Early exposure to antibiotics in the neonatal intensive care unit alters the taxonomic and functional infant gut microbiome

Abstract

Introduction

The infant gut microbiome is thought to play a key role in developing metabolic and immunologic pathways. Antibiotics have been shown to disrupt the human microbiome, but the impact they have on infants during this key window of development remains poorly understood. Through this study, we further characterize the effect antibiotics have on the gut microbiome of infants by looking at metagenomic sequencing data over time.

Materials and methods

Stool samples were collected on infants from a large tertiary care neonatal intensive care unit. After DNA extraction, metagenomics libraries were generated and sequenced. Taxonomic and functional analyses were then performed. Further directed specimen sequencing for fungal species was also performed.

Results

A total of 51 stool samples from 25 infants were analyzed: seven infants were on antibiotics during at least one of their collection time points. Antibiotics given at birth altered the microbiome (PERMANOVA R² = 0.044, p = .002) but later courses did not (R² = 0.023, p = .114). Longitudinal samples collected while off antibiotics were more similar than those collected during a transition on or off antibiotics (mean Bray–Curtis distance 0.29 vs. 0.63, Wilcoxon p = .06). Functional analysis revealed four microbial pathways that were disrupted by antibiotics given at-birth (p < .1, folate synthesis, glycerolipid metabolism, fatty acid biosynthesis, and glycolysis). No functional changes associated with current antibiotic use were identified. In a limited sample set, we saw little evidence of fungal involvement in the overall infant microbiome.

Conclusion

Through this study, we have further characterized the role antibiotics have in the development of the infant microbiome. Antibiotics given at birth were associated with alterations in the microbiome and had significant impact on the functional pathways involved in folate synthesis and multiple metabolic pathways. Later courses of antibiotics led to stochastic dysbiosis and a significant decrease in Escherichia coli. Further characterization of the infant mycobiome is still needed.

Keywords:

• Bifidobacterium
• Escherichia coli
• FishTaco
• metagenomics
• mycobiome

Availability of data and materials

Sequence data has been deposited to the NCBI Sequence Read Archive under BioProject accession number PRJNA521878. All code and intermediate files to reproduce the analyses are available at https://github.com/fanli-gcb/NICU_microbiome.

Author contributions

JMB, FL, and GMA conceptualized the study. JMB, CC, HP, TC, and NF enrolled subjects and assisted with specimen processing. JMB, SZ, and LH performed sequencing and other experiments involved in this study. FL performed formal analysis of the data. JMB and FL wrote the original draft of the paper and created all of the figures. SC and PSP assisted throughout with the study design and final manuscript preparation. GMA provided leadership and oversight. All authors read, edited, and approved the final manuscript.

Statement of ethics

The parents or guardians of the research subjects have given their written informed consent. The study protocol has been approved by CHLA’s committee on human research.

Acknowledgments

The authors thank the Single Cell, Sequencing, and CyTOF (SC2) Core Laboratory at the Saban Research Institute of Children’s Hospital Los Angeles for their assistance in performing the studies presented.

Disclosure statement

No potential conflict of interest was reported by the authors.

Additional information

Funding

JMB and this project were supported through a Children’s Hospital Los Angeles [K12 HD 052954-09] Young Investigator Award.