Early prediction of the bactericidal and bacteriostatic effect of imipenem and doxycycline using tabletop scanning electron microscopy.

Abstract

Our work aims at establishing a proof-of-concept for a method that allows the early prediction of the bactericidal and bacteriostatic effects of antibiotics on bacteria using scanning electron microscopy (SEM) as compared to traditional culture-based methods. We tested these effects using Imipenem (bactericidal) and Doxycycline (bacteriostatic) with several strains of sensitive and resistant Escherichia coli. We developed a SEM-based predictive score based on three main criteria: Bacterial Density, Morphology/Ultrastructure, and Viability. We determined the results for each of these criteria using SEM micrographs taken with the TM4000Plus II-Tabletop-SEM (Hitachi, Japan) following an optimized, rapid, and automated acquisition and analysis protocol. We compared our method with the traditional culture colony counting gold standard method and classic definitions of the two effects. Our method revealed total agreement with the CFU method and classic definition by visualizing the effect of the antibiotic at 60 minutes and 120 minutes using SEM. This early prediction allows a rapid and early identification of the bactericidal and bacteriostatic effects as compared to culture that would take a minimum of 18 hours. This has several future applications in the development of SEM-automated assays coupled to machine learning models that identify the antibiotic effect and facilitate determination of bacterial susceptibility.