Abstract:To find new biomarkers and establish histopathology protein fingerprint models for early detection of colorectal cancer (CRC), laser capture microdissection (LCM) was utilized to obtain 3 groups of cells of interest--CRC tissues, their adjacent normal colorectal tissues, and their adjacent adenomatous polyps tissues--from the same 12 CRC patients. Each sample was then detected by surface-enhanced laser desorption ionization time-of-flight mass spectrometry (SELDI-TOF-MS) technology and CM10 protein chip as well as bioinformatics tools. Model 1 formed by 15 protein peaks could be used to distinguish CRC tissues from normal tissues. The diagnostic pattern constructed using support vector machine (SVM) including the 15 proteins showed maximum Youden index (YI). Model 2 formed by 14 protein peaks could be used to distinguish CRC tissues from adenomatous polyps tissues. The two patterns were validated and the results showed that the sensitivity and specificity were both 100.0%. Model 3 formed by 15 protein peaks could distinguish adenomatous polyps tissues from normal tissues with a sensitivity of 92.3% and specificity of 100.0%. The protein peaks m/z 3570 and 5224 were identified in screening for changes during cancer progression. Peak 5224 was significantly upregulated in CRC. However, peak 3570 was significantly downregulated in CRC. LCM technology coupled with SELDI protein chip and bioinformatics approaches could effectively screen the differentially expressed protein profiles and establish molecular diagnosis models with high sensitivity and specificity for CRC.

Protein Fingerprint of Colorectal Cancer, Adenomatous Polyps, and Normal Mucosa Using ProteinChip Analysis on Laser Capture Microdissected Cells.