Colorectal carcinogenesis involves the overexpression of many immediate-early response genes associated with growth and inflammation, which significantly alters downstream protein synthesis and small-molecule metabolite production. We have performed a serum metabolic analysis to test the hypothesis that the distinct metabolite profiles of malignant tumors are reflected in biofluids. In this study, we have analyzed the serum metabolites from 64 colorectal cancer (CRC) patients and 65 healthy controls using gas chromatography time-of-flight mass spectrometry (GC-TOFMS) and Acquity ultraperformance liquid chromatography-quadrupole time-of-flight mass spectrometry (Acquity UPLC-QTOFMS). Orthogonal partial least-squares discriminate analysis (OPLS-DA) models generated from GC-TOFMS and UPLC-QTOFMS metabolic profile data showed robust discrimination from CRC patients and healthy controls. A total of 33 differential metabolites were identified using these two analytical platforms, five of which were detected in both instruments. These metabolites potentially reveal perturbation of glycolysis, arginine and proline metabolism, fatty acid metabolism and oleamide metabolism, associated with CRC morbidity. These results suggest that serum metabolic profiling has great potential in detecting CRC and helping to understand its underlying mechanisms.

Missing values exist widely in mass-spectrometry (MS) based metabolomics data. Various methods have been applied for handling missing values, but the selection can significantly affect following data analyses. Typically, there are three types of missing values, missing not at random (MNAR), missing at random (MAR), and missing completely at random (MCAR). Our study comprehensively compared eight imputation methods (zero, half minimum (HM), mean, median, random forest (RF), singular value decomposition (SVD), k-nearest neighbors (kNN), and quantile regression imputation of left-censored data (QRILC)) for different types of missing values using four metabolomics datasets. Normalized root mean squared error (NRMSE) and NRMSE-based sum of ranks (SOR) were applied to evaluate imputation accuracy. Principal component analysis (PCA)/partial least squares (PLS)-Procrustes analysis were used to evaluate the overall sample distribution. Student’s t-test followed by correlation analysis was conducted to evaluate the effects on univariate statistics. Our findings demonstrated that RF performed the best for MCAR/MAR and QRILC was the favored one for left-censored MNAR. Finally, we proposed a comprehensive strategy and developed a public-accessible web-tool for the application of missing value imputation in metabolomics (https://metabolomics.cc.hawaii.edu/software/MetImp/).

After our serum metabonomic study of colorectal cancer (CRC) patients recently published in J. Proteome Res., we profiled urine metabolites from the same group of CRC patients (before and after surgical operation) and 63 age-matched healthy volunteers using gas chromatography-mass spectrometry (GC-MS) in conjunction with a multivariate statistics technique. A parallel metabonomic study on a 1,2-dimethylhydrazine (DMH)-treated Sprague-Dawley rat model was also performed to identify significantly altered metabolites associated with chemically induced precancerous colorectal lesion. The orthogonal partial least-squares-discriminant analysis (OPLS-DA) models of metabonomic results demonstrated good separations between CRC patients or DMH-induced model rats and their healthy counterparts. The significantly increased tryptophan metabolism, and disturbed tricarboxylic acid (TCA) cycle and the gut microflora metabolism were observed in both the CRC patients and the rat model. The urinary metabolite profile of postoperative CRC subjects altered significantly from that of the preoperative stage. The significantly down-regulated gut microflora metabolism and TCA cycle were observed in postoperative CRC subjects, presumably due to the colon flush involved in the surgical procedure and weakened physical conditions of the patients. The expression of 5-hydroxytryptophan significantly decreased in postsurgery samples, suggesting a recovered tryptophan metabolism toward healthy state. Abnormal histamine metabolism and glutamate metabolism were found only in the urine samples of CRC patients, and the abnormal polyamine metabolism was found only in the rat urine. This study assessed the important metabonomic variations in urine associated with CRC and, therefore, provided baseline information complementary to serum/plasma and tissue metabonomics for the complete elucidation of the underlying metabolic mechanisms of CRC.