Antibiotic interactions shape short-term evolution of resistance in E. faecalis

Ziah Dean; Jeff Maltas; Kevin B Wood

doi:10.1371/journal.ppat.1008278

. 2020 Mar 2;16(3):e1008278. doi: 10.1371/journal.ppat.1008278

Antibiotic interactions shape short-term evolution of resistance in E. faecalis

Ziah Dean ¹, Jeff Maltas ¹, Kevin B Wood ^1,^2,^*

Editor: Michael Otto³

PMCID: PMC7093004 PMID: 32119717

Abstract

Antibiotic combinations are increasingly used to combat bacterial infections. Multidrug therapies are a particularly important treatment option for E. faecalis, an opportunistic pathogen that contributes to high-inoculum infections such as infective endocarditis. While numerous synergistic drug combinations for E. faecalis have been identified, much less is known about how different combinations impact the rate of resistance evolution. In this work, we use high-throughput laboratory evolution experiments to quantify adaptation in growth rate and drug resistance of E. faecalis exposed to drug combinations exhibiting different classes of interactions, ranging from synergistic to suppressive. We identify a wide range of evolutionary behavior, including both increased and decreased rates of growth adaptation, depending on the specific interplay between drug interaction and drug resistance profiles. For example, selection in a dual β-lactam combination leads to accelerated growth adaptation compared to selection with the individual drugs, even though the resulting resistance profiles are nearly identical. On the other hand, populations evolved in an aminoglycoside and β-lactam combination exhibit decreased growth adaptation and resistant profiles that depend on the specific drug concentrations. We show that the main qualitative features of these evolutionary trajectories can be explained by simple rescaling arguments that correspond to geometric transformations of the two-drug growth response surfaces measured in ancestral cells. The analysis also reveals multiple examples where resistance profiles selected by drug combinations are nearly growth-optimized along a contour connecting profiles selected by the component drugs. Our results highlight trade-offs between drug interactions and resistance profiles during the evolution of multi-drug resistance and emphasize evolutionary benefits and disadvantages of particular drug pairs targeting enterococci.

Author summary

Antibiotics are increasingly used in combinations to combat difficult bacterial infections, including those caused by the opportunistic pathogen E. faecalis. Some pairs of drugs exhibit particularly strong inhibitory effects when used simultaneously, as each drug magnifies the effects of the other, a phenomenon known as synergy. Other drug combinations may counteract one another, an effect called antagonism. These antibiotic “interactions” form the basis of powerful antimicrobial therapies, but they may also impact the emergence of drug resistance in surprising ways. In this work, we studied the evolution of antibiotic resistance using laboratory evolution of E. faecalis populations exposed to different two-drug combinations. We found that cells adapt more rapidly to certain drug combinations, while other antibiotic pairs markedly slow resistance. The modulated adaptation results from the interplay of drug interactions and the evolved resistance to each individual drug. Despite a wide range of observed evolutionary behavior, we find that our results can be intuitively understood using simple rescaling arguments that describe how resistance mutations impact the effective concentration of each drug. These findings highlight trade-offs between drug interactions and resistance profiles during the evolution of multi-drug resistance and emphasize evolutionary benefits and disadvantages of particular drug pairs targeting enterococci.

Introduction

The rapid rise of antibiotic resistance poses a growing threat to public health [1, 2]. The discovery of new antimicrobial agents is a long and difficult process, underscoring the need for new approaches that optimize the use of currently available drugs. In recent years, significant efforts have been devoted to designing evolutionarily sound strategies that balance short-term drug efficacy with the long-term potential to develop resistance. These approaches describe a number of different factors that could modulate resistance evolution, including interactions between bacterial cells [3–8], synergy with the immune system [9], spatial heterogeneity [10–15], epistasis between resistance mutations [16, 17], precise temporal scheduling [18–21], and statistical correlations between resistance profiles for different drugs [22–31].

Drug combinations are an especially promising and widely used strategy for slowing resistance [32], and there has been significant work devoted to identifying and predicting the effects of different drug cocktails [33–39]. From a clinical perspective, synergistic interactions–where the combined effect of the drugs is greater than expected based on the effects of the drugs alone [33]–have long been considered desirable because they provide strong antimicrobial effects at reduced concentrations. By contrast, drug pairs that interact antagonistically–effectively weakening one another in combination–have been traditionally avoided. Work over the last decade has challenged this conventional wisdom by demonstrating that synergistic interactions have a potentially serious drawback: they may accelerate the evolution of resistance [40–42]. Similarly, antagonistic interactions can slow or even reverse the evolution of resistance [43]. These results indicate that drug interactions underlie a natural trade-off between short-term efficacy and long-term evolutionary potential [44]. In addition, recent work has shown that cross-resistance (or collateral sensitivity) between drugs in a combination may also significantly modulate resistance evolution [26, 27, 45]. As a whole, these studies show that drug interactions and collateral effects may combine in complex ways to influence evolution of resistance in multi-drug environments.

Antibiotic combinations are often the most effective therapies for enterococcal infections, which lead to significant morbidity and mortality [46–51]. E. faecalis is among the most commonly isolated enterococcal species and underlies a host of human infections, including infective endocarditis, infections of the urinary tract or blood stream, and infections related to surgical devices and medical implants. Multiple combination therapies have been proposed or are currently in use for E. faecalis infections [46, 52, 53]. While synergistic combinations are the standard–particularly for high inoculum infections–relatively little is known about how different combinations affect the potential for, and rate of, resistance evolution.

To address these questions, we use large scale laboratory evolution to measure growth adaptation and phenotypic resistance in populations of E. faecalis exposed to four different two-drug combinations over multiple days. The drug pairs include several clinically relevant combinations–for example, two β-lactams or a β-lactam and an amimoglycoside–and exhibit a range of interactions, from synergistic to strongly antagonistic (suppressive). Our results reveal rich and at times surprising evolutionary behavior. In all cases, we find that different dosing combinations lead to significantly different rates of growth adaptation, even when the level of inhibition is constant. In some cases, differences in growth adaptation appear to be driven by selection for distinct cross-resistance profiles, while in other cases, strong interactions between the drugs lead to different adaption rates but highly similar profiles. Despite this apparent diversity, we show that qualitative features of these evolutionary trajectories can be understood using simple rescaling arguments that link resistance profiles in evolving populations to geometric transformations of the two-drug response surface in ancestral cells. Our results represent a quantitative case study in the evolution of multidrug resistance in an opportunistic pathogen and highlight both potential limitations and unappreciated evolutionary benefits of different drug combinations.

Results

Selection of antibiotic pairs with different interactions

We first set out to identify a set of two-drug combinations that include a range of interaction types: synergistic, antagonistic, and suppressive (i.e. strongly antagonistic so that the effect of the combination is less than that of one of the drugs alone). To do so, we measured the per capita growth rate of E. faecalis V583 populations in liquid cultures exposed to multiple drug pairs at 90-100 dosage combinations per pair. We estimated per capita growth rate in early exponential phase from optical density (OD) time series acquired using an automated microplate reader and plate stacker (Methods). The type of interaction for each drug pair is defined by the shape of the contours of constant growth (“isoboles”) describing the growth response surface g(D₁, D₂), where D_i is the concentration of drug i. Linear contours of constant growth represent additive (non-interacting) pairs–for example, the effect of one unit of drug 1 or drug 2 alone is the same as that of a combination with half a unit of each drug. Deviations from additivity include synergy (antagonism), which corresponds to contours with increased concavity (convexity), rendering an equal mixture of the two drugs more (less) effective than in the non-interacting scenario. While other metrics exist for quantifying drug interactions (see, for example, [33]), we choose this Loewe null model [54] because, as we will see, its simple geometric interpretation provides useful intuition for interpreting evolutionary trajectories [41, 43]. Based on these interaction measures and with an eye towards choosing clinically relevant combinations when possible, we decided to focus on 4 drug combinations: ceftriaxone (CRO) and ampicillin (AMP); ampicillin and streptomycin (STR); ceftriaxone and ciprofloxacin (CIP); and tigecycline (TGC) and ciprofloxacin. We describe these combinations in more detail below.

