Predicting the re-distribution of antibiotic molecules caused by inter-species interactions in microbial communities

Abstract

The use of antimicrobials without imposing selection on resistant mutants is conjectured (1, 2) to stop the rise of multi-drug resistance, but proof is still elusive. Here I present experimental evidence, underpinned by a mathematical model, showing that antimicrobial sensitivity can be predictably manipulated to achieve the sustained drug efficacy expected from evolution-proof therapies. The model relies on neighbouring microbial species often found in polymicrobial environments. The neighbours can act as drug or carbon sink depending on their drug sensitivity, changing the relative abundance of drug molecules within a focal species and influencing its sensitivity to antimicrobials. Aided by this theory, I doubled the sensitivity of Escherichia coli MC4100 to tetracycline in 24h sensitivity tests. Importantly, the effect was maintained after 168h of serial passages akin to those used in evolutionary biology (3). My results show that evolutionary-proof therapy design is, indeed, possible. My theory provides a framework to design synthetic neigh-bours that maximise drug efficacy, while minimising selection on resistance, opening a new venue in drug therapy design.