Maternal adiposity alters the human milk metabolome: associations between nonglucose monosaccharides and infant adiposity

Abstract

ABSTRACT Background Human milk composition is altered by maternal obesity. The association between milk metabolites and infant outcomes has not been thoroughly investigated. Objectives This study aimed to quantify maternal adiposity-related differences in the human milk metabolome and to identify metabolites associated with infant adiposity during the first 6 mo postpartum using untargeted metabolomics. Method Maternal anthropometrics were assessed ≤14 weeks of gestation. Human milk samples were collected at 0.5 mo (n = 159), 2 mo (n = 131), and 6 mo (n = 94) postpartum from normal weight (NW, BMI = 18.5–24.9 kg/m2) and obese (OB, BMI >30 kg/m2) mothers. GC-time-of-flight-MS was used to identify metabolic signatures that discriminate NW and OB women. Partial least squared (PLS)-discriminant analysis, and PLS-regression models were assessed to examine relations between metabolites and maternal BMI and fat mass. Metabolites altered by maternal obesity were used in linear mixed effect models to predict infant adiposity. Results Multivariate modeling identified 23, 17, and 10 metabolites that described maternal adiposity indices at 0.5 mo, 2 mo, and 6 mo postpartum, respectively. Monosaccharides and sugar alcohols were the most representative annotated metabolite classes that were increased in milk from OB women and included: mannose, ribose, lyxose, lyxitol (0.5 mo); mannose, ribitol, glycerol, isothreonic acid, lyxitol (2 mo); lyxitol and isothreonic acid (6 mo). Other discriminant metabolites included: 1-monostearin, xylonolactone, shikimic acid, pseudo uridine, and dodecanol (0.5 mo); N-acetyl-D-hexosamine and fumaric acid (2 mo); uric acid and tyrosine (6 mo). Mannose, lyxitol, and shikimic acid predicted higher infant adiposity over the first 6 mo of life. Conclusions This study reports on 1 of the largest cohorts to date examining the metabolic profiles in human milk comparing NW and OB women. Maternal adiposity was associated with increased amounts of milk nonglucose monosaccharides. Human milk metabolomics may be useful in predicting infant adiposity. These trials were registered at www.clinicaltrials.gov as NCT01131117 and NCT02125149.