Search

Sporadic colorectal cancer is characterized by a multistep progression from normal epithelium to precancerous low-risk and high-risk adenomas to invasive cancer. Yet, the underlying molecular mechanisms of colorectal carcinogenesis are not completely understood. Within the “Metabolomic profiles throughout the continuum of colorectal cancer” (MetaboCCC) consortium we analyzed data generated by untargeted, mass spectrometry-based metabolomics using plasma from 88 colorectal cancer patients, 200 patients with high-risk adenomas and 200 patients with low-risk adenomas recruited within the “Colorectal Cancer Study of Austria” (CORSA). Univariate logistic regression models comparing colorectal cancer to adenomas resulted in 442 statistically significant molecular features. Metabolites discriminating colorectal cancer patients from those with adenomas in our dataset included acylcarnitines, caffeine, amino acids, glycerophospholipids, fatty acids, bilirubin, bile acids and bacterial metabolites of tryptophan. The data obtained discovers metabolite profiles reflecting metabolic differences between colorectal cancer and colorectal adenomas and delineates a potentially underlying biological interpretation.

The identification of patients at high-risk for colorectal cancer (CRC) recurrence remains an unmet clinical need. The aim of this study was to investigate associations of metabolites with risk of recurrence in stage II/III CRC patients. A targeted metabolomics assay (128 metabolites measured) was performed on pre-surgery collected EDTA plasma samples from n = 440 newly diagnosed stage II/III CRC patients. Patients have been recruited from four prospective cohort studies as part of an international consortium: Metabolomic profiles throughout the continuum of CRC (MetaboCCC). Cox proportional hazard models were computed to investigate associations of metabolites with recurrence, adjusted for age, sex, tumor stage, tumor site, body mass index, and cohort; false discovery rate (FDR) was used to account for multiple testing. Sixty-nine patients (15%) had a recurrence after a median follow-up time of 20 months. We identified 13 metabolites that were nominally associated with a reduced risk of recurrence. None of the associations were statistically significant after controlling for multiple testing. Pathway topology analyses did not reveal statistically significant associations between recurrence and alterations in metabolic pathways (e.g., sphingolipid metabolism p = 0.04; pFDR = 1.00). To conclude, we did not observe statistically significant associations between metabolites and CRC recurrence using a well-established metabolomics assay. The observed results require follow-up in larger studies.

Demographic, lifestyle and biospecimen-related factors at the time of blood collection can influence metabolite levels in epidemiological studies. Identifying the major influences on metabolite concentrations is critical to designing appropriate sample collection protocols and considering covariate adjustment in metabolomics analyses. We examined the association of age, sex, and other short-term pre-blood collection factors (time of day, season, fasting duration, physical activity, NSAID use, smoking and alcohol consumption in the days prior to collection) with 133 targeted plasma metabolites (acylcarnitines, amino acids, biogenic amines, sphingolipids, glycerophospholipids, and hexoses) among 108 individuals that reported exposures within 48 h before collection. The differences in mean metabolite concentrations were assessed between groups based on pre-collection factors using two-sided t-tests and ANOVA with FDR correction. Percent differences in metabolite concentrations were negligible across season, time of day of collection, fasting status or lifestyle behaviors at the time of collection, including physical activity or the use of tobacco, alcohol or NSAIDs. The metabolites differed in concentration between the age and sex categories for 21.8% and 14.3% metabolites, respectively. In conclusion, extrinsic factors in the short period prior to collection were not meaningfully associated with concentrations of selected endogenous metabolites in a cross-sectional sample, though metabolite concentrations differed by age and sex. Larger studies with more coverage of the human metabolome are warranted.

Cachexia is a multifactorial syndrome that is characterized by loss of skeletal muscle mass in cancer patients. The biological pathways involved remain poorly characterized. Here, we compare urinary metabolic profiles in newly diagnosed colorectal cancer patients (stage I−IV) from the ColoCare Study in Heidelberg, Germany. Patients were classified as cachectic (n = 16), pre-cachectic (n = 13), or non-cachectic (n = 23) based on standard criteria on weight loss over time at two time points. Urine samples were collected pre-surgery, and 6 and 12 months thereafter. Fat and muscle mass area were assessed utilizing computed tomography scans at the time of surgery. N = 152 compounds were detected using untargeted metabolomics with gas chromatography−mass spectrometry and n = 154 features with proton nuclear magnetic resonance spectroscopy. Thirty-four metabolites were overlapping across platforms. We calculated differences across groups and performed discriminant and overrepresentation enrichment analysis. We observed a trend for 32 compounds that were nominally significantly different across groups, although not statistically significant after adjustment for multiple testing. Nineteen compounds could be identified, including acetone, hydroquinone, and glycine. Comparing cachectic to non-cachectic patients, higher levels of metabolites such as acetone (Fold change (FC) = 3.17; p = 0.02) and arginine (FC = 0.33; p = 0.04) were observed. The two top pathways identified were glycerol phosphate shuttle metabolism and glycine and serine metabolism pathways. Larger subsequent studies are needed to replicate and validate these results.