Laboratory evolution across iso-inhibitory dosage combinations

Our goal was to compare evolutionary adaptation for different dosage combinations of each drug pair. The rate of adaptation is expected to depend heavily on the level of growth inhibition in the initial cultures, which sets the selection pressure favoring resistant mutants. To control for initial inhibition levels, we chose four dosage combinations for each drug pair—two corresponding to single drug treatments and two to drug combinations—that lie (approximately) along a contour of constant inhibition (i.e. constant per capita growth rate). We then evolved 24 replicate populations in each dosage combination for 3-4 days (20-30 generations) with daily dilutions into fresh media and drug. The evolution experiments were performed in microwell plates, allowing us to measure time series of cell density (OD) for each population over the course of the adaptation. In addition, we characterized the phenotypic resistance of 6 randomly chosen populations per condition by measuring standard dose-response curves and estimating the half-maximal inhibitory concentration (IC₅₀) of each drug in the combination. Together, these measurements provide a detailed quantitative picture of both growth adaptation and changes in phenotypic resistance to each drug over the course of the evolution experiment.

Dual β-lactam combination accelerates growth adaption but selects for similar resistance profiles as adaption to component drugs

Cell wall synthesis inhibitors, including β-lactams, are among the most frequently used antibiotics for E. faecalis infections [55]. While E. faecalis often exhibit sensitivity to amino-pencillins, such as ampicillin (AMP), they are intrinsically resistant to cephalosporins (e.g. ceftriaxone (CRO)). Despite the limited utility of ceftriaxone alone, it combines with ampicillin to form a powerful synergistic pair, making it an attractive option for E. faecalis harboring high-level aminoglycoside resistance. Dual β-lactam combinations like CRO-AMP have been particularly effective in treatment of endocarditis infections in the clinical setting [46, 52].

As expected, we found that the CRO-AMP combination is strongly synergistic in the ancestral V583 E. faecalis strain (Fig 1a, left panel). We selected four dosage combinations (labeled A-D) along the concave contour of constant inhibition and evolved replicate populations to each condition. Growth curves on day 0, the first day of evolution, show similar levels of inhibition for each combination, with growth initially increasing but later collapsing or plateauing. The growth curves on later days sometimes differ between replicates and between conditions (Fig 1a, right panels; S1 Fig). To quantify growth, we estimated the effective growth rate during an intermediate range of OD (0.1< OD<0.4) for each day and each condition using nonlinear least squares fitting to an exponential function (Fig 1b, left panels). For ancestor cells in the absence of drug, this regime corresponds to exponential growth (Fig 1a, right panels; S1 Fig) and this metric estimates the per capita growth rate. In cases where growth over this region is non-monotonic, this metric instead provides an effective measure of growth that decreases when population density declines, even if initial growth is rapid.

In all four conditions, growth increases significantly by day 2. Notably, late stage (days 2-3) growth rate is higher, on average, when both drugs are present (conditions B and C) than it is for the single-drug conditions (A and D). To further quantify these trends, we estimated the average adaptation rate for each population using linear regression for each growth rate time series (Fig 1b, right panel). While the 24 replicate populations in each condition show considerable variability–as expected, perhaps, for a stochastic evolutionary process–growth adaptation is significantly faster for the two combinations (B and C) than for the single drug treatments (A and D). These qualitative results do not depend sensitively on the specific OD window used to estimate growth (S2 and S3 Figs) and also hold when adaptation is estimated with a nonlinear function (S4 and S5 Figs). As an alternate way to visualize this adaptation, we also plotted the median growth curve (on final day of adaptation) calculated across all populations exposed to the same condition. The trends from this simple analysis are consistent with those from adaptation rate calculations–specifically, growth following adaptation to the drug combos is faster than growth following adaptation to the individual drugs (S6 Fig).

One might naively expect that increased growth adaption in the drug combination indicates that populations in these conditions evolve higher levels of resistance to one or both drugs. To test this hypothesis, we used replicate dose-response measurements to estimate the half-maximal inhibitory concentration (IC₅₀) of both CRO and AMP in 6 randomly selected populations for each condition at the end of each day (see S7–S14 Figs for dose-response curves and S15 Fig for estimates of relative error). We then quantified resistance as the (log₂-scaled) fold change in IC₅₀ between the evolved and ancestral populations; positive values indicate increased resistance and negative values indicate increased sensitivity relative to the ancestral strain (Fig 1c). Note that resistance measured at the end of day 0 is considered to be resistance on day 1, as it is the expected resistance of the population at the start of day 1. Populations evolved under all four conditions exhibit similar patterns of phenotypic resistance, with IC₅₀’s to each drug rising rapidly and plateauing after 1-2 days. We note in passing that the growth of the population in condition D is near 0 on day 1, even though resistance to ampicillin is already seen in the IC₅₀ measurements; this peculiar finding may suggest that the population is heterogeneous, with resistant subpopulations comprising only a small fraction of the population on day 1. In all conditions, populations tend to show higher increases in CRO resistance than AMP resistance. These results are initially surprising because they indicate that mutants with nearly identical phenotypic resistance profiles nevertheless exhibit markedly different patterns of growth adaptation that depend on the specific dosage combination.

Aminoglycoside/ β-lactam and β-lactam/fluoroquinolone combinations slow growth adaptation and select for resistant profiles distinct from those evolved to the component drugs

In addition to dual β-lactam therapies, combinations involving an aminoglycoside with a cell wall inhibiting antibiotic are commonly used for treating drug resistant E. faecalis [46]. In particular, the ampicillin and streptomycin (STR) combination has been used as a first line of treatment for E. faecalis infective endocarditis [46, 55]. Unfortunately, enterococci isolates are increasingly exhibiting high-level resistance to aminoglycosides, which has been shown to reduce the synergistic effect of the combination therapy [55]. While aminoglycoside resistance is a growing problem, the adaptation of aminoglycoside-resistant E. faecalis to combination therapies remains poorly understood.

Consistent with previous findings, we did not observe synergy between AMP and STR in the ancestral V583 strain, which exhibits considerable aminoglycoside resistance. In fact, the drug pair exhibits marked antagonism, as evidenced by the convex growth contours on the response surface (Fig 2a, left panel). Growth curves from populations evolved to four different conditions along a growth contour show considerable differences at day 2 between the single-drug conditions (conditions A and D) and the two-drug conditions (B and C, Fig 2a, right panel; S16 Fig), with single-drug populations reaching a higher growth rate (Fig 2b, left panels) and a dramatically increased rate of adaptation (Fig 2a, right panel; S17 Fig). At the end of the experiment (after day 2 of adaptation), populations grown in single drugs tend to show resistance to the selecting drug but little cross resistance to the other drug (Fig 2c, red and blue curves). Populations grown in a combination of both drugs show similar resistance profiles to those selected by the single drug that is dominant within the mixture. For example, the day 3 resistance profiles for condition B are similar to those from condition A (dominated by AMP), while those for condition C are similar to condition D (dominated by STR). It is interesting to note that the temporal dynamics (i.e. change in resistance over time for each drug) for all four conditions do show different qualitative trends, even when the final resistance profiles are similar (Fig 2c).

Fig 2 — a) Left panel: per capita growth rate of ancestral populations as a function of ampicillin (AMP) and streptomycin (STR) concentrations. Circles correspond to different selecting conditions along a contour of constant inhibition. Drug concentrations (AMP,STR) are (0.3, 0) (red, A); (0.3, 1000) (magenta, B); (0.19, 1800) (cyan, C); and (0, 1800) (blue, D). Right panels: growth curves for the first (black) and last (green) days of evolution for each condition. b) Left panels: per capita growth rate over time for each condition. Dashed lines are drug-free growth curves in ancestor strains. Shaded region is OD range over which per capita growth rate is estimated. Right panel: average rate of growth adaptation over the course of the evolution. Adaptation rate in the single drug conditions (red and blue conditions, combined) is greater than that in the combined drug conditions (magenta and cyan, combined), 1-sided t-test (unequal variance), p < 10⁻⁴. Error bars are ± standard error of the mean (SEM). c. Resistance to AMP (top panels) and STR (bottom panels) over time for isolates from different conditions. Resistance is defined as the log₂-scaled fold change in IC₅₀ of the resistant isolate relative to ancestral cells (positive is increased resistance, negative is increased sensitivity). In all plots, light transparent lines correspond to individual populations and darker lines to the mean across populations.

