Development of an in vitro biofilm model for the study of the impact of fluoroquinolones on sewer biofilm microbiota

Abstract

Sewer biofilms are likely to constitute hotspots for selecting and accumulating antibiotic-resistant bacteria (ARB) and antibiotic resistance genes (ARGs). This study aimed to optimize culture conditions to obtain in vitro biofilms, mimicking the biofilm collected in sewers, to study the impact of fluoroquinolones (FQs) on sewer biofilm microbiota. Biofilms were grown on coupons in CDC Biofilm Reactors®, continuously fed with nutrients and inoculum (1/100 diluted wastewater). Different culture conditions were tested: (i) initial inoculum: diluted wastewater with or without sewer biofilm, (ii) coupon material: concrete vs. polycarbonate, and (iii) time of culture: 7 versus 14 days. This study found that the biomass was highest when in vitro biofilms were formed on concrete coupons. The biofilm taxonomic diversity was not affected by adding sewer biofilm to the initial inoculum nor by the coupon material. Pseudomonadales, Burkholderiales and Enterobacterales dominated in the sewer biofilm composition, whereas in vitro biofilms were mainly composed of Enterobacterales. The relative abundance of qnrA, B, D and S genes was higher in in vitro biofilms than sewer biofilm. The resistome of sewer biofilm showed the highest Shannon diversity index compared to wastewater and in vitro biofilms. A PCoA analysis showed differentiation of samples according to the nature of the sample, and a Procrustes analysis showed that the ARG changes observed were linked to changes in the microbial community. The following growing conditions were selected for in vitro biofilms: concrete coupons, initial inoculation with sewer biofilm, and a culture duration of 14 days. Then, biofilms were established under high and low concentrations of FQs to validate our in vitro biofilm model. Fluoroquinolone exposure had no significant impact on the abundance of qnr genes, but high concentration exposure increased the proportion of mutations in gyrA (codons S83L and D87N) and parC (codon S80I). In conclusion, this study allowed the determination of the culture conditions to develop an in vitro model of sewer biofilm; and was successfully used to investigate the impact of FQs on sewer microbiota. In the future, this setup could be used to clarify the role of sewer biofilms in disseminating resistance to FQs in the environment.

Keywords: antibiotic resistance; biofilm; bioreactor; fluoroquinolone; sewer; wastewater.

Figure 1

Bar chart representing the relative abundance of the ten most abundant orders in field samples (wastewater and sewer biofilm) and in vitro biofilms. The samples were categorized based on sampling time (D07 and D14), experiment set (first and second), and coupon material (C, concrete; PC, polycarbonate).

Figure 2

(A) Alpha diversity (Shannon index) of field samples (wastewater, purple and sewer biofilm, blue) and in vitro biofilms as a function of the experiment set (first, green and second, orange) and the day of sampling (D₇, circle, and D₁₄, triangle). Statistical differences were assessed with ANOVA (p < 0.01) and Tukey post-hoc test (^*p < 0.05). (B) Visualization by principal coordinate analysis (PCoA) of the beta-diversity analysis carried out using the Bray-Curtis distance, constrained by the type of samples (wastewater, purple; sewer biofilm, blue; first set of experiment, green and second set of experiment, orange).

Figure 3

Percentages of resistance to ciprofloxacin of heterotrophic bacteria (A,C,E) and E. coli (B,D,F) in in vitro biofilms at D07 and D14. Experiments were conducted without exposure to FQs (A,B) with two coupon materials (concrete, red and polycarbonate, blue) and with exposure to FQs at low concentration (2.5 μg/L) (C,D) and high concentration (5,000 μg/L) (E,F). CIP, ciprofloxacin; NOR, norfloxacin. p-values for C-F were calculated with Dunnett’s multiple comparison test (^*p < 0.05, ^**p < 0.01, ^***p < 0.001).

Figure 4

(A) Relative abundance of qnr genes in field samples (wastewater and sewer biofilm, blue) and in vitro biofilms from the 1st and 2nd experiment at D07 and D14, depending on the addition (red) or not (green) of sewer biofilm at the initial inoculation of bioreactors, and the material of the coupons (C, concrete; PC, polycarbonate). (B) Relative abundance of qnr genes in in vitro biofilms not exposed (Control) and exposed to ciprofloxacin (CIP) and norfloxacin (NOR) at high (5,000 μg/L, purple) and low (2.5 μg/L, orange) concentration. Statistical differences were assessed with Kruskal–Wallis test and Dunn post-hoc test for (A), and Dunnett’s multiple comparison test for (B) (^*p < 0.05; ^***p < 0.001).

Figure 5

(A) Shannon diversity index based on antibiotic resistance genes and efflux pump genes counts derived from metagenomic data (^*p < 0.05 Mann–Whitney U test). (B) PCoA based on Bray-Curtis of ARG and EPG counts for metagenomic data. The ARG and EPG counts were rarified with the lowest 16S rRNA counts.

Figure 6

(A) Relative abundance of ARGs grouped according to drug family resistance. For MS, and MLS, M, macrolide; L, lincosamide; S, streptogramin. (B) Relative-abundance of EPGs grouped according to resistance profile. ARG and EPG that showed an average reads ≥10 in at least one of the sample categories (sewer biofilm, wastewater or in vitro biofilm) were included. The samples were grouped by the type of sample, the experiment (1st to 4th) and coupon material (C, concrete; PC, polycarbonate).