Background:Multidrug-resistant (MDR) Klebsiella pneumoniae poses significant global health challenges, with polymyxins often serving as a last-resort treatment. However, polymyxin resistance can emerge with monotherapy. While the polymyxin-zidovudine combination showed superior antimicrobial effectiveness and minimised resistance development, the metabolic responses of K. pneumoniae to this combination remain unclear. Here, we present iKpCairns, a detailed manually-curated model, to systematically explore the metabolic flux alterations under combination therapy.
Methods:iKpCairns was constructed from a comprehensive collection of experimental data with extensive manual curation for an MDR K. pneumoniae clinical isolate KpCairns. Growth capabilities and gene essentiality were analysed with experimentally validated Biolog assay and a transposon mutant library. RNA-seq was performed with samples collected from a polymyxin B-zidovudine time-kill study. Context-specific genome-scale models were then generated by integrating transcriptomics data with iKpCairns. Comparison of reaction flux distributions and pathway analysis were conducted with Kullback-Leibler divergence and PageRank algorithm.
Results:iKpCairns contains 1,438 metabolites, 2,980 reactions, and 1,975 genes, and demonstrates high accuracies in predicting nutrient utilisation (84.5%) and gene essentiality (86.3%). After integration with transcriptomics data, we generated 2 context-specific GSMMs. Comparison of flux distributions shows that a total of 89 metabolic fluxes were significantly increased and 135 were significantly decreased after 4h combination treatment; these included enhanced fluxes through lipopolysaccharide biosynthesis, tricarboxylic cycle and fatty acid biosynthesis, and altered fluxes in nucleotide and amino acid metabolism. These flux changes suggest that synergistic effects stem from induced changes in cell membrane formation, central metabolism, and energy production, while the metabolic alterations in nucleotide and amino acid metabolism may potentially reduce resistance emergence.
Conclusions:This study leverages context-specific metabolic modelling to delineate the mechanistic basis of synergistic killing against MDR K. pneumoniae. Such insights pave the way for developing robust, model-driven hypotheses to better understand pathogen responses under antibiotic treatment and to design more effective treatment strategies.