Metabolic profiling of adherence to diet, physical activity and body size recommendations for cancer prevention

Abstract

Maintaining a healthy body weight, eating well and being physically active lowers cancer risk by 30%. However, the biology underlying these relationships is not well understood. We examined cross-sectional associations between metabolites and cancer preventive behaviors as well as the relevance to cancer-related pathways among 120 participants (50% men, mean BMI 26.6 kg/m², mean age 54 years) with no history of smoking or cancer. Participants completed questionnaires, physical measurements and provided blood samples. Non-targeted nuclear magnetic resonance captured 223 metabolite measures. Factor analysis was performed separately for amino acid, fatty acid and lipoprotein groups. Multivariable-adjusted linear regression was used to evaluate associations between cancer preventive recommendations and metabolite-containing factors (p-value < 0.05, false discovery rate <0.20). An inflammation-related metabolite (glycoprotein acetylation) loaded strongly on a factor that was associated with excess adiposity (body fat ≥25% (men) or ≥30% (women) ß (SE) = 0.74 (0.18)) and not meeting physical activity recommendations (ß (SE) = 0.40 (0.20)). Insulin sensitivity-related metabolites including monounsaturated and polyunsaturated fats were lower among participants not meeting recommendations for adiposity, fruits and vegetables and physical activity while branched chain amino acids were higher. Cancer preventive behaviors were associated with complex metabolic signatures, including alterations in pathways known to be involved in cancer pathogenesis.

Figure 1

Factor loading plot for fatty acids metabolites. Grey bars indicate positive loadings, white bars indicate negative loadings, dashed lines indicate the threshold of 0.4 for identifying metabolites that load on a factor.

Figure 2

Factor loading plot for amino acid factors 1–4. Grey bars indicate positive loadings, white bars indicate negative loadings, dashed lines indicate the threshold of 0.4 for identifying metabolites that load on a factor. Abbreviations: FA; fatty acid, LA; linoleic acid, MUFA; monounsaturated fatty acids, AA; amino acid, C; cholesterol, CE; cholesterol ester, D; diameter, DHA; docosahexaenoic acid, PUFA; polyunsaturated fatty acids, SFA; saturated fatty acids, UnSat; estimated degree of unsaturation.

Figure 3

Factor loading plot for lipoprotein factors 1–4. Grey bars indicate positive loadings, white bars indicate negative loadings, dashed lines indicate the threshold of 0.4 for identifying metabolites that load on a factor. Abbreviations: C; cholesterol, CE; cholesterol ester, D; diameter, FC; free cholesterol, HDL; high density lipoprotein, IDL-intermediate density lipoprotein, L; lipids, LDL; low density lipoprotein, P; particles, PC; phosphatidylcholine, PG; phosphoglycerides, PL; phospholipid, SM; sphingomyelins, TG; triglyceride, VLDL; very low density lipoprotein.

Figure 4

Factor loading plot for lipoprotein factors 5–12. Grey bars indicate positive loadings, white bars indicate negative loadings, dashed lines indicate the threshold of 0.4 for identifying metabolites that load on a factor. r indicates lipid ratio, e.g. rM-VLDL-C is total cholesterol to total lipids ratio in medium VLDL. Abbreviations: C; cholesterol, CE; cholesterol ester, D; diameter, FC; free cholesterol, HDL; high density lipoprotein, IDL-intermediate density lipoprotein, L; lipids, LDL; low density lipoprotein, P; particles, PC; phosphatidylcholine, PG; phosphoglycerides, PL; phospholipid, SM; sphingomyelins, TG; triglyceride, VLDL; very low density lipoprotein.