study were (1) to test whether cytokeratins were also distinct on the protein level, (2) to evaluate these biomarkers in a series of well-characterised CRCs, (3) to apply hierarchical cluster analysis to immunohistochemical data. Methods: Tissue microarrays (TMA) comprising 468 CRC specimens from 203 patients were constructed to evaluate CK5, CK7, CK8, CK13, CK14, CK16, CK17, CK18, CK19 and CK20. In total, 2919 samples were analyzed. Results: Unsupervised hierarchical clustering discovered subgroups represented by reduced CK8 and CK20 expression, that differed by a shorter patients survival. The evaluation of the specific biomarkers by Kaplan-Meier analysis showed that reduced CK8 expression (p < 0.01) was significantly associated with shorter patients' survival, but was not an independent factor correlated with tumour stage (pT), grading (G) and nodal stage (pN). Conclusions: Reduced coexpression of CK8 and CK20 may indicate an epithelial-mesenchymal transition (EMT) representing an important step in the development of more aggressive CRCs. In addition, multiplex analysis of TMAs together with immunohistochemistry (IHC) supplemented by hierarchical clustering are a useful, promising and very powerful tool for the identification of tumour subgroups with diagnostic and prognostic signatures.