The bacteriome and metabolome of human breast milk and their association with infant growth

Abstract

Human breast milk is a complex species-specific biological fluid universally known as the optimal postnatal source of nutrition for infants and therefore, recommended by the World Health Organization as the exclusive food for infants in the first six months of life. Despite the importance of human breast milk in infant health, study of its composition, especially the bacteriome (bacterial communities) and metabolome (complete set of metabolites), and their relationship to infant health and growth have not yet been comprehensively characterized. This is particularly true in low- and middle-income countries. We therefore conducted a cross-sectional study, nested within an existing birth cohort, the Drakenstein Child Health Study, to describe the bacteriome and metabolome of human breast milk samples collected between 6-10 weeks postpartum from lactating women living in South Africa. The determinants of these components of human breast milk and their role in infant growth were also investigated. Four commercial DNA extraction kits were compared for DNA extraction from human breast milk samples. The kit showing the best results, including quality and quantity of DNA, as well as best reproducibility, was chosen for further extractions. Using 16S rRNA gene amplicon next generation sequencing, a reproducible bioinformatics sequencing pipeline, and robust multivariate statistical analysis, we confirmed the presence of a diverse bacterial community in human breast milk and identified a core bacteriome, present in 80% of the samples. The bacteriome was shown to cluster into three different profile groups (biotypes) according to the predominant bacterial genus present. Bacterial interactions were suggested by the finding of positive correlations between the relative abundances of bacteria usually found in the oral or skin microbiota. Apart from study site (a proxy for ethnicity in this study), infant birth length and maternal age, no other associations were found between potential sociodemographic and psychosocial determinants and the composition of the human breast milk bacteriome. Using Nuclear Magnetic Resonance spectroscopy, we quantified forty-nine metabolites in all human breast milk samples. A subset of women with low levels of lactose concentrations were identified. Low lactose was associated with an increase in metabolites associated with mixed acid fermentation and microbial dysbiosis (staphylococcal-predominant biotype). Low-lactose (vs normal lactose) human breast milk correlated with a reduced median duration of exclusive breastfeeding and reduced infant growth (reduced weight and length z-scores) during the period of exclusive breastfeeding. These results suggest that bacterial fermentation of lactose results in low-lactose breast milk, which in turn impacts on breastfeeding outcome. Taken together, the results presented in this thesis provide a better understanding of human breast milk composition among lactating mothers living in South Africa, their potential determinants and their role in infant growth. Knowledge about the composition of human breast milk may provide opportunity for diagnostic and therapeutic interventions and help promote (exclusive) breastfeeding for the recommended period to improve infant health.