Unpredictability of the fitness effects of antimicrobial resistance mutations across environments inEscherichia coli

Abstract

The evolution of antimicrobial resistance (AMR) in bacteria is a major public health concern. When resistant bacteria are highly prevalent in microbial populations, antibiotic restriction protocols are often implemented to reduce their spread. These measures rely on the existence of deleterious fitness effects (i.e., costs) imposed by AMR mutations during growth in the absence of antibiotics. According to this assumption, resistant strains will be outcompeted by susceptible strains that do not pay the cost during the period of restriction. Hence, the success of a given intervention depends on the magnitude and direction of fitness effects of mutations, which can vary depending on the genetic and environmental context. However, the fitness effects of AMR mutations are generally studied in laboratory reference strains and estimated in a limited number of environments, usually a standard laboratory growth medium. In this study, we systematically measure how three sources of variation impact the fitness effects of AMR mutations: the type of resistance mutation, the genetic background of the host, and the growth environment. We demonstrate that while AMR mutations are generally costly in antibiotic-free environments, their fitness effects vary widely and depend on complex interactions between the AMR mutation, genetic background, and environment. We test the ability of the Rough Mount Fuji genotype-fitness model to reproduce the empirical data in simulation. We identify model parameters that reasonably capture the variation in fitness effects due to genetic variation. However, the model fails to accommodate variation when considering multiple growth environments. Overall, this study reveals a wealth of variation in the fitness effects of resistance mutations owing to genetic background and environmental conditions, that will ultimately impact their persistence in natural populations.

Most azole antifungal resistance mutations in the drug target provide cross-resistance and carry no intrinsic fitness cost

“…It is important to note that the experimental conditions used here do not necessarily reflect the selective pressure of the natural or clinical environments. Therefore, the proportion of tradeoff mutations could be more important if one considered other potential environments that could make the function of Erg11 even more critical or that could further enhance the impact of resistance amino acid substitutions, for instance elevated temperatures 17,54 .…”

“…It is important to note that the experimental conditions used here do not necessarily reflect the selective pressure of the natural or clinical environments. Therefore, the proportion of tradeoff mutations could be more important if one considered other potential environments that could make the function of Erg11 even more critical or that could further enhance the impact of resistance amino acid substitutions, for instance elevated temperatures 17,60 . Also, it is possible that the fitness cost of some resistance mutations that we measured here is larger in C. albicans than it is in S. cerevisiae, although we consider that unlikely since Erg11 is essential for S. cerevisiae but not for C. albicans according to some reports 61,62 .…”

Explaining the stable coexistence of drug-resistant and -susceptible pathogens: the Resistance Acquisition Purifying Selection model

“…If treatment itself is associated with strongly reduced transmission (as is the case in HIV), then this could in principle be modeled as a strong reduction of the R0 of resistant strains, but at least in the simulations it may be more realistic to include in the model the treatment state of hosts, where the transmission probability of hosts on treatment is significantly lower than those not on treatment [34,35] . The cost of resistance may also vary between strains and may be affected by compensatory mutations [31,36] .…”

Distinguishing mutants that resist drugs via different mechanisms by examining fitness tradeoffs

“…Another way to phrase this observation is to say that adaptive mutations often have collateral effects in environments other than the one in which they originally evolved (Pál et al, 2015). But these effects, referred to in some studies as pleiotropic effects, can be unpredictable and context dependent (Bakerlee et al, 2021; Chen et al, 2023; Geiler-Samerotte et al, 2020; Hinz et al, 2023; Jerison et al, 2020). In simpler terms, some mutants that resist Drug A will suffer a tradeoff in Drug B, but others may suffer a tradeoff in Drug C. To sum, observations from many fields suggest that the mutations that provide a benefit in one environment do not always suffer similar tradeoffs.…”

“…Instead of using a mechanistic understanding to predict a microbe’s fitness, here we use how fitness varies across environments to distinguish mutants that likely affect fitness via different mechanisms. This inverted approach to investigating the mechanisms by which mutations affect fitness has broad applications; it could be used to characterize dominant negative mutations (Flynn et al, 2024; Padhy et al, 2023), mutations with collateral fitness effects (Mehlhoff et al, 2020; Mehlhoff and Ostermeier, 2023), and in other high-throughput mutational scanning studies (Flynn et al, 2020; Fowler and Fields, 2014; Hinz et al, 2023; Starr et al, 2017).…”