Impaired Gut Microbial Community Development in Undernourished Children

The healthy growth of children is typically considered from an anthropometric perspective: i.e. changes in height and weight over time. Another feature of postnatal development involves the acquisition of our microbial communities, the largest of which resides in our gut. Malnutrition (undernutrition) in children, and its severity, is defined by the degree to which their anthropometric scores deviate from median values established by a World Health Organization reference cohort of 8440 individuals living in six countries. Epidemiologic studies have shown that moderate to severe forms of acute undernutrition are not due to food insecurity alone. The human gut microbiota can be thought of as a microbial ‘organ’ that plays important roles in extracting and metabolizing food ingredients, providing metabolites to the host and shaping development of the immune system. The central hypotheses of my thesis are that this microbial organ undergoes definable stages in its development following birth, that features of its developmental program are shared across biologically unrelated individuals living in distinct geographic locales and representing distinctive cultural traditions, that this developmental program is disrupted in undernourished children, and that such disruption is not merely an effect of undernutrition but is causally related to it. My thesis consists of three parts. The first part is a ‘Perspective’ describing the hypotheses described above, and describing approaches that might be useful for linking the identification of bacterial taxa that define normal development of the gut microbiota during the first several years of postnatal life to (i) an analysis of how this developmental program may be linked to the risk for, or the expression of the manifestations of undernutrition, and (ii) how knowledge of complementary feeding practices could be applied to developing new ways to sponsor robust development of the microbiota in individuals where this program has already been perturbed. In the second part, I define normal gut microbiota development in unrelated children with healthy growth phenotypes who live in an urban slum of Dhaka, Bangladesh. I did so by applying a machine-learning method (Random Forests) to bacterial 16S rRNA datasets generated from fecal samples collected monthly from birth through 24 months of life. I identified a group of ‘age-discriminatory’ bacterial strains whose changing representation in gut microbiota over time provide a signature of the developmental biology of the gut microbial community. I used this Random Forests-derived model to create two metrics that define the state of maturation of a given child’s microbiota relative his/her chronologic age; ‘relative microbiota maturity index’ and ‘microbiota-for-age Z score’. Using these metrics, I found that children with severe acute malnutrition (SAM) have immature gut microbiota (i.e. the configuration of their gut communities is younger than expected based on their chronologic age) and that this immaturity is incompletely and only transiently improved by two commonly used therapeutic food interventions. In the third part, I expand my Random Forests-based modeling of gut microbiota development by studying members of birth cohorts living in India, South Africa, Peru, and Brazil; finding that features of microbiota development (age-discriminatory strains) are shared across populations representing diverse geographic locations and cultural traditions. I also present a preclinical model for identifying complementary foods that could be used to repair the persistent microbiota immaturity present in children with SAM. This model was created by (i) culturing nine age-discriminatory bacterial strains as well as seven SAM-associated strains from the fecal microbiota of Bangladeshi children, (ii) introducing these strains into germ-free mice, (iii) feeding the animals different sequences of a prototypic Bangladeshi diet supplemented with different combinations of commonly consumed complementary foods, and (iv) analyzing 16S rRNA datasets generated from the recipient animal’s fecal microbiota in order to identify foods that promote the representation of age-indicative but not SAM-associated strains. A follow-up study in gnotobiotic mice of one of the lead complementary foods discovered from these analyses confirmed that it promotes microbiota maturation, as well as sponsoring an increase in butyrate levels and the representation of colonic regulatory T cells