Characterizing metabolic changes in human colorectal cancer

Colorectal cancer (CRC) remains a leading cause of cancer death worldwide, despite the fact that it is a curable disease when diagnosed early. The development of new screening methods to aid in early diagnosis or identify precursor lesions at risk for progressing to CRC will be vital to improving the survival rate of individuals predisposed to CRC. Metabolomics is an advancing area that has recently seen numerous applications to the field of cancer research. Altered metabolism has been studied for many years as a means to understand and characterize cancer. However, further work is required to establish standard procedures and improve our ability to identify distinct metabolomic profiles that can be used to diagnose CRC or predict disease progression. The present study demonstrates the use of direct infusion traveling wave ion mobility mass spectrometry to distinguish metabolic profiles from CRC samples and matched non-neoplastic epithelium as well as metastatic and primary tumors at different stages of disease (T1-T4). By directly infusing our samples, the analysis time was reduced significantly, thus increasing the speed and efficiency of this method compared to traditional metabolomics platforms. Partial least squares discriminant analysis was used to visualize differences between the metabolic profiles of sample types and to identify the specific m/z features that led to this differentiation. Identification of the distinct m/z features was made using the human metabolome database. We discovered alterations in fatty acid biosynthesis and oxidative, glycolytic, and polyamine pathways that distinguish tumors from non-malignant colonic epithelium as well as various stages of CRC. Although further studies are needed, our results indicate that colonic epithelial cells undergo metabolic reprogramming during their evolution to CRC, and the distinct metabolites could serve as diagnostic tools or potential targets in therapy or primary prevention. Graphical Abstract Colon tissue biopsy samples were collected from patients after which metabolites were extracted via sonication. Two-dimensional data were collected via IMS in tandem with MS (IMMS). Data were then interpreted statistically via PLS-DA. Scores plots provided a visualization of statistical separation and groupings of sample types. Loading plots allowed identification of influential ion features. Lists of these features were exported and analyzed for specific differences. Direct comparisons of the ion features led to the identification and comparative analyses of candidate biomarkers. These differences were then expressed visually in charts and tables.