We observed qualitatively similar behavior in a combination of a β-lactam and fluoroquinolone (ciprofloxacin). Ciprofloxacin (CIP) is not typically used in the treatment of enterococci, though it has been used with β-lactams in the treatment of enterococcal endocarditis with high-level aminoglycoside resistance [56]. In vitro studies also demonstrate efficacy of ciprofloxacin in multiple combinations against E. faecalis biofilms [57]. We investigated resistance evolution in a combination of CIP with CRO, which is not a clinically used combination but exhibits less dramatic antagonism than the AMP-STR combination (Fig 3a, left panel), making it a potentially interesting proof-of-principle example of resistance involving fluoroquinolones.

Fig 3 — a) Left panel: per capita growth rate of ancestral populations as a function of ceftriaxone (CRO) and ciprofloxacin (CIP) concentrations. Circles correspond to different selecting conditions along a contour of constant inhibition. Drug concentrations (CRO, CIP) are (48.5, 0) (red, A); (34.7, 0.21) (magenta, B); (6.93, 0.34) (cyan, C); and (0, 0.46) (blue, D). The latter point is shifted for visualization. Right panels: growth curves for the first (black) and last (green) days of evolution for each condition. Dashed lines are drug-free growth curves in ancestor strains. Shaded region is OD range over which per capita growth rate is estimated. b) Left panels: per capita growth rate over time for each condition. Right panel: average rate of growth adaptation over the course of the evolution. Adaptation rate in the single drug conditions (red and blue conditions, combined) is greater than that in the combined drug conditions (magenta and cyan, combined), 1-sided t-test (unequal variance), p < 10⁻⁴. Error bars are ± standard error of the mean (SEM). c. Resistance to CRO (top panels) and CIP (bottom panels) over time for isolates from different conditions. Resistance is defined as the log₂-scaled fold change in IC₅₀ of the resistant isolate relative to ancestral cells (positive is increased resistance, negative is increased sensitivity). In all plots, light transparent lines correspond to individual populations and darker lines to the mean across populations.

As with the AMP-STR combination, we observe slowed growth adaptation in combinations of CRO-CIP relative to that in the component drugs alone (Fig 3b; S18 and S19 Figs). Adaptation to CRO alone leads to strong CRO resistance and slight CIP sensitivity (Fig 3c, red). On the other hand, adaptation to CIP alone leads to CIP resistance along with significant increases in CRO sensitivity (Fig 3c, blue). Populations evolved to combinations show different resistance profiles at different concentrations, though for both mixtures the collateral sensitivities are eliminated (Fig 3c, magenta and cyan).

Tigecycline suppresses growth adaptation and eliminates evolution of fluoroquinolone resistance

The fourth combination we investigated was comprised of a protein synthesis inhibitor (tigecycline, TGC) and fluoroquinolone (CIP). TGC is a relatively new broad-spectrum antibiotic used for soft-tissue infections [58]; it also shows in-vitro synergy in combination with multiple antibiotics [59, 60]. We found that the TGC-CIP combination exhibits a particularly interesting type of interaction known as suppression (Fig 4a, left panel), where the combined effect of both drugs can be smaller than the effect of one drug (in this case, TGC) alone at the same concentration. Suppressive interactions between protein and DNA synthesis inhibitors in E. coli have been previously linked to sub-optimal regulation of ribosomal genes [61] as well as inverted selection for sensitive cells [43]. Because the growth contours in this case show non-monotonic behavior, we performed evolution experiments in replicates of 8 for 11 different concentration combinations that fall along the isobole (Fig 4a; S20 and S21 Figs). We find that growth adaptation decreases approximately monotonically as TGC concentration is increased, eventually approaching a minimum as TGC eclipses a critical concentration TGC_crit ≈ 0.03 μg/mL (Fig 4c). Furthermore, while we observed TGC resistance only in rare cases, populations adapted to TGC below the critical concentration show approximately constant levels of CIP resistance, while those at higher concentrations show essentially no CIP resistance (Fig 4d). It is particularly striking that populations evolved in conditions A (red) and I (light blue) exhibit such different evolutionary behavior, as both are exposed to nearly identical CIP concentrations and, by design, start at similar levels of inhibition. Yet evolution in condition A leads to fast growth adaptation and strong CIP resistance, while evolution in condition I (light blue) leads to little adaptation and no CIP resistance. In effect, the addition of TGC eliminates CIP resistance without modulating the overall efficacy of the (initial) combination.

Fig 4 — a) Left panel: per capita growth rate of ancestral populations as a function of tigecycline (TGC) and ciprofloxacin (CIP) concentrations. Circles correspond to different selecting conditions along a contour of constant inhibition. b) growth curves for the first (black) and last (green) days of evolution for six of the 11 selecting conditions. c) Average rate of growth adaptation over the course of the evolution for each condition. d. Resistance to CIP (left) and TGC (right) for isolates on the final day of evolution. Resistance is defined as the log₂-scaled fold change in IC₅₀ of the resistant isolate relative to ancestral cells (positive is increased resistance, negative is increased sensitivity). In all plots, light transparent lines or small points correspond to individual populations. Darker lines and larger circles represent means taken across populations. Error bars are ± standard error of the mean (SEM).

Geometric rescaling of ancestral growth surface explains condition-dependent growth adaptation when resistant profiles are unchanged

To interpret the observed evolutionary dynamics, we hypothesized that mutations conferring resistance modulate the effective drug concentration experienced by the population. In some cases–for example, resistance due to efflux pumps–this effective concentration change corresponds to a genuine physical change in intracellular drug concentration. More generally, though, this hypothesis assumes that resistant cells growing in external drug concentration C behave similarly to wild type (drug-sensitive) cells experiencing a reduced effective concentration C’ < C. Similar rescaling arguments were pioneered in [41, 43], where they were used to predict correlations between the rate of resistance evolution and the type of drug interaction.

Our results indicate that adaptation rates in CRO-AMP (Fig 1) and TGC-CIP (Fig 4) combinations can vary significantly across dosage combinations, even when resistance profiles are essentially unchanged. For example, populations adapted to CRO-AMP combinations show an average increase in IC₅₀ for CRO and AMP of about 2^3.5 ≈ 11.6 fold and 2^1.3 ≈ 2.5 fold, respectively. To understand how this level of resistance might be expected to impact growth, we rescaled the concentrations of CRO and AMP that lie along the contour of constant growth that passes through the four experimental dosage combinations (A-D). For each point on the contour, the concentrations of CRO and AMP are reduced by factors of 11.6 and 2.5, respectively. For example, the points corresponding to conditions A-D are mapped to the points shown in Fig 5a (squares). The new rescaled contour, which includes the rescaled locations of the original points A-D, does not in general correspond to a contour of constant growth on the original growth surface; therefore, growth of the adapted cells is expected to differ along the contour. More specifically, to predict growth of mutants selected along the original contour of constant growth, one simply needs to find the rescaled contour, plot it atop the original (ancestral) growth surface, and read off the values of growth along that rescaled contour. In the CRO-AMP combination, this rescaling approach predicts that growth is slightly (approximately 10 percent or less) higher when the drugs are combined (e.g. conditions B and C) than when they are used individually (Fig 5a, top panel), in qualitative agreement with our experiments (Fig 1). More generally, the rescaling suggests that adaptation in all conditions should lead to dramatically increased growth, a consequence of the steep dose-dependence of the synergistic response surface.

Fig 5 — Main panels: contour plots show growth in ancestral (WT) strain as function of drug concentrations for AMP-CRO (a, left) and TGC-CIP (b, right). Circles indicate selecting conditions (note that for visualization purposes, several selecting conditions are outside of the axis limits). Squares indicate effective concentrations achieved by rescaling true concentrations by the observed (mean) fold change in IC₅₀ for each drug. All arrows in a single panel correspond to the same rescaling: AMP→ AMP/2.5, CRO→ CRO/11.6 (panel a) and CIP→ CIP/3.0, TGC→ TCG (panel b). The rescaling factors correspond to cross resistance to AMP and CRO, with CRO resistance larger than AMP resistance (panel a) and to CIP resistance with no change in TGC resistance (panel b), which correspond to the mean values observed experimentally over all populations in AMP-CRO (see Fig 1) and the CIP-resistant populations in TGC-CIP (see Fig 4). Upper panels: predicted change in growth (relative to condition A) as one moves along the original contour, starting at condition A (one drug only). Predicted growth rate is calculated using 2d interpolation to estimate the growth along the rescaled contour on the ancestral growth surface.

