Joining the resistance

See allHide authors and affiliations

Science Translational Medicine  24 Jan 2018:
Vol. 10, Issue 425, eaar7519
DOI: 10.1126/scitranslmed.aar7519


The effect of antibiotic resistance genes on the gut microbiome is examined in preterm infants before and after antibiotic administration.

It is increasingly apparent that the gut microbiome plays a variety of roles that contribute to human health and disease and that gut microbiome alterations early in life may have far-reaching health consequences. A particularly concerning function of the gut microbiome is its ability to act as a reservoir for antibiotic resistance genes. Preterm infants represent a group at high risk for developing such resistance, as almost all receive antibiotics. Rahman and colleagues evaluated how gut microbial resistance genes influence microbial community shifts in this infant population.

The authors assembled bacterial genomes derived from shotgun sequencing of stool samples longitudinally collected over 3 months from approximately 100 preterm infants, all of whom received antibiotics in the first week of life. Approximately one-third of the infants received subsequent antibiotics, and they did not show the same decreases in known gut microbial resistance genes over time as the infants who did not. Feeding regimen was the most influential environmental factor effecting resistance gene acquisition. Compared with breast feeding, bottle-feeding led to accumulation of the resistance gene class D beta-lactamase, particularly in Clostridium difficile. They found that this gene is linked to other genes that may be involved in nutrient metabolism and hypothesized that these linked genes—not the beta-lactamase—may confer a survival advantage in the bottle-fed compared with the breast-fed nutritional environment. Additionally, they predicted accurately a microbe’s propensity to increase in abundance following administration of specific antibiotics using a machine learning model validated in a subset of the data. This methodology was able to use whole bacterial genomes, not just bacterial marker genes, to predict clinical response

The study is limited by both power, given the small sample size, and generalizability, as the infants studied were at high risk for health problems and radical microbiota alterations. Additionally, the analysis did not account for genes carried on plasmids. However, the authors provide a proof-of-principle examination of reconstructed bacterial genomes that suggests this data can be interrogated to predict response to particular environmental and clinical conditions, such as feeding and antibiotic regimens. Future clinical applications could include using these models to inform antibiotic choices and avoid regimens that could increase the survival of antibiotic-resistant pathogens.

Highlighted Article

Stay Connected to Science Translational Medicine

Navigate This Article