Overexpression of antibiotic resistance genes in hospital effluents over time

Abstract

Objectives: Effluents contain a diverse abundance of antibiotic resistance genes that augment the resistome of receiving aquatic environments. However, uncertainty remains regarding their temporal persistence, transcription and response to anthropogenic factors, such as antibiotic usage. We present a spatiotemporal study within a river catchment (River Cam, UK) that aims to determine the contribution of antibiotic resistance gene-containing effluents originating from sites of varying antibiotic usage to the receiving environment.

Methods: Gene abundance in effluents (municipal hospital and dairy farm) was compared against background samples of the receiving aquatic environment (i.e. the catchment source) to determine the resistome contribution of effluents. We used metagenomics and metatranscriptomics to correlate DNA and RNA abundance and identified differentially regulated gene transcripts.

Results: We found that mean antibiotic resistance gene and transcript abundances were correlated for both hospital ( ρ = 0.9, two-tailed P <0.0001) and farm ( ρ = 0.5, two-tailed P <0.0001) effluents and that two β-lactam resistance genes ( bla GES and bla OXA ) were overexpressed in all hospital effluent samples. High β-lactam resistance gene transcript abundance was related to hospital antibiotic usage over time and hospital effluents contained antibiotic residues.

Conclusions: We conclude that effluents contribute high levels of antibiotic resistance genes to the aquatic environment; these genes are expressed at significant levels and are possibly related to the level of antibiotic usage at the effluent source.

Figure 1

(a) Mean normalized ARG abundance across the three sample types: hospital effluent, farm effluent and background sample of river source water. The ARG abundance for each sample was normalized to the number of 16S sequences before averaging values for each sample type. Error bars depict standard errors for mean values. (b) Bubble plot showing the normalized abundance of MGEs compared with the number of bacterial species in each sample. The bubble size corresponds to the normalized ARG abundance in each sample.

Figure 2

Mean ARG and ARG transcript abundance for all hospital effluent and farm effluent samples. Mean ARG and ARG transcript abundances were highly correlated for hospital (ρ = 0.9, two-tailed P <0.0001) and farm (ρ = 0.5, two-tailed P <0.0001) effluents. β-Lactam resistance genes bla_GES and bla_OXA are indicated by plus and diamond symbols, respectively. Asterisks indicate overexpressed genes (bla_GES and bla_OXA) as determined by t-test and FDR of log DNA:RNA abundance values.

Figure 3

Monthly change in the relative abundance of β-lactam ARG transcripts compared with the relative β-lactam antibiotic usage for the month preceding sample collection. The asterisk indicates detection of β-lactam antibiotics in hospital effluent using LC–MS.