Similarly, populations adapted in TGC-CIP show an increase in CIP IC₅₀ of approximately 2^1.6 ≈ 3 fold, but only when adaptation occurs below a critical TGC concentration. If we apply the same rescaling approach–that is, we reduce the concentration of CIP by 3-fold for all points along the contour–we again get a series of new points that no longer fall on a single growth contour (Fig 5b, squares). Furthermore, the predicted growth for points on the new contour decreases monotonically with TGC concentration before plateauing near point G, near the critical concentration where experimental growth adaptation approaches its minimum value (Fig 5b). Intuitively, then, it becomes clear why selection for ciprofloxacin resistance is only favored below this critical concentration: for higher concentrations of TGC, the rescaled points fall very nearly on the same contour as the original point. That is, when the original contour becomes approximately vertical, rescaling the CIP concentration is no longer expected to increase growth (see, for example, point I).

Resistance profiles selected in different AMP-STR and CRO-CIP combinations are (nearly) growth-optimized along contours connecting the profiles selected by component drugs

Rescaling arguments may also help us to understand why particular resistance profiles appear to be preferentially selected under different initial conditions, as we observed with AMP-STR and CRO-CIP. The resistance profiles on the final day of adaptation fall at different points in the two dimensional space describing resistance to each drug (Fig 6a and 6b, left panels). When profiles arising from adaptation to the same condition are averaged together, the resulting profiles (large circles) fall approximately on a smooth contour. We set out to determine how profiles at different points along these contours would be expected–based on rescaling–to impact growth in each of the four selecting conditions used experimentally.

Fig 6 — a: AMP-STR, b. CRO-CIP. Left panels: resistance profiles for individual populations (small circles) and the mean across populations (large filled circles) for each selecting condition. Color scheme is the same as in previous figures (i.e. red is condition A, magenta condition B, cyan condition C, and blue condition D). Solid black lines are linear (a) or quadratic (b) fits to the averaged resistance profiles, which create smooth contours in the 2-d resistance space. Right panels: each curve shows the predicted growth for populations with a range of resistance profiles (falling on the contours, left panels) grown at one specific selecting condition (A, red; B, magenta; C, cyan; D, blue). The horizontal axis corresponds to position along the contours (left panels). Filled circles correspond to the locations of experimentally observed (mean) resistance profiles.

To approximate these contours, we used a line segment (for AMP-STR) or a quadratic contour (for CRO-CIP) that connects conditions with extreme resistance. In both cases, the contours approximately run between red and blue points, corresponding to adaptation in the single drug conditions. To investigate how different resistance levels along these contours are expected to affect growth in each selecting condition, we rescaled the drug concentrations corresponding to each selecting condition by a range of rescaling factors that lie along the contours connecting their end points (labeled X and Y). As a result, each of the original selecting conditions (more specifically the points defined by the drug concentrations describing each condition) is mapped to a smooth curve in the two-drug concentration space. Points along that curve indicate how the original selecting concentrations are mapped to rescaled drug concentrations by the resistance profiles lying along the contour connecting X and Y. From that new curve, then, one can read off the corresponding growth rates from the ancestral growth surface, leading to predicted growth rates for mutants with any particular profile along the contour.

For each selecting condition, there is an optimal resistance profile (along the contour) that leads to the maximum possible expected growth. Not surprisingly, in the case of AMP-STR, maximum growth in condition A (AMP only) occurs at point X, where AMP resistance is highest. Similarly, maximum growth in conditions D (STR only) and C occur at point Y, which has the largest STR resistance. On the other hand, the optimal resistance profile for condition B lies just short of the midpoint on the line segment connecting X and Y (Fig 6a, right panel). Interestingly, the (mean) resistance profiles observed experimentally (circles) are predicted to give growth rates within approximately 15% of the optimal value.

In the case of the CRO-CIP combination, the optimal resistance profiles (along the contour) for conditions A (CRO only) and D (CIP only) lie at the endpoints X and Y, respectively (Fig 6b, right panel), which have the highest resistance levels to the component drugs. By contrast, the optimal profiles for conditions B (magenta) and C (cyan) fall at different points along the contour, reflecting trade-offs between resistance levels and collateral sensitivities to the component drugs. Once again, the observed (mean) resistance profiles are predicted to give growth rates near the optimal values for each condition (particularly for conditions A, B, and D). Given that there is a finite number of genetic mutations possible under these short-term conditions, one would not expect that phenotypic profiles exist for all points along the contours. Nevertheless, these results suggest that for these two drug pairs, the resistant profiles selected in the combinations are expected to give growth benefits that are nearly optimal among all those possible along the contour.

Discussion

Using laboratory evolution, we have shown that adaptation of E. faecalis populations to drug combinations can differ substantially from adaptation to the component drugs. While the evolutionary trajectory of any particular population is difficult to predict, the results as a whole point to simple trends that can be explained with rescaling arguments linking growth of adapted populations to growth of the ancestral population at properly rescaled drug dosages. These arguments show, for example, how identical resistance profiles yield different growth rates for different selecting conditions. The analysis also suggests that, in multiple cases, the profiles selected by different dosage combinations are very nearly growth-optimized along the contour connecting profiles selected by the individual component drugs. Given the inherent stochasticity of individual evolutionary trajectories, it is remarkable that such simple principles can be used to understand the mean behavior across large experimental ensembles.

It is important to point out several limitations to our study. First, our goal was not to investigate the specific molecular mechanisms involved in drug adaption, but instead to provide a quantitative picture of resistance evolution that does not require extensive molecular-level knowledge, which many times is not available. However, the richness of the observed phenotypes points to complex and potentially interesting genetic changes that can be partially resolved with modern sequencing technologies. For example, cross-resistance observed between ceftriaxone and ampicillin may be due to mutations in penicillin-binding proteins, which are common resistance determinants for both drugs [55]. We also note that our experiments were performed just below the minimum inhibitory drug concentrations, allowing for slowed but nonzero rates of proliferation. Previous work indicates that drug interactions may modulate evolution in different ways at higher drug concentrations [27]. In addition, our experiments were performed in planktonic populations, while many of the high-inoculum infections requiring combination treatment are likely to involve surface-associated biofilms, where spatial heterogeneity and complex community dynamics can dramatically alter the response to antibiotics. In E. faecalis, for example, population density can significantly modulate growth dynamics [8], while sub-inhibitory doses of cell wall inhibitors may actually promote biofilm growth [62]. Recent work in other bacterial species also shows that evolutionary adaptation may differ between biofilm and planktonic communities [63, 64]. Investigating adaptation to drug combinations in these different regimes, both at clinically-relevant concentrations and in biofilms, remains an interesting avenue for future work.

There are also several notable technical limitations. To minimize batch effects due to day-to-day variability (due to small changes in, for example, drug stock, media composition or temperature), we perform all experiments for a given drug pair together during the same 3-day period using the same reagent batches. We do see considerable variability between evolutionary replicates, even when they are performed side-by-side during the same period, with the same batch of media, and under the same selection conditions, giving us hope that these conditions capture at least some of the complexity of the possible evolutionary outcomes. However, it is possible that additional trajectories could arise, and perhaps even dominate, if the experiments were repeated multiple times across different days or different growth conditions. In addition, it is clear that all growth curves are not purely exponential, and in fact the per capita growth rate can change with time. Our growth rate estimates should therefore be thought of as an effective growth rate that reduces the population dynamics each day to a single number. Similarly, the rescaling analysis assumes that the drug resistance in the population can be captured by a single IC₅₀ (for each drug), essentially neglecting clonal interference in favor of a single dominant resistant phenotype. In addition, the rescaling analysis does not incorporate fitness cost (which we did not measure). As a result of these limitations, the trends predicted by rescaling can only be evaluated qualitatively, though these limitations could potentially be overcome with significantly more experimental data, leading to more quantitative rescaling predictions. For example, resistance phenotyping of individual isolates from each population could provide insight into population heterogeneity during adaptation. Nevertheless, given the potential complexity of evolutionary trajectories, it is encouraging that simple rescaling arguments can qualitatively capture the coarse-grained features we measured. Future studies that aim to overcome the technical limitations of this work may be able to further evaluate quantitative agreement between specific evolutionary trajectories and the predictions of rescaling.

