Metatranscriptomic Analysis of the Chicken Gut Resistome Response to In-Feed Antibiotics and Natural Feed Additives

Abstract

The emergence of resistance against common antibiotics in the gut microbiota is a major issue for both human and livestock health. This highlights the need for understanding the impact of such application on the reservoir of antibiotic resistance genes in poultry gut and devising means to circumvent the potential resistome expansion. Phytogenic feed additives (PFAs) are potential natural alternative to antibiotic to improve animal health and performance, supposedly via positively affecting the gut microbial ecosystem, but there is little systematic information available. In this time-course study, we applied a shotgun meta-transcriptomics approach to investigate the impact of a PFA product as well as the commonly used antibiotic, zinc bacitracin either at AGP concentration or therapeutic concentration on the gut microbiome and resistome of broiler chickens raised for 35 days. Over the course of the trial, PFA treatments increased the abundance of Firmicutes such as Lactobacillus and resulted in a lower abundance of Escherichia, while the latter group increased significantly in the feces of chickens that received either AGP or AB doses of bacitracin. Tetracycline resistance and aminoglycoside resistance were the predominant antibiotic resistance gene (ARG) classes found, regardless of the treatment. PFA application resulted in a decrease in abundance of ARGs compared to those in the control group and other antibiotic treatment groups. In summary, the findings from this study demonstrate the potential of phytogenic feed additives could be an alternative to antibiotics in poultry farming, with the added benefit of counteracting antimicrobial resistance development.

Keywords: antibiotic resistance genes; chicken; gut microbiome; metatranscriptomics; phytogenic feed additives; resistome.

FIGURE 1

(A) Richness and (B) diversity (Shannon) of ARGs across feeding treatments (CON, PFA, AGP, AGP + PFA, AB, AB + PFA) and sampling points (D14, D21, D35). Bars in the boxplot represent interquartile ranges (25th to 75th percentile). The horizontal black line represents the median. Whiskers show intervals going from –1.5 to + 1.5 of the interquartile range. Dots indicate values falling outside of the interquartile range. Significance was tested with a Mann–Whitney U test (“*” = 0.05). Colors indicate different treatments.

FIGURE 2

NMDS ordination based on Bray-Curtis dissimilarity metric represents bacterial compositional differences between the treatments.

FIGURE 3

Heatmap representing the number of unique ARGs found in each treatment that were not found in any other treatment at any given timepoint.

FIGURE 4

Relative transcript abundances of the top 10 ARG classes (A) and mechanisms (B) in chicken fecal samples from all feeding treatments (CON, PFA, AGP, AGP + PFA, AB, AB + PFA) and sampling points (D3, D14, D21, D35). The ARG classes and mechanisms with a relative abundance of <1% of the total reads were grouped into “Other”. To facilitate comparisons, the D3 timepoint was represented as a single feeding treatment alongside the other six because all samples at D3 could be considered replicates.

FIGURE 5

Differential expression of antibiotic resistance genes after administration of treatments and compared against the control (CON). X-axis: treatments by timepoint; “CON” represents the control sample. Y-axis: gene IDs as per the MEGARes2 database, together with their group and their class. The color reflects the detected log2foldChange (LFC) value. Only significantly (p < 0.05) differentially expressed genes are shown.

FIGURE 6

(A) Richness and (B) diversity (Shannon) of taxa across treatments and over sampling points. Bars in the boxplot represent interquartile ranges (25th to 75th percentile). The horizontal black line represents the median. Whiskers show intervals going from –1.5 to + 1.5 of the interquartile range. Dots indicate values falling outside of the interquartile range. Significance was tested with a Mann–Whitney U test (“*” = 0.05). Colors indicate different treatments.

FIGURE 7

Relative abundance of the most abundant taxa (≥1% of the total reads) in each treatment at each time point at the genus (A) and phylum (B) levels; taxa representing < 1% of the total reads were grouped together under the label “Other”.

FIGURE 8

Differential abundance of bacterial genera compared against the control (CON). X-axis: treatments by timepoint; “CON” represents the control sample. Y-axis: genera. The color reflects the detected log2foldChange (LFC) value. Only significant (p < 0.05) differences in taxa abundance are shown.