Abstract

Today, undernutrition affects approximately 150 million children worldwide. During the first two years of life, the healthy gut microbial community undergoes a series of predictable developmental transitions. These transitions are delayed in undernourished children resulting in microbiota “immaturity.” In a foundational study, immature gut microbial communities from undernourished children and normally developing communities from chronologically age-matched healthy children living in the same area were transplanted into groups of germ-free mice consuming the same diet – one representative of what consumed by the study population. Compared mice with a mature gut microbiome, mice with an immature gut microbiome had impaired weight gain, altered bone morphology, and metabolic abnormalities thus providing preclinical evidence of a causal of role of the gut community in the pathogenesis of undernutrition. This led to a central hypothesis that this thesis is examining: disruption of the normal co-development of the gut microbiome and host is a contributing cause of undernutrition and repairing gut microbial immaturity may an effective therapeutic approach for this pervasive global health challenge. A microbiota-directed therapeutic food (MDCF-2) has been designed that targets age- and growth-associated members of the gut microbial community whose abundances and expressed functions are perturbed in children with undernutrition. This MDCF capacity to repair the microbiota of undernourished children was first tested in gnotobiotic mice and pigs and subsequently in a 3-month-long randomized controlled trial involving12-18-month Bangladeshi children with moderate acute malnutrition (MAM). Compared to children treagted with a ready-to-use supplementary food (RUSF), those consuming MDCF-2 had significantly greater ponderal growth (as measured by weight-for-length z-score [WLZ]), even though MDCF-2 was 15% less calorically dense than RUSF. This MDCF-2 mediated recovery was coupled with greater levels of plasma proteins involved in musculoskeletal development and neurodevelopment; in a 2-year post-treatment follow up, a significant improvement in linear growth was also documented. To understand the role that members of the microbial community played in this recovery, metagenome associated genomes (MAGs) were first reconstructed using datasets generated from deep shotgun sequencing of fecal microbiomes serially sampled from participants in the randomized controlled clinical study prior to, during and after treatment with MDCF-2 or RUSF. Pooling MAGs from all participants followed by dereplication yielded 1000 high quality MAGs. Linear mixed effects models were then used to identify WLZ-associated MAGs. Followup in silico, mcSEED-based, reconstructions of metabolic pathways represented in these MAGs revealed a prominent enrichment in those related to carbohydrate utilization in WLZ-associated MAGs. Microbial RNA-Seq of the same fecal samples used for shotgun DNA- sequencing disclosed that the WLZ response to MDCF-2 was correlated with expression of genes in carbohydrate utilization pathways involved in metabolism of MDCF-2 glycans (and with the results of mass spectrometry-based quantification of microbial products of degradation of these glycans in the fecal samples). Importantly, differential expression of carbohydrate utilization pathways in MDCF-2-treated microbiomes was llinked to a substantial degree to two MAGs, positively associated with WLZ, that were assigned to Prevotella copri. Numerous P. copri strains were subsequently isolated from these fecal samples One strain, P. copri BgF5_2, had genomic features that were most similar to the two WLZ-associated P. copri MAGs – these shared features included polysaccharide utilization loci (PULs) that target MDCF-2 glycans. When this strain was introduced into germ-free mice as part of a defined microbial community composed of cultured, genome-sequenced, Bangladeshi bacterial strains representing age- and WLZ-associated MAGs, mice consuming MDCF-2 gained more weight compared to mice that lacked Prevotella and were significantly more efficient in metabolizing MDCF-2 glycans. P. copri BgF5_2 contains an uncommon PUL predicted to degrade an uncommon plant polysaccharide: glucomannan. This locus encodes a multidomain CAZyme previously not seen in human-derived microbes; it was expressed in E. coli purified and shown to metabolize glucomannan to its constituent glucose and mannose monomers. In vitro, P. copri BgF5_2 was able to grow in medium where the only carbon source was glucomannan. When glucomannan was added to the diets of gnotobiotic mice containing the defined community, it increased the fitness of P. copri BgF5_2, altered its metabolic features, and impacted aspects of intestinal mucosal energy metabolism. Combined this body of work demonstrates that P. copri (1) is associated with weight gain in undernourished children, (2) can be enriched by specific dietary supplements, in this case glucomannan, and (3) affects host physiology. It also illustrates a clinical translational medicine pipeline for discovering and testing therapeutic foods designed to repair the gut microbiomes and growth faltering of children with undernutrition.

Committee Chair

Jeffrey Gordon

Committee Members

Barbara Warner; Clay Semenkovich; Daniel Goldberg; Megan Baldridge; Scott Hultgren

Degree

Doctor of Philosophy (PhD)

Author's Department

Biology & Biomedical Sciences (Molecular Genetics & Genomics)

Author's School

Graduate School of Arts and Sciences

Document Type

Dissertation

Date of Award

4-15-2026

Language

English (en)

Author's ORCID

https://orcid.org/0000-0003-4335-6894

Download

Included in

Microbiology Commons

Share

COinS