Diverse mutant selection windows shape spatial heterogeneity in evolving populations

Abstract

Mutant selection windows (MSWs), the range of drug concentrations that select for drug-resistant mutants, have long been used as a model for predicting drug resistance and designing optimal dosing strategies in infectious disease. The canonical MSW model only offers comparisons between two phenotypes at a time: sensitive and resistant. The fitness landscape model with N alleles allows comparisons between N genotypes simultaneously, but does not encode drug dosing data. In clinical settings, there may be a wide, spatio-temporal range of drug concentrations selecting for a variety of adaptive genotypes. There is a need for a more robust model of the pathogen response to therapy to predict resistance and design optimal dosing strategies. Fitness seascapes, which model genotype-by-environment interactions, permit multiple MSW comparisons simultaneously by encoding genotype-specific dose-response data. In this work, we show how N-allele fitness seascapes embed N ∗ 2N unique MSW comparisons. In three clinically relevant pharmacokinetic models, we show how fitness seascapes reveal heterogeneous MSWs, extending the MSW model to more accurately reflect the selection of drug resistant genotypes. Our work highlights the importance and utility of considering dose-dependent fitness seascapes in evolutionary medicine.

A low-footprint, fluorescence-based bacterial time-kill assay for estimating dose-dependent cell death dynamics

“…Individual colonies were then counted by hand after 24 hrs incubation. While AB and CFU cell count estimates for the no-drug condition aligned closely, the CFU estimate for the 10 µg/mL CTX condition diverged sharply after the initial measurement (CFU = ∼ 3 ∗ 10 3 cells µL -1 , AB = ∼ 1.8 ∗ 10 5 cells µL -1 after 80 minutes).…”

“…In order to validate the estimated cell count from the fluorescence assay, we performed a standard colony-forming unit (CFU) assay in parallel for 0 and 10 μg/mL CTX ( Fig S1 ). For CFU estimates, samples were diluted 10 4 and 10 5 times and plated on antibiotic-free LB agar plates. Individual colonies were then counted by hand after 24 hrs incubation.…”

“…To optimize the time each condition spent in the dynamic range of the assay, we used different starting concentrations of cells for the two above- and below-MIC conditions. For the above-MIC conditions (conditions where the cell count was not expected to increase), we initialized the experiment with ∼2×10 5 cells/mL, which we obtained by diluting the overnight culture by 4x with fresh media. For the below-MIC conditions, we initialized the experiment with ∼2 * 10 3 cells/mL by diluting the overnight culture by 100x.…”

Reinforcement learning informs optimal treatment strategies to limit antibiotic resistance

“…We assume that selection under drug therapy represents a strong-selection and weak mutation regime in order to compute transition matrices for our models. It is likely that other selection regimes emerge in cases of real-world pharmacokinetics or spatial regimes where the drug concentration fluctuates dramatically ( 54 , 55 ). While we relax some of the strongest assumptions in the SSWM regime via a previously studied phenomenological model ( 27 , 56 ), we still do not capture the possibility of deleterious or multiple simultaneous mutations to fix.…”

Reinforcement Learning informs optimal treatment strategies to limit antibiotic resistance

“…We assume that selection under drug therapy represents a strong-selection and weak mutation regime in order to compute transition matrices for our models. While this is likely true in most cases, it is possible that other selection regimes emerge in cases of real world pharmacokinetics or spatial regimes where the drug concentration fluctuates dramatically 48,49 . In addition, we chose to keep drug concentration constant throughout are analysis, largely owing to the lack of robust empirical data linking genotype to phenotype under dose varying conditions (sometimes called a fitness seascape) 50 .…”