Bugs, brains and babies: the role of gut microbiota in preterm brain development

Abstract

The gut microbiota is increasingly recognised as a modulator of brain and behaviour. However, to date there is limited knowledge of the relationships between the early-life gut microbiota and brain structure in preterm infants, who are at risk for both altered brain development and gut bacterial dysbiosis.

This thesis aimed: to develop a brain imaging phenotype capturing diffuse white matter injury, a prevailing feature of the encephalopathy of prematurity (EoP) along with deep grey matter volume loss; to review the evidence that gut microbiota diversity and community composition associate with human brain development in infancy and childhood; to characterise and understand the drivers of preterm early-life gut microbiota diversity and community composition; and to investigate associations between the preterm gut microbiota and features of the EoP at term-equivalent age.

In the first study, using brain diffusion MRI data of 79 term and 141 preterm infants recruited to the Theirworld Edinburgh Birth Cohort (TEBC), I applied data reduction to diffusion MRI metrics across white matter tracts to derive single- and multi-metric latent general factors capturing diffuse white matter disease. I found that the general factors are influenced by preterm birth, suggesting they provide useful measures of global white matter microstructure to investigate upstream determinants of brain dysconnectivity among preterm children. I used these metrics in subsequent microbiota-brain analyses.

In the second study, I identified twenty studies that suggested the microbiome-gut-brain axis may operate across three domains in infancy and early childhood: general neurocognitive development, socio-emotional behaviours, and brain structure and function. However, there was substantial variation in the bacteria-brain/behaviour relationships reported, and I identified sources of clinical and methodological heterogeneity in the studies. I showed that there is very little understanding about microbiota-brain relationships in preterm infants in the extant literature.

In the third study, in 147 infants recruited to TEBC, I found that the majority of meconium samples were characterised by high relative abundances of Staphylococcus or Streptococcus, which minimally differed between term and preterm neonates. The microbiota community of preterm infants by the time of discharge from hospital had high relative abundances of Bifidobacterium or Enterobacteriaceae; some samples also had high abundances of Escherichia/Shigella, Klebsiella, or Enterococcus. The preterm infant meconium sample composition was strongly influenced by mode of delivery, while the main drivers of gut microbiota composition prior to discharge from the hospital were the degree of prematurity, sex, and antibiotic exposure.

Finally, I took a whole-brain approach to investigate the associations between gut microbiota diversity and community composition prior to hospital discharge and brain imaging features of the EoP at term-equivalent age in 79 preterm neonates recruited to TEBC. Using dimensionality reduction, I identified four main axes of variance in the microbiota community composition data, driven by the abundances of Bifidobacterium, Escherichia/Shigella-Enterobacteriaceae, Klebsiella, or Enterococcus. I found that microbiota richness and the main axes of variance in community composition correlated with brain microstructure, particularly in the deep grey and white matter. This study provided the first evidence that the gut microbiota associates with common neuroimaging features of preterm brain dysmaturation.

This thesis provides evidence that general factors of diffusion MRI are useful for capturing global white matter changes associated with preterm birth; that the gut microbiota plays a role in a range of neurodevelopmental domains; and that the degree of prematurity is one of the main drivers of the preterm infant gut microbiota, which in turn may contribute to deep grey and white matter dysmaturation at term-equivalent age.