LC-MS/MS metabolomics unravels the resistant phenotype of carbapenemase-producing Enterobacterales.

Abstract

Introduction: The degree of antimicrobial resistance demonstrated by carbapenemase-producing Enterobacterales (CPE) represents a growing public health challenge. Conventional methods for detecting CPE involve culture-based techniques with lengthy incubation steps. There is a need to develop rapid and accurate methods for the detection of resistance, for implementation into clinical diagnostics.

Objectives: With cellular phenotype closely linked to the metabolome, the acquisition of resistance should result in detectable differences in microbial metabolism. Accordingly, we sought to profile the metabolome of Enterobacterales isolates belonging to both CPE and non-CPE groups to identify metabolites linked to CPE.

Methods: We used liquid chromatography-mass spectrometry to profile the endo- and exometabolome of 32 Klebsiella pneumoniae and Escherichia coli isolates to identify metabolites which could predict CPE in antibiotic-free conditions after 6 h of growth.

Results: Using supervised machine learning and multivariate analysis algorithms (partial least squares-discriminant analysis, k-nearest neighbour and random forest), we identified 21 metabolite biomarkers which displayed high performance metrics for the prediction of CPE (AUROCs ≥ 0.845). Results revealed a range of alterations between the metabolomes of CPE and non-CPE isolates. Pathway analysis revealed enrichment of microbial pathways including arginine metabolism, ATP-binding cassette transporters, purine metabolism, biotin metabolism, nucleotide metabolism, and biofilm formation, providing mechanistic insight into the resistance phenotype of CPE.

Conclusion: Our models demonstrate the ability to distinguish CPE from non-CPE in under 7 h using metabolite biomarkers, showing potential for the development of a targeted diagnostic assay.