Use Of Targeted Metabolomics In The Relationship Between Healthy Lifestyle Index And Hepatocellular Carcinoma In The EPIC Study
Nada ASSI, International Agency for Research on Cancer (IARC), France
LEITZMANN M. 2
, STEPIEN M. 1
, SCALBERT A. 1
, JENAB M. 1
, VIALLON V. 3,4,5
, FERRARI P. 1
1 Nutrition and Metabolism Section, International Agency for Research on Cancer (IARC), Lyon, France
2 Department of Epidemiology and Preventive Medicine, Regensburg University, Regensburg, Germany
3 Université de Lyon, 69622 Lyon, France
4 UMRESTTE, Université Lyon 1, 69373 Lyon, France
5 UMRESTTE, IFSTTAR, 69675 Bron, France
Purpose: Metabolomics has the potential to disclose pathological processes leading to a better understanding of disease development. The “meeting-in-the-middle” principle is particularly suitable to explore such data, as it may reveal exposure-specific biomarkers that are predictive of morbid conditions. Using this concept, components of a healthy lifestyle index (HLI) were explored with respect to serum metabolites in a nested case-control study on hepatocellular carcinoma (HCC) within the EPIC cohort.
Methods: Using 147 HCC cases and 147 matched controls, Partial Least Squares (PLS) analysis related 7 HLI variables (“predictors”, including diet, body mass index (BMI), physical activity, lifetime alcohol consumption, smoking, diabetes and hepatitis) to 132 metabolites (“responses”), acquired using standard targeted metabolite profiling protocols (BiocratesKit). A series of multiple PLS was further applied using each HLI variable separately. All resulting PLS scores were related to HCC risk in conditional logistic regression models.
Results: One overall PLS factor was retained. Its HLI component was associated with low levels of lifetime alcohol, BMI, smoking and diabetes and high levels of physical activity. Its metabolic counterpart was positively related to sphingolipids C14:1, C16:1, C22:2, and negatively with glutamate, hexoses, and glycerophospolipids C32:0, C32:1. Both components displayed a decreased HCC risk with odds ratios (OR) equal to 0.46 (95% CI: 0.43-0.95, P = 0.03) and 0.81 (0.72-0.91, P<0.001) for a 1 SD change in the components’ scores. Specific metabolic signatures of BMI, smoking and lifetime alcohol were found to be statistically significantly associated with an increased HCC risk, with OR=1.35 (1.12-1.62, P<0.01), 1.31 (1.13-1.51, P<0.001) and 1.47 (1.03-1.27, P=0.01), respectively.
Conclusions: In a setting with targeted metabolomics, this study explored relations between metabolites, lifestyle variables and HCC risk, using an integrative approach to maximize the informative potential of high-throughput data with respect to cancer risk.
Funding: INCA, EDISS Lyon.