Most importantly, we stress that our results are based on in vitro laboratory experiments, which provide a well-controlled but potentially artificial–and certainly simplified–environment for evolutionary selection. While in vitro studies form the basis for many pharmacological regimens, the ultimate success or failure of new therapies must be evaluated using in vivo model systems and, ultimately, controlled clinical trials. We hope the results presented here offer a provocative look at evolution of E. faecalis in multidrug environments, but it is clear that these findings are not directly transferable to the clinic. Our results do include some clinically relevant antibiotic combinations, though we also sought a wide range of drug interaction types, including multiple antagonistic combinations that are unlikely a priori choices for clinical use. In addition, the relatively high levels of resistance of the ancestor strain to some drugs–such as aminoglycosides–would make them unlikely clinical choices to these populations. It is notable, however, that the resulting non-synergistic combinations often produced considerably slower growth adaptation, consistent with previous results that highlight potential benefits for non-standard combinations [40, 41, 43].

Finally, our results reveal that simple rescaling arguments—similar to those originally introduced in [41, 43]—can be used to understand many features of evolution in two-drug environments. Extending and formalizing these qualitative findings using stochastic models where fast evolutionary dynamics are coupled to geometric rescaling on adiabatically changing interaction landscapes is an exciting avenue for theoretical work that may provide both general insight as well as specific, experimentally testable predictions for how resistance evolves in multi-drug environments.

Materials and methods

Strains, media, and growth conditions

All experiments were performed on the V583 strain of E. faecalis, a fully sequenced clinical isolate. Overnight seed cultures were inoculated from a single colony and grown in sterilized brain heart infusion (BHI) medium at 37C with no shaking. Antibiotic stock solutions (Table 1) were prepared using sterilized Millipore water, diluted and aliquoted into single use micro-centrifuge tubes and stored at -20C or -80C. All drugs and media were purchased from Dot Scientific, Sigma-Aldrich or Fisher Scientific.

Table 1. Antibiotics used in this study.

Antibiotic	Class	Mechanism
Ampicillin (AMP)	β-lactam (penicillins)	cell wall synthesis inhibitor
Ceftriaxone (CRO)	β-lactam (cephalosporin)	cell wall synthesis inhibitor
Ciprofloxacin (CIP)	fluoroquinolone	DNA synthesis inhibitor
Streptomycin (STR)	aminoglycoside	protein synthesis inhibitor
Tigecycline (TGC)	glycylcycline	protein synthesis inhibitor

Open in a new tab

Laboratory evolution and growth measurement

Evolution experiments were seeded by diluting overnight cultures of ancestral V583 cells 400x into individual wells of a 96-well microplate containing appropriate drug concentrations. All plates for a multi-day evolution experiment were prepared in advance by adding appropriate drug concentrations to 200 μL BHI and storing at -20C for not more than 4 days. Each day, a new plate was thawed and inoculated with 2 μL (100X dilution) from the previous day’s culture. Plates were then sealed with BIO-RAD Microseal film to minimize evaporation and prevent cross contamination. Optical density at 600 nm (OD) was measured for each population every 20-25 minutes using an EnSpire Multimode Plate Reader with multi-plate stacker attachment located in a temperature-controlled (30C) warm room. Control wells containing ancestral cells and BHI medium were included on each plate as a growth control and for background subtraction, respectively. All dilutions and daily transfers were performed inside a ThermoFisher 1300 Series A2 safety cabinet to minimize contamination. Samples from each population were stocked in 15 percent glycerol and stored at -80C.

Estimating per capita growth rate and drug response surfaces

We estimated per capita growth rate (g) from OD time series by fitting the early exponential phase portion of the background subtracted curves (0.1<OD<0.4) to an exponential function (OD ∼ e^gt) using nonlinear least squares (MATLAB 7.6.0 curve fitting toolbox, Mathworks). The OD range [0.1, 0.4] was chosen because it spans a large region of early exponential phase growth in ancestor cells grown without drug. If the growth rate was estimated to be negative or the growth curve did not reach 0.1 over the course of the experiment, growth was set to 0. See S22 Fig for examples, including those where growth is clearly not exponential and therefore g should be interpreted as an effective growth parameter, not the true per capita growth. We normalized all growth rates by the growth rate of ancestral cells in the absence of drugs performed on the same day, with one exception: ancestor growth curves for CRO-AMP and CRO-CIP were shared between the two experiments, which were performed on consecutive days. Ancestor growth rates varied slightly day-to-day, with a minimum of 0.74 and a maximum of 0.88 hr⁻¹ (doubling times of 47 and 56 mins, respectively). We visualized two-drug growth response surfaces by smoothing (2-d cubic spline interpolation) to reduce experimental noise and displaying smoothed surfaces as two-dimensional heat maps. When relevant–for example, for rescaling analysis–growth in unsampled regions of the growth surface was estimated with 2d interpolation.

Phenotypic resistance profiling

Experiments to estimate the half-maximal inhibitory concentration (IC₅₀) for each population were performed in replicates of 3 in 96-well plates. Prior to IC₅₀ testing, frozen stocks for each population were swabbed and grown overnight in drug-free medium. These overnight cultures were then diluted 100X into new plates containing fresh media and a gradient of 8 drug concentrations. We chose linearly spaced drug concentrations (rather than the more conventional 2-fold dilutions [65]) to increase precision of IC₅₀ estimates. After 20 hours of growth the optical density at 600 nm (OD600) was again measured and used to create a dose response curve. To quantify drug resistance, the resulting dose response curve was fit to a Hill-like function f(x) = (1+ (x/K)^h)⁻¹ using nonlinear least squares fitting, where K is the half-maximal inhibitory concentration (IC₅₀) and h is a Hill coefficient describing the steepness of the dose-response relationship. To reduce day-to-day fluctuations, control dose response curves (in replicates of 6) in ancestral cells were measured side-by-side each day with dose response curves of the adapted populations. The IC₅₀’s from these controls were used as the normalizing factor in calculating resistance relative to ancestor (i.e. log₂-scaled fold change). See S7–S14 Figs for examples. Relative error (standard error of mean / mean) for IC₅₀ is typically on the order of ten percent and often much smaller (S15 Fig).

Estimated growth surfaces for ancestral cells as well as all estimated growth rates and IC₅₀ values are available in ref [66].

Supporting information

S1 Fig. Growth curves for adaptation in AMP-CRO combinations.

OD time series for populations grown in conditions A (top 2 rows, red), B (magenta), C (cyan), and D (last 2 rows, blue) for combinations of ceftriaxone (CRO) and ampicillin (AMP). Light gray curve is ancestor control in absence of drug. Days are green (0), blue (1), red (2), and magenta (3). Shaded region corresponds to OD range [0.1,0.4] over which growth rate is estimated.

(EPS)

Click here for additional data file.^{(1.1MB, eps)}

S2 Fig. Adaptation rate trends do not depend sensitively on choice of OD windows for growth rate fits.

Adaptation rates for populations grown in conditions A (red), B (magenta), C (cyan), and D (blue) for combinations of ceftriaxone (CRO) and ampicillin (AMP). Upper insets show the OD time series of ancestor strains in the absence of drug, with shaded region the OD range over which exponential growth rate is estimated (OD = [0.1,0.4], upper left; OD = [0.05,0.3], upper right; OD = [0.15,0.35], lower left; OD = [0.15,0.5], lower right). In all cases, adaptation rate in the single drug conditions (red and blue conditions, combined) is less than that in the combined drug conditions (magenta and cyan, combined), 1-sided t-test (unequal variance), p < 0.05.

(EPS)

Click here for additional data file.^{(371.3KB, eps)}

S3 Fig. Growth rate time series do not depend sensitively on choice of OD windows for growth rate fits.

Per capita growth rate over time for each condition, A (red), B (magenta), C (cyan), and D (blue), for combinations of ceftriaxone (CRO) and ampicillin (AMP). Rows 1-4 correspond to the four choices of OD windows in S2 Fig.

(EPS)

Click here for additional data file.^{(553.8KB, eps)}

S4 Fig. Growth rate time series and fits to linear and saturating functions.

Growth rate time series (circles) and both linear and saturating fits to determine mean adaptation rate (lines) for populations grown in conditions A (top 3 rows, red), B (magenta), C (cyan), and D (last 3 rows, blue) for combinations of ceftriaxone (CRO) and ampicillin (AMP). Saturating fits correspond to a functional form y = y_max x/(x+ K_inv), where y_max is the maximum possible growth rate (set to 1.2, slightly higher than 1 to allow for potential increases in growth relative to ancestor) and K_inv is defined as the inverse adaptation rate.

(EPS)

Click here for additional data file.^{(759KB, eps)}

S5 Fig. Adaptation rate trends do not depend sensitively on linear or saturating fits to growth rate time series.

Adaptation rates for populations grown in conditions A (red), B (magenta), C (cyan), and D (blue) for combinations of ceftriaxone (CRO) and ampicillin (AMP) for linear fits (left) and saturating fits (right). Note that the vertical axis scales differ, as do the definition of adaptation rate. In the linear case, adaptation rate is in units of growth rate per unit time, while in the saturating case adaptation rate is in units of inverse time. In all cases, adaptation rate in the single drug conditions (red and blue conditions, combined) is less than that in the combined drug conditions (magenta and cyan, combined), 1-sided t-test (unequal variance), p < 10⁻³.

(EPS)

Click here for additional data file.^{(166.3KB, eps)}

S6 Fig. Median growth curves within each condition capture qualitative trends of growth rate adaptation.

Median growth curves–which consist of the median OD across all populations within a given condition at each time point–for the four drug combinations on the final day of adaptation. Color scheme is same as in the main text: red and blue are single drug conditions, magenta and cyan correspond to drug combinations. Shaded region corresponds to OD range over which exponential growth rates are estimated in the main text.

(EPS)

Click here for additional data file.^{(183.4KB, eps)}

S7 Fig. Dose response curves to estimate AMP resistance following CRO adaptation.

Dose response curves for six populations (”mutants”, rows) over 3 days (columns) for populations adapted to condition A (CRO-only). Dots are measurements (3 technical replicates), lines are estimated fits to Hill-like dose response function. Shaded region indicates mean IC₅₀ ± 2 standard errors across replicates. Light gray curves and x symbols correspond to ancestor strain control.

(EPS)

Click here for additional data file.^{(809.7KB, eps)}

S8 Fig. Dose response curves to estimate AMP resistance following AMP-CRO adaptation.

Dose response curves for six populations (”mutants”, rows) over 3 days (columns) for populations adapted to condition B (AMP-CRO). Dots are measurements (3 technical replicates), lines are estimated fits to Hill-like dose response function. Shaded region indicates mean IC₅₀ ± 2 standard errors across replicates. Light gray curves and x symbols correspond to ancestor strain control.

(EPS)

Click here for additional data file.^{(771.2KB, eps)}

S9 Fig. Dose response curves to estimate AMP resistance following AMP-CRO adaptation.

Dose response curves for six populations (”mutants”, rows) over 3 days (columns) for populations adapted to condition C (AMP-CRO). Dots are measurements (3 technical replicates), lines are estimated fits to Hill-like dose response function. Shaded region indicates mean IC₅₀ ± 2 standard errors across replicates. Light gray curves and x symbols correspond to ancestor strain control.

(EPS)

Click here for additional data file.^{(708.4KB, eps)}

S10 Fig. Dose response curves to estimate AMP resistance following AMP adaptation.

Dose response curves for six populations (”mutants”, rows) over 3 days (columns) for populations adapted to condition D (AMP-only). Dots are measurements (3 technical replicates), lines are estimated fits to Hill-like dose response function. Shaded region indicates mean IC₅₀ ± 2 standard errors across replicates. Light gray curves and x symbols correspond to ancestor strain control.

(EPS)

Click here for additional data file.^{(798KB, eps)}

S11 Fig. Dose response curves to estimate CRO resistance following CRO adaptation.

(EPS)

Click here for additional data file.^{(801.8KB, eps)}

S12 Fig. Dose response curves to estimate CRO resistance following AMP-CRO adaptation.

(EPS)

Click here for additional data file.^{(747.9KB, eps)}

S13 Fig. Dose response curves to estimate CRO resistance following AMP-CRO adaptation.

(EPS)

Click here for additional data file.^{(714.7KB, eps)}

S14 Fig. Dose response curves to estimate CRO resistance following AMP adaptation.

Dose response curves for six populations (”mutants”, rows) over 3 days (columns) for populations adapted to condition A (AMP-only). Dots are measurements (3 technical replicates), lines are estimated fits to Hill-like dose response function. Shaded region indicates mean IC₅₀ ± 2 standard errors across replicates. Light gray curves and x symbols correspond to ancestor strain control.

(EPS)

Click here for additional data file.^{(787KB, eps)}

S15 Fig. Relative error in resistance measurements for AMP and CRO.

Histogram of relative error, defined as the standard error of the mean divided by the mean, for IC₅₀ estimates from 72 populations (6 examples per condition, 4 conditions, 3 days) per drug adapted to AMP-CRO conditions. In each case, standard error is calculated over three technical replicates.

(EPS)

Click here for additional data file.^{(52.5KB, eps)}

S16 Fig. Growth curves for adaptation in AMP-STR combinations.

OD time series for populations grown in conditions A (top 2 rows, red), B (magenta), C (cyan), and D (last 2 rows, blue) for combinations of ampicillin (AMP) and streptinomycin (STR). Light gray curve is ancestor control in absence of drug. Days are green (0), blue (1), and red (2). Shaded region corresponds to OD range [0.1,0.4] over which growth rate is estimated.

(EPS)

Click here for additional data file.^{(971.5KB, eps)}

S17 Fig. Growth rate time series for adaptation in AMP-STR combinations.

Growth rate time series (circles) and linear fits to determine mean adaptation rate (lines) for populations grown in conditions A (top 3 rows, red), B (magenta), C (cyan), and D (last 3 rows, blue) for combinations of ampicillin (AMP) and streptinomycin (STR).

(EPS)

Click here for additional data file.^{(344.2KB, eps)}

S18 Fig. Growth curves for adaptation in CRO-CIP combinations.

(EPS)

Click here for additional data file.^{(1.1MB, eps)}

S19 Fig. Growth rate time series for adaptation in CRO-CIP combinations.

Growth rate time series (circles) and linear fits to determine mean adaptation rate (lines) for populations grown in conditions A (top 2 rows, red), B (magenta), C (cyan), and D (last 2 rows, blue) for combinations of ceftriaxone (CRO) and ciprofloxacin (CIP).

(EPS)

Click here for additional data file.^{(434.8KB, eps)}

S20 Fig. Growth curves for adaptation in TGC-CIP combinations.

OD time series for populations grown in conditions A-K (11 rows, 8 replicates per condition). Light gray curve is ancestor control in absence of drug. Days are green (0), blue (1), and red (2). Shaded region corresponds to OD range [0.1,0.4] over which growth rate is estimated.

(EPS)

Click here for additional data file.^{(843.2KB, eps)}

S21 Fig. Growth rate time series for adaptation in TGC-CIP combinations.

Growth rate time series for each population (light lines) and the mean across populations for a given condition (A-K) for combinations of tigecycline (TGC) and ciprofloxacin (CIP).

(EPS)

Click here for additional data file.^{(377.4KB, eps)}

S22 Fig. Example exponential fits to growth curves.

OD time series for populations grown in conditions A (top 2 rows), B (rows 3-4), C (rows 5-6), and D (last 2 rows) for combinations of ampicillin (AMP) and streptinomycin (STR). Light gray curve is ancestor control in absence of drug. Days are green (0), blue (1), and red (2). Lines show fits to exponential functions. Shaded region corresponds to OD range [0.1,0.4] over which growth rate is estimated. Note that growth rates less than zero (see several examples on top row) were set to 0 for all subsequent analysis.

(EPS)

Click here for additional data file.^{(585.6KB, eps)}

Acknowledgments

We would like to thank Naji Kabbani and Zahra Arabzada for their technical support and contributions to the early stages of this project.

Data Availability

We have posted all data to Data Dryad (https://datadryad.org/) where it is freely available without restriction.

Funding Statement

This work is supported by the National Science Foundation (NSF No. 1553028 to KW), the National Institutes of Health (NIH No. 1R35GM124875-01 to KW), and the Hartwell Foundation for Biomedical Research (to KW). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

References

1. Davies J, Davies D. Origins and evolution of antibiotic resistance. Microbiol Mol Biol Rev. 2010;74(3):417–433. 10.1128/MMBR.00016-10 [DOI] [PMC free article] [PubMed] [Google Scholar]
2. Levy SB, Marshall B. Antibacterial resistance worldwide: causes, challenges and responses. Nature medicine. 2004;10(12s):S122 10.1038/nm1145 [DOI] [PubMed] [Google Scholar]
3. Read AF, Day T, Huijben S. The evolution of drug resistance and the curious orthodoxy of aggressive chemotherapy. Proceedings of the National Academy of Sciences. 2011;108(Supplement 2):10871–10877. 10.1073/pnas.1100299108 [DOI] [PMC free article] [PubMed] [Google Scholar]
4. Hansen E, Woods RJ, Read AF. How to use a chemotherapeutic agent when resistance to it threatens the patient. PLoS biology. 2017;15(2):e2001110 10.1371/journal.pbio.2001110 [DOI] [PMC free article] [PubMed] [Google Scholar]
5. Meredith HR, Srimani JK, Lee AJ, Lopatkin AJ, You L. Collective antibiotic tolerance: mechanisms, dynamics and intervention. Nature chemical biology. 2015;11(3):182–188. 10.1038/nchembio.1754 [DOI] [PMC free article] [PubMed] [Google Scholar]
6. Sorg RA, Lin L, Van Doorn GS, Sorg M, Olson J, Nizet V, et al. Collective resistance in microbial communities by intracellular antibiotic deactivation. PLoS biology. 2016;14(12):e2000631 10.1371/journal.pbio.2000631 [DOI] [PMC free article] [PubMed] [Google Scholar]
7. Tan C, Smith RP, Srimani JK, Riccione KA, Prasada S, Kuehn M, et al. The inoculum effect and band-pass bacterial response to periodic antibiotic treatment. Molecular systems biology. 2012;8(1):617 10.1038/msb.2012.49 [DOI] [PMC free article] [PubMed] [Google Scholar]
8. Karslake J, Maltas J, Brumm P, Wood KB. Population density modulates drug inhibition and gives rise to potential bistability of treatment outcomes for bacterial infections. PLoS computational biology. 2016;12(10):e1005098 10.1371/journal.pcbi.1005098 [DOI] [PMC free article] [PubMed] [Google Scholar]
9. Gjini E, Brito PH. Integrating antimicrobial therapy with host immunity to fight drug-resistant infections: classical vs. adaptive treatment. PLoS computational biology. 2016;12(4):e1004857 10.1371/journal.pcbi.1004857 [DOI] [PMC free article] [PubMed] [Google Scholar]
10. Zhang Q, Lambert G, Liao D, Kim H, Robin K, Tung Ck, et al. Acceleration of emergence of bacterial antibiotic resistance in connected microenvironments. Science. 2011;333(6050):1764–1767. 10.1126/science.1208747 [DOI] [PubMed] [Google Scholar]
11. Baym M, Lieberman TD, Kelsic ED, Chait R, Gross R, Yelin I, et al. Spatiotemporal microbial evolution on antibiotic landscapes. Science. 2016;353(6304):1147–1151. 10.1126/science.aag0822 [DOI] [PMC free article] [PubMed] [Google Scholar]
12. Greulich P, Waclaw B, Allen RJ. Mutational pathway determines whether drug gradients accelerate evolution of drug-resistant cells. Physical Review Letters. 2012;109(8):088101 10.1103/PhysRevLett.109.088101 [DOI] [PubMed] [Google Scholar]
13. Hermsen R, Deris JB, Hwa T. On the rapidity of antibiotic resistance evolution facilitated by a concentration gradient. Proceedings of the National Academy of Sciences. 2012;109(27):10775–10780. 10.1073/pnas.1117716109 [DOI] [PMC free article] [PubMed] [Google Scholar]
14. Moreno-Gamez S, Hill AL, Rosenbloom DI, Petrov DA, Nowak MA, Pennings PS. Imperfect drug penetration leads to spatial monotherapy and rapid evolution of multidrug resistance. Proceedings of the National Academy of Sciences. 2015;112(22):E2874–E2883. 10.1073/pnas.1424184112 [DOI] [PMC free article] [PubMed] [Google Scholar]
15. De Jong MG, Wood KB. Tuning spatial profiles of selection pressure to modulate the evolution of drug resistance. Physical review letters. 2018;120(23):238102 10.1103/PhysRevLett.120.238102 [DOI] [PMC free article] [PubMed] [Google Scholar]
16. Trindade S, Sousa A, Xavier KB, Dionisio F, Ferreira MG, Gordo I. Positive epistasis drives the acquisition of multidrug resistance. PLoS genetics. 2009;5(7):e1000578 10.1371/journal.pgen.1000578 [DOI] [PMC free article] [PubMed] [Google Scholar]
17. Borrell S, Teo Y, Giardina F, Streicher EM, Klopper M, Feldmann J, et al. Epistasis between antibiotic resistance mutations drives the evolution of extensively drug-resistant tuberculosis. Evolution, medicine, and public health. 2013;2013(1):65–74. 10.1093/emph/eot003 [DOI] [PMC free article] [PubMed] [Google Scholar]
18. Yoshida M, Reyes SG, Tsudo S, Horinouchi T, Furusawa C, Cronin L. Time-programmable dosing allows the manipulation, suppression and reversal of antibiotic drug resistance in vitro. Nat Commun. 2017;8 10.1038/ncomms15589 [DOI] [PMC free article] [PubMed] [Google Scholar]
19. Meredith HR, Lopatkin AJ, Anderson DJ, You L. Bacterial temporal dynamics enable optimal design of antibiotic treatment. PLoS computational biology. 2015;11(4):e1004201 10.1371/journal.pcbi.1004201 [DOI] [PMC free article] [PubMed] [Google Scholar]
20. Nichol D, Jeavons P, Fletcher AG, Bonomo RA, Maini PK, Paul JL, et al. Steering evolution with sequential therapy to prevent the emergence of bacterial antibiotic resistance. PLoS computational biology. 2015;11(9):e1004493 10.1371/journal.pcbi.1004493 [DOI] [PMC free article] [PubMed] [Google Scholar]
21. Fuentes-Hernandez A, Plucain J, Gori F, Pena-Miller R, Reding C, Jansen G, et al. Using a sequential regimen to eliminate bacteria at sublethal antibiotic dosages. PLoS biology. 2015;13(4):e1002104 10.1371/journal.pbio.1002104 [DOI] [PMC free article] [PubMed] [Google Scholar]
22. Imamovic L, Sommer MOA. Use of collateral sensitivity networks to design drug cycling protocols that avoid resistance development. Sci Transl Med. 2013;5:204ra132 10.1126/scitranslmed.3006609 [DOI] [PubMed] [Google Scholar]
23. Kim S, Lieberman TD, Kishony R. Alternating antibiotic treatments constrain evolutionary paths to multidrug resistance. Proc Natl Acad Sci USA. 2014;111:14494–14499. 10.1073/pnas.1409800111 [DOI] [PMC free article] [PubMed] [Google Scholar]
24. Pál C, Papp B, Lázár V. Collateral sensitivity of antibiotic-resistant microbes. Trends in microbiology. 2015;23(7):401–407. 10.1016/j.tim.2015.02.009 [DOI] [PMC free article] [PubMed] [Google Scholar]
25. Barbosa C, Trebosc V, Kemmer C, Rosenstiel P, Beardmore R, Schulenburg H, et al. Alternative evolutionary paths to bacterial antibiotic resistance cause distinct collateral effects. Molecular biology and evolution. 2017;34(9):2229–2244. 10.1093/molbev/msx158 [DOI] [PMC free article] [PubMed] [Google Scholar]
26. Barbosa C, Beardmore R, Schulenburg H, Jansen G. Antibiotic combination efficacy (ACE) networks for a Pseudomonas aeruginosa model. PLoS biology. 2018;16(4):e2004356 10.1371/journal.pbio.2004356 [DOI] [PMC free article] [PubMed] [Google Scholar]
27. Rodriguez de Evgrafov M, Gumpert H, Munck C, Thomsen TT, Sommer MO. Collateral resistance and sensitivity modulate evolution of high-level resistance to drug combination treatment in Staphylococcus aureus. Molecular biology and evolution. 2015;32(5):1175–1185. 10.1093/molbev/msv006 [DOI] [PubMed] [Google Scholar]
28. Nichol D, Rutter J, Bryant C, Hujer AM, Lek S, Adams MD, et al. Antibiotic collateral sensitivity is contingent on the repeatability of evolution. Nature communications. 2019;10(1):334 10.1038/s41467-018-08098-6 [DOI] [PMC free article] [PubMed] [Google Scholar]
29.Maltas J, Wood KB. Pervasive and diverse collateral sensitivity profiles inform optimal strategies to limit antibiotic resistance. bioRxiv. 2019. [DOI] [PMC free article] [PubMed]
30. Podnecky NL, Fredheim EGA, Kloos J, Sorum V, Primicerio R, Roberts AP, et al. Conserved collateral antibiotic susceptibility networks in diverse clinical strains of Escherichia coli. Nature Communications. 2018;9 10.1038/s41467-018-06143-y [DOI] [PMC free article] [PubMed] [Google Scholar]
31. Imamovic L, Ellabaan MMH, Machado AMD, Citterio L, Wulff T, Molin S, et al. Drug-driven phenotypic convergence supports rational treatment strategies of chronic infections. Cell. 2018;172(1-2):121–134. 10.1016/j.cell.2017.12.012 [DOI] [PMC free article] [PubMed] [Google Scholar]
32. Baym M, Stone LK, Kishony R. Multidrug evolutionary strategies to reverse antibiotic resistance. Science. 2016;351(6268):aad3292 10.1126/science.aad3292 [DOI] [PMC free article] [PubMed] [Google Scholar]
33. Greco WR, Bravo G, Parsons JC. The search for synergy: a critical review from a response surface perspective. Pharmacological reviews. 1995;47(2):331–385. [PubMed] [Google Scholar]
34. Zhao XM, Iskar M, Zeller G, Kuhn M, Van Noort V, Bork P. Prediction of drug combinations by integrating molecular and pharmacological data. PLoS computational biology. 2011;7(12):e1002323 10.1371/journal.pcbi.1002323 [DOI] [PMC free article] [PubMed] [Google Scholar]
35. Wood K, Nishida S, Sontag ED, Cluzel P. Mechanism-independent method for predicting response to multidrug combinations in bacteria. Proceedings of the National Academy of Sciences. 2012;109(30):12254–12259. 10.1073/pnas.1201281109 [DOI] [PMC free article] [PubMed] [Google Scholar]
36. Zimmer A, Katzir I, Dekel E, Mayo AE, Alon U. Prediction of multidimensional drug dose responses based on measurements of drug pairs. Proceedings of the National Academy of Sciences. 2016;113(37):10442–10447. 10.1073/pnas.1606301113 [DOI] [PMC free article] [PubMed] [Google Scholar]
37. Zimmer A, Tendler A, Katzir I, Mayo A, Alon U. Prediction of drug cocktail effects when the number of measurements is limited. PLoS biology. 2017;15(10):e2002518 10.1371/journal.pbio.2002518 [DOI] [PMC free article] [PubMed] [Google Scholar]
38. Tendler A, Zimmer A, Mayo A, Alon U. Noise-precision tradeoff in predicting combinations of mutations and drugs. PLoS computational biology. 2019;15(5):e1006956 10.1371/journal.pcbi.1006956 [DOI] [PMC free article] [PubMed] [Google Scholar]
39. Russ D, Kishony R. Additivity of inhibitory effects in multidrug combinations. Nature microbiology. 2018;3(12):1339 10.1038/s41564-018-0252-1 [DOI] [PMC free article] [PubMed] [Google Scholar]
40. Michel JB, Yeh PJ, Chait R, Moellering RC, Kishony R. Drug interactions modulate the potential for evolution of resistance. Proceedings of the National Academy of Sciences. 2008;105(39):14918–14923. 10.1073/pnas.0800944105 [DOI] [PMC free article] [PubMed] [Google Scholar]
41. Hegreness M, Shoresh N, Damian D, Hartl D, Kishony R. Accelerated evolution of resistance in multidrug environments. Proceedings of the National Academy of Sciences. 2008;105(37):13977–13981. 10.1073/pnas.0805965105 [DOI] [PMC free article] [PubMed] [Google Scholar]
42. Pena-Miller R, Laehnemann D, Jansen G, Fuentes-Hernandez A, Rosenstiel P, Schulenburg H, et al. When the most potent combination of antibiotics selects for the greatest bacterial load: the smile-frown transition. PLoS biology. 2013;11(4):e1001540 10.1371/journal.pbio.1001540 [DOI] [PMC free article] [PubMed] [Google Scholar]
43. Chait R, Craney A, Kishony R. Antibiotic interactions that select against resistance. Nature. 2007;446(7136):668 10.1038/nature05685 [DOI] [PubMed] [Google Scholar]
44. Torella JP, Chait R, Kishony R. Optimal drug synergy in antimicrobial treatments. PLoS computational biology. 2010;6(6):e1000796 10.1371/journal.pcbi.1000796 [DOI] [PMC free article] [PubMed] [Google Scholar]
45. Munck C, Gumpert HK, Wallin AIN, Wang HH, Sommer MO. Prediction of resistance development against drug combinations by collateral responses to component drugs. Science translational medicine. 2014;6(262):262ra156–262ra156. 10.1126/scitranslmed.3009940 [DOI] [PMC free article] [PubMed] [Google Scholar]
46. Beganovic M, Luther MK, Rice LB, Arias CA, Rybak MJ, LaPlante KL. A review of combination antimicrobial therapy for Enterococcus faecalis bloodstream infections and infective endocarditis. Clinical Infectious Diseases. 2018;67(2):303–309. 10.1093/cid/ciy064 [DOI] [PMC free article] [PubMed] [Google Scholar]
47.Clewell DB, Gilmore MS, Ike Y, Shankar N. Enterococci: from commensals to leading causes of drug resistant infection. Massachusetts Eye and Ear Infirmary; 2014. [PubMed]
48. Baddour LM, Wilson WR, Bayer AS, Fowler VG Jr, Tleyjeh IM, Rybak MJ, et al. Infective endocarditis in adults: diagnosis, antimicrobial therapy, and management of complications: a scientific statement for healthcare professionals from the American Heart Association. Circulation. 2015;132(15):1435–1486. 10.1161/CIR.0000000000000296 [DOI] [PubMed] [Google Scholar]
49. Chirouze C, Athan E, Alla F, Chu VH, Corey GR, Selton-Suty C, et al. Enterococcal endocarditis in the beginning of the 21st century: analysis from the International Collaboration on Endocarditis-Prospective Cohort Study. Clinical microbiology and infection. 2013;19(12):1140–1147. 10.1111/1469-0691.12166 [DOI] [PubMed] [Google Scholar]
50. Huycke MM, Sahm DF, Gilmore MS. Multiple-drug resistant enterococci: the nature of the problem and an agenda for the future. Emerging infectious diseases. 1998;4(2):239 10.3201/eid0402.980211 [DOI] [PMC free article] [PubMed] [Google Scholar]
51. Hancock LE, Gilmore MS. Pathogenicity of enterococci In: Gram-Positive Pathogens, Second Edition American Society of Microbiology; 2006. p. 299–311. [Google Scholar]
52. Fernández-Hidalgo N, Almirante B, Gavaldà J, Gurgui M, Peña C, De Alarcón A, et al. Ampicillin plus ceftriaxone is as effective