Environmental remodeling of human gut microbiota and antibiotic resistome in livestock farms

The human gut microbiota, a dynamic ecosystem of commensal microbes, significantly influences host health and physiology. While generally stable and resilient to short-term perturbations, the extent of recovery after environmental changes remains unclear. The gut microbiota is shaped by both host genetics and environment, with environmental factors often predominating. Changes in diet, geography, and antibiotic use readily alter microbiota composition and colonization resistance. Swine farms, with their routine antibiotic use, present high-risk environments for the enrichment and exchange of antibiotic resistance genes (ARGs) and bacteria. Antibiotic-resistant bacteria and ARGs can spread to humans through various routes, including contaminated meat and farm dust. This study aimed to longitudinally investigate the impact of swine farm environments on the gut microbiome and resistome of healthy students undergoing occupational exposure.

Numerous studies have highlighted the stability of the human gut microbiome over time and its resilience to short-term disturbances like antibiotic treatment. However, the long-term effects of major environmental shifts on the microbiome remain less explored. The influence of environmental factors such as diet, geography, and antibiotic exposure on the human gut microbiota has been widely documented, showcasing their ability to rapidly alter microbial composition and compromise colonization resistance against pathogens. The swine farm environment has been identified as a hotspot for the selection and spread of antibiotic resistance, due to the prevalent use of antibiotics in livestock. Prior research has demonstrated the potential for environmental transmission of ARGs and their bacterial hosts between livestock and humans, raising concerns about human health risks associated with exposure to these environments.

Fourteen healthy male veterinary students participated in a 3-month internship at three different large-scale swine farms in China. Fecal samples were collected at seven time points: T0 (pre-exposure), T1-T3 (during farm stay), and T4-T6 (post-exposure). 16S rRNA gene sequencing was performed on 91 fecal samples to characterize temporal changes in gut microbial community structure. Whole-metagenome shotgun sequencing (WGS) was conducted on 42 fecal samples (T0, T3, T6) from students and samples from farm workers. WGS was also used to analyze environmental samples (dust, feces, sewage, soil) from each farm. Microbial community composition was analyzed using multivariate methods (PERMANOVA, dbRDA). AR genes were identified and quantified using ShortBRED against a custom database based on CARD. SourceTracker was employed to assess microbial transmission from the environment to the human gut. Comparative genomic analysis was performed on draft genomes of high-abundance species identified in both student and environmental samples. Culture-based methods were used to analyze the genetic relatedness of *E. coli* strains. Phenotypic resistance testing was done for multiple antibiotics. Finally, a dynamic Bayesian network (DBN) model was created to predict the duration of environmental effects on gut microbial structure.

Exposure to the swine farm environment resulted in significant changes in the students' gut microbiota within one month, with a shift towards a composition more similar to that of full-time farm workers. This involved a decrease in Bacteroidetes and an increase in Proteobacteria, particularly Gammaproteobacteria. The changes partially reversed after three months of leaving the farm environment. Metagenomic analysis revealed a modest increase in the relative abundance of ARGs in the students' gut resistome during farm stay, with a 3.7% average increase. The types of AR genes impacted included β-lactamases, aminoglycoside resistance proteins, and chloramphenicol acetyltransferases. A substantial proportion (two-thirds) of new genes detected in the students' gut microbiota during the farm stay were also present in the environmental microbiome, strongly indicating microbial transmission. SourceTracker identified 142 species transmission events from various swine farm environments to the students' gut, including pathogenic species such as *Ruminococcus* spp., *Escherichia* spp., and *Pseudomonas putida*. Comparative genome analysis confirmed that these species were shared between students and their surrounding environment. Analysis of *E. coli* strains further revealed multiple clonal spread events between students, farm workers, and the farm environment. 25% of the ARGs in the students' microbiota co-localized with mobile genetic elements, indicating potential for horizontal gene transfer. Importantly, 41% of ARGs found in both human gut and environmental samples shared similar genetic contexts, including extended-spectrum β-lactamases (*bla*_CTX-M), *tet(X)*, and *qnrS*, suggesting direct gene sharing. Phenotypic resistance testing showed increases in resistance rates to multiple antibiotics among *E. coli* strains from students' samples, persisting to some degree after returning from the farm. The DBN model predicted that the observed changes in gut microbiota composition and resistome would partially reverse within 4-6 months after returning home.

This study provides direct evidence that acute environmental changes can induce persistent remodeling of the human gut microbiota and resistome. The significant sharing of ARGs and bacteria between the swine farm environment and human hosts highlights the potential for environmental exposure to contribute to the spread of antibiotic resistance. The observation that some clinically relevant AR genes persisted even after the students left the farm environment warrants further investigation. While the study sample size is relatively small and limited to male participants, the consistent trends across different farms suggest a considerable impact of the swine farm environment on the human gut microbiome. Other confounding factors, such as diet and stress, could also influence the microbiome. The potential occupational hazards associated with working in such environments necessitate further research into the precise mechanisms of AR gene exchange and transmission to quantify these risks. Quantitative risk models are needed to better inform preventive measures.

This research demonstrates that short-term exposure to swine farm environments can significantly and persistently alter the human gut microbiome and resistome, highlighting the potential for environmental transmission of antibiotic resistance. While some changes revert over time, certain clinically relevant AR genes persist. Future studies should focus on expanding the study population, investigating the roles of various environmental reservoirs in AR gene transmission, and developing more precise quantitative risk models to guide effective mitigation strategies.

The study's relatively small sample size, limited to male participants, restricts the generalizability of findings. The cross-sectional design of the environmental samples prevents a full understanding of temporal dynamics within the swine farm ecosystem. The influence of confounding factors, such as dietary changes and stress, could not be fully controlled for. More comprehensive investigation is needed to fully understand the complex interplay of factors shaping the microbiome under this particular type of environmental change.