Background: Colistin is a critical last-resort antibiotic against Gram-negative bacteria. Colistin binds to the lipopolysaccharide (LPS), displacing divalent cations to alter the net negative charge and increase cell permeability, leading to cell death1. Modification of the bacterial cell membrane, and subsequently increased resistance amongst Enterobacterales, is caused by genes such as pmrA/B, phoQ and mcr-12. Induction of these genes can occur via high concentrations of iron (Fe3+), increased acidity, and changes in magnesium (Mg2+) and calcium (Ca2+) concentrations1. Rapid testing for colistin susceptibility often fails to detect phenotypic resistance as these genes are not always expressed, and membrane modifications are not instantaneous2. Our study aimed to determine whether supplementation with known inducers of LPS modifying genes would lead to earlier phenotypic change, thereby improving the accuracy of rapid susceptibility tests.
Methods: Six Enterobacterales isolates were standardised to 5×105 cells/mL and exposed to colistin sulphate concentrations (0.06 to 4 mg/L) determined by the European Committee on Antimicrobial Susceptibility Testing (EUCAST) wild-type minimum inhibitory concentration (MIC) distributions3. Colistin and iron III chloride (FeCl3), hydrochloric acid (HCl) and ethylenediaminetetraacetic acid (EDTA) were co-incubated at ranges within the natural tolerance of Escherichia coli. Isolates were analysed using the Attune™ NxT flow cytometer following two hours of exposure, and a 24-hour broth microdilution (BMD) measured the MIC and bacterial tolerance to inducers. Results were compared to a known susceptible control strain, ATCC 25922, and an mcr-1 expressing control strain, ATCC 13846.
Results: Addition of iron III chloride caused an alteration of the autofluorescence phenotype from bimodal to trimodal, with increasing concentrations inducing the persistence of an intermediate population. EDTA at low concentrations did not alter the phenotype, however MIC values by BMD decreased with increasing concentrations and bacterial growth ceased at 10mM. Tolerance to HCl varied across isolates, with MIC values fluctuating by BMD and staining uptake decreasing at higher concentrations.
Conclusion: Supplementation with known inducers of LPS modifying genes resulted in substantial changes in the phenotype via flow cytometry, however these were not comparable with MIC values by BMD. Further studies are required to induce bacterial LPS modification, to expedite colistin susceptibility testing.