Extended-spectrum-beta-lactamase (ESBL)/AmpC-producing Escherichia coli strains are widely found in E. coli isolates from broiler feces, largely due to the presence of the blaCTX-M-1 gene on IncI1 plasmids. Plasmid carriage is theorized to cause fitness loss and thus should decrease under conditions of reduced antibiotic use. However, in vitro studies showed plasmid carriage to increase in the absence of antimicrobials, due to plasmid conjugation. We investigated whether this translates to increased levels of plasmid in the gastrointestinal tracts of chickens, where conjugation rates may be different and subtle differences in growth rates may have a larger impact on colonization. Eight groups of five chickens were orally inoculated at 4 days of age with a 0.5-ml volume containing 106 CFU/ml E. coli cells, of which 0%, 0.1%, 10%, or 100% carried the IncI1 plasmid with the gene blaCTX-M-1. At 13 time points during 41 days, fecal samples were taken from each chicken. E. coli strains with and without plasmids were quantified. Trends in E. coli subpopulations were analyzed using generalized linear mixed models, and population dynamics were studied by fitting to a mechanistic model. Trends in E. coli subpopulations were different between groups rather than between individual chickens, suggesting substantial levels of E. coli exchange between chickens in a group. The IncI1 plasmid carrying blaCTX-M-1 was transferred with conjugation coefficients at levels higher than those observed in vitro. Across groups, the plasmids disappeared or were established independently of the initial fraction of plasmid-carrying E. coli, but no major increase occurred as observed in vitro. Differences in growth rates were observed, but competitive exclusion of plasmid-carrying variants was counteracted by conjugation.
IMPORTANCE Bacteria that produce extended-spectrum beta-lactamases are resistant to an important class of antimicrobials in human and veterinary medicine. Reduction in antibiotic use is expected to decrease the prevalence of resistance. However, resistance genes often lie on plasmids which can be copied and transferred to other bacteria by conjugation, so in vitro resistance was observed to increase in the absence of antimicrobials. We sought to determine whether this also occurs in the chicken gut and if competitive exclusion by similar E. coli variants without the resistance occurred. We studied the excretion of E. coli carrying IncI1 plasmids with the blaCTX-M-1 resistance gene in small groups of broiler chickens, after inoculating the chickens with E. coli suspensions containing different fractions of plasmid-carrying cells. Our results showed little variation between chickens within groups but large differences between groups that were independent of the ratio of variants with and without the plasmid and with persistence or extinction of the plasmid. However, there was no major plasmid increase as observed in vitro. We conclude that in vivo studies with sufficient independent replications are important for intervention studies on plasmid-mediated antimicrobial resistance.