Association of the Infant Gut Microbiome with Temperament at Nine Months of Age: A Michigan Cohort Study

Tengfei Ma, Sihan Bu, Adannaya C. Nzerem, Nigel Paneth, Jean M. Kerver, Cybil Nicole Cavalieri, Sarah S. ComstockMicroorganisms 2024 · DOI

TL;DR

This study investigated the link between the infant gut microbiome at 3 months and temperament at 9 months in a Michigan cohort. While no significant associations were found after stringent statistical correction, exploratory analyses suggested that certain gut microbial communities might be linked to temperament, with potential moderation by vitamin D intake and infant sex.

What Problem Does This Paper Address?

The gut-brain axis, the bidirectional communication pathway between the gut microbiome and the brain, is increasingly recognized for its role in development and behavior. While animal studies and some human observational studies have hinted at this connection, particularly in relation to neurodevelopmental disorders, there's a significant gap in understanding these associations in healthy infants during their crucial early developmental stages. Specifically, how the composition of the gut microbiome in early infancy might predict later temperament – the innate behavioral and emotional predispositions of an infant – remains largely unexplored. This study aims to address this by prospectively examining the relationship between the infant gut microbiome and temperament at 9 months of age.

Methods

The researchers utilized 16S rRNA gene sequencing to profile the gut microbiome composition from fecal samples collected from infants at 3 months of age. They then employed Dirichlet multinomial mixture (DMM) modeling, an unsupervised Bayesian clustering method, to group infants based on their distinct gut microbial community structures. These clusters were then analyzed for associations with infant temperament, measured using the Infant Behavior Questionnaire-Revised Very Short Form (IBQ-RVSF) at approximately 9.4 months of age. Multivariate linear regression models were used to assess these associations, adjusting for various maternal and infant covariates. Additionally, differential abundance (DA) tools (DESeq2, ANCOM-BC, and MaAslin2) were used to identify specific bacterial taxa associated with temperament. Exploratory analyses also investigated the potential moderating roles of infant sex and vitamin D intake.

Datasets

The study utilized data from the Michigan Archive for Research on Child Health (MARCH) cohort, a prospective pregnancy cohort. Fecal samples were collected from 159 singleton infants at 3 months of age. Infant temperament data was collected from the mothers of these infants at approximately 9 months of age.

Key Insights and Ideas

  • The gut microbiome is a dynamic landscape in early infancy: The study identified three distinct gut microbial community clusters, highlighting the heterogeneity of infant gut microbiomes even at a young age.
  • Temperament is complex and multifactorial: The IBQ-RVSF captured three key temperament dimensions: positive affect/surgency, negative emotionality, and orienting/regulatory capacity, providing a nuanced view of infant behavior.
  • Potential for gut-brain axis influence, but with caveats: While initial analyses suggested associations between certain microbial clusters and temperament (particularly negative emotionality), these did not hold up after rigorous statistical correction for multiple comparisons. This underscores the challenge of establishing robust causal links in complex biological systems.
  • Vitamin D as a potential moderator: Exploratory analyses hinted that vitamin D intake might attenuate the association between specific gut microbial clusters and negative emotionality, suggesting a protective role.
  • Sex-specific effects: The study also found that the association between gut microbiota clusters and negative emotionality appeared to be driven by male infants, suggesting sex-dependent pathways in the gut-brain axis.
  • Individual taxa matter: Beyond community-level clustering, specific bacterial genera like Firmicutes and Clostridioides showed differential abundance related to temperament traits, though these findings also required careful interpretation due to statistical stringency.

Key Results

  • Three distinct gut microbial clusters (A, B, and C) were identified, with cluster B showing significantly lower alpha diversity (Shannon and Chao 1 indices).
  • Univariate analyses suggested associations between cluster A and higher positive affect/surgency and negative emotionality scores compared to cluster C. However, these associations were not statistically significant after FDR correction in multivariate models.
  • Differential abundance testing indicated that Firmicutes was associated with higher positive affect/surgency, while Clostridioides was associated with lower scores.
  • In a contingency analysis, infants with poor vitamin D intake and belonging to cluster A had significantly higher negative emotionality scores compared to cluster C. This association was attenuated in infants receiving vitamin D supplements.
  • Male infants in clusters A and B showed significantly higher negative emotionality scores compared to male infants in cluster C, an association not observed in female infants.

Limitations and Future Work

The authors acknowledge several limitations:

  • Limited sample size: While 159 participants is a reasonable cohort size, a larger sample might be needed to achieve statistical power for detecting subtle associations, especially after stringent corrections.
  • Potential for bias: The cohort had a high proportion of white and college-educated mothers, which could introduce bias.
  • Single time-point sampling: Fecal samples were collected at only one time point (3 months). The gut microbiome is dynamic, and a single snapshot may not fully capture its temporal influence on development.
  • IBQ-RVSF limitations: The Infant Behavior Questionnaire-Revised Very Short Form, while widely used, may not fully capture all aspects of infant temperament, such as soothability and cuddliness.
  • 16S rRNA sequencing limitations: This method provides genus-level information but lacks strain-level resolution and cannot reveal the metabolic capabilities of the microbes.

Future work should focus on:

  • Replication: The findings warrant replication in larger, more diverse cohorts.
  • Vitamin D and sex moderation: Further investigation into the moderating roles of vitamin D and sex is crucial.
  • Shotgun metagenomics and metabolomics: Employing more advanced techniques like shotgun metagenomic sequencing and metabolomic analysis could provide deeper insights into the functional roles of the gut microbiome.
  • Longitudinal studies: Collecting multiple fecal samples over time would better capture the dynamic nature of the infant gut microbiome and its developmental trajectory.

Summary updated

Abstract

Though studies in animals and humans link the gut microbiota to brain development and control of behavior, little research has examined this connection in healthy infants. This prospective study could determine associations between infant gut microbiota at 3 months, and infant temperament at 9 months, in a prospective pregnancy cohort (Michigan Archive for Research on Child Health; <i>n</i> = 159). Microbiota profiling with 16S rRNA gene sequencing was conducted on fecal samples obtained at 3 months of age. Based on the relative abundance of gut microbiotas, three groups were identified, and each group was characterized by different microbes. Infant temperament outcomes were reported by mothers using the Infant Behavior Questionnaire-Revised Very Short Form at a mean age of 9.4 months. Fully adjusted multivariate linear regression models showed that certain clusters were associated with higher negative emotionality scores, prominently among infants who had poor vitamin D intake. However, no associations were evident between gut microbiota clusters and temperament scales after FDR correction. After using three differential abundance tools, <i>Firmicutes</i> was associated with higher positive affect/surgency scores, whereas <i>Clostridioides</i> was associated with lower scores. An association between the gut microbiota and early infancy temperament was observed; thus, this study warrants replication, with a particular focus on vitamin D moderation.
BibTeX
@inproceedings{ma2024association,
  author = {Tengfei Ma and Sihan Bu and Adannaya C. Nzerem and Nigel Paneth and Jean M. Kerver and Cybil Nicole Cavalieri and Sarah S. Comstock},
  title = {{Association of the Infant Gut Microbiome with Temperament at Nine Months of Age: A Michigan Cohort Study}},
  booktitle = {Microorganisms},
  year = {2024},
  doi = {10.3390/microorganisms12010214},
}
Figure