Pathogenic Escherichia coli Possess Elevated Growth Rates under Exposure to Sub-Inhibitory Concentrations of Azithromycin

Abstract

Antimicrobial resistance (AMR) has been identified by the World Health Organization (WHO) as one of the ten major threats to global health. Advances in technology, including whole-genome sequencing, have provided new insights into the origin and mechanisms of AMR. However, our understanding of the short-term impact of antimicrobial pressure and resistance on the physiology of bacterial populations is limited. We aimed to investigate morphological and physiological responses of clinical isolates of E. coli under short-term exposure to key antimicrobials. We performed whole-genome sequencing on twenty-seven E. coli isolates isolated from children with sepsis to evaluate their AMR gene content. We assessed their antimicrobial susceptibility profile and measured their growth dynamics and morphological characteristics under exposure to varying concentrations of ciprofloxacin, ceftriaxone, tetracycline, gentamicin, and azithromycin. AMR was common, with all organisms resistant to at least one antimicrobial; a total of 81.5% were multi-drug-resistant (MDR). We observed an association between resistance profile and morphological characteristics of the E. coli over a three-hour exposure to antimicrobials. Growth dynamics experiments demonstrated that resistance to tetracycline promoted the growth of E. coli under antimicrobial-free conditions, while resistance to the other antimicrobials incurred a fitness cost. Notably, antimicrobial exposure heterogeneously suppressed bacterial growth, but sub-MIC concentrations of azithromycin increased the maximum growth rate of the clinical isolates. Our results outline complex interactions between organism and antimicrobials and raise clinical concerns regarding exposure of sub-MIC concentrations of specific antimicrobials.

Keywords: Escherichia coli; antimicrobial resistance; azithromycin; fitness cost; growth rate; morphology; physiology; resistome; tetracycline.

Figure 1

Bacterial morphology associated with antimicrobial resistance (AMR) and resistome. (a) The morphological features of E. coli after 3 h of being treated with ciprofloxacin. Brightfield images of E. coli in different morphologies, including, from left to right, cell lysis, ellipsoid, rod-like shape, and elongation, were acquired with a Nikon Ni-E upright microscope and a 100× objective lens. Morphology of the reference strain is in rod-like shape in drug-free condition. (b) The phylogenetic structure of E. coli with corresponding susceptibility, AMR genes, and morphology. (Left) The unrooted phylogenetic tree constructed from core genomes of the 28 E. coli isolates. (Right) Sequence type (ST) of those isolates obtained by mapping the whole-genome reads to a database of seven housekeeping genes. (Middle) Five heatmap blocks, each of which demonstrates (first column) antimicrobial susceptibility, (second column) morphology under exposure to antimicrobials, and (remaining columns) the presence (black) and absence (gray) of resistance genes against quinolone (CIP: ciprofloxacin), macrolide (AZI: azithromycin), $\beta$-lactam (CRO: ceftriaxone), tetracycline (TE), and aminoglycoside (CN: gentamicin). NF: non-identified ST.

Figure 2

The dynamics of the growth features. From top to bottom, lag-phase period ($\lambda$), maximum growth rate ($\mu$), maximum cell density (A), and area under the curve (AUC) corresponding to antimicrobial concentration relative to MICs of 28 E. coli isolates from clinical samples. From left to right, ciprofloxacin (CIP), azithromycin (AZI), ceftriaxone (CRO), tetracycline (TE), and gentamicin (CN). The horizontal axis is antimicrobial concentrations relative to MICs of each isolate. The vertical axis represents the values of growth features. Bold lines are loess smooth functions and thin lines are the growth features of each isolate.

Figure 3

A comparison between the growth features for susceptible isolates and those for resistant isolates. From top to bottom, maximum cell density (A), maximum growth rate ($\mu$), lag-phase period ($\lambda$), and area under the curve (AUC) of resistant isolates (dark gray) and those of susceptible isolates (light gray) without antimicrobial (left) and at half MICs (right). *: p-value < 0.05, **: p-value < 0.01, ***: p-value < 0.001, ****: p-value < 0.0001.

Figure 4

A comparison between growth features of resistant and susceptible isolates under exposure to azithromycin and tetracycline. The horizontal axis is antimicrobial concentrations relative to MICs. The vertical axis is values of the four growth features, including lag time ($\lambda$), maximum growth rate ($\mu$), maximum cell density (A), and area under the curve (AUC). For each isolate, growth features were plotted against relative concentrations of antimicrobials as thin lines. Bold red (resistant isolates) and blue (susceptible isolates) lines indicate loess smooth regression. Gray shading represents 95% confident intervals.

Figure 5

The impact of the ermB gene on growth dynamics of azithromycin-resistant E. coli. (a) Growth features of E. coli in the presence (blue thin lines) and absence (red thin lines) of ermB. Vertical axis is values of $\lambda$: lag time, $\mu$: maximum growth rate, A: maximum cell density, and AUC: area under the curve. Horizontal axis is antimicrobial concentration relative to MICs. Bold red and blue lines are loess smooth regression. Gray shading is 95% confident intervals. (b) Maximum growth rate ($\mu$) of ermB^- E. coli (red boxes) compared to that of ermB⁺ E.coli (blue boxes). Horizontal axis is the absolute concentration of azithromycin (mg/L). (c) Comparison MICs of ermB⁺ and ermB^- E. coli. *: p-value < 0.05.