Probing Antibiotic Inhibition in Small Bacterial Populations With Combinatorial Droplet Microfluidics

Bacterial infections often involve small, local populations of bacteria, yet antibiotic treatment decisions are generally based on bulk population susceptibility assays. Stochastic variability among local small populations can influence susceptibility, limiting the predictive capability of bulk assays. Therefore there is a need to better understand antibiotic response in small populations. Droplet-based microfluidics enables the high-throughput production of tens of thousands of picolitre droplets, in which small populations of bacteria (e.g., 8 cells) can be encapsulated and their responses to different environmental conditions tracked. Here, we use a combinatorial droplet-generation platform, combined with microscopy and image analysis, to interrogate the responses of small populations of Escherichia coli to different bulk-determined sub-inhibitory concentrations of the antibiotics tetracycline, streptomycin, and ampicillin within a single experiment. We observe qualitatively distinct small-population responses for these antibiotics. For the bacteriostatic ribosome-targeting antibiotic tetracycline, growth varies nonmonotonically at low antibiotic concentrations. For the bactericidal ribosome-targeting antibiotic streptomycin, we observe apparent bistability, some replicate populations growing while others die. For the bactericidal cell-wall targeting antibiotic ampicillin, we observe stochastic bacterial filamentation. Our study shows how distinct phenomena impacting antibiotic susceptibility may emerge in small bacterial populations, laying a foundation for deeper studies into potential treatment implications.