Transcriptomic data to support in vitro semimechanistic PK/PD modelling of Polymyxin B against Acinetobacter baumannii
Résumé
Background
Time-kill experiments (TK) with Acinetobacter baumannii and polymyxin B (PMB), usually show CFU decrease followed by regrowth even at concentrations up to 16xMIC, potentially due to mutation-driven resistance or phenotypic adaptation. This dynamic is described by PD
models with heterogeneous subpopulations or adaptive resistance. The aim of this study was to use transcriptomics to characterize the adaptive resistance of A.baumannii exposed to PMB in vitro, to inform semi-mechanistic PK/PD modelling.
Methods
Two clinical A.baumannii isolates1 susceptible (AB121) and resistant (AB122) obtained before and after colistin treatment were used to perform TK experiments (3 replicates) in 200mL MHB with a starting inoculum at 108 CFU/mL. PMB concentrations showing bacterial decrease and rapid regrowth at 1h were selected for both strains (1mg/L for AB121and 32mg/L for AB122). 2ml aliquots of bacterial cultures were harvested at t = 0, 1, 2, 4
hours. Bulk RNA was extracted and depleted of stable RNAs (rRNA/tRNA) prior to sequencing. Data analysis was performed using EdgeR and WGCNA to identify strain- and treatment-specific expression patterns.
Results
For both strains, the expression of genes related to general metabolism and osmotic pressure pathways increased with time. However, gene expression differed between the two isolates after PMB exposure (Figure 1). For AB121, an overexpression of genes involved in pyruvate/acetyl-CoA metabolism, fatty acid biosynthesis and efflux pumps was observed, suggesting adaptation mechanism(s) related to the membrane. In contrast, for AB122 most of the genes overexpressed were associated with the transcription/translation and osmotic regulation pathways. AB122 carries a 10 AA insertion in pmrB, absent in AB121, and known to confer colistin resistance through constitutive eptA expression. The eptA activation was
already observed at T0, but the presence of subpopulations with different levels of expression could explain CFU regrowth.
Conclusions
This study has shown that transcriptomics may provide evidences to support the choice of PD models structure. However, the relative importance of the various highlighted pathways as well as their potential interconnections may be challenging to assess and to implement into PK/PD models.
Origine | Fichiers produits par l'(les) auteur(s) |
---|