Antibiotics (ABX) are among the most prescribed medications around the world, yet treatment often results in strong perturbations of the gut microbiome via off-target anti-microbial activity. When the host microbial community is depleted, there is an increased risk of opportunistic infection or dominance by native antibiotic-resistant microbes, e.g. Enterococcus faecalis. In healthy subjects, return to a stable microbial community is typically achieved in the weeks following treatment; however, the trajectory and completeness of this recovery is variable. Therefore, promoting a faster return of the microbiome to a “healthy” state is key in reducing microbially-linked antibiotic side effects.

One modifiable factor which potentially influences the trajectory of antibiotic recovery is host dietary factors. However, studying this in humans is challenging, due to the difficulties of dietary tracking and compliance, as well as recruitment. To address this, a colony dog model was utilized. Forty dogs starting with clinical indications requiring antibiotic treatment were recruited and fed either a standard maintenance (n = 20) or a veterinary therapeutic food (n = 20) enriched in fiber shown previously to support the growth of gut commensals. Fecal samples were periodically collected during and after ABX treatments and associated metadata were collected, including fecal metagenomic, metatranscriptomic, and metabolomic profiling, as well fecal moisture and ammonia content.

In our preliminary analyses, fecal moisture was significantly lower in the prescription food fed group from 11 days after treatment onwards (q < 0.1), indicating a response to the food. This was accompanied by significantly higher Shannon diversity at ABX treatment end, as well as 1, 3, and 4 weeks after (q < 0.1). Because pre-treatment samples were not available, a synthetic baseline was created by collecting metagenomes from healthy colony dogs as part of the One Health Microbiome Resource. A dysbiosis score was defined as the median Bray-Curtis dissimilarity between an experimental sample and the background, representing a whole-microbiome score of health within this cohort. The dysbiosis score was significantly lower in the prescription food fed group at 0-, 1-, and 3-week timepoints following the end of treatment. This was accompanied by a higher relative abundance at the end of treatment of an Erysipelotrichaceae species (q = 0.01) in the prescription group, as determined by MaAsLin 3. Future work will use time-series modeling and taxon-clustering techniques, as well as integration of functional and metabolomic data, to further explore the outcomes and microbial mechanisms of fiber supplementation during and after antibiotics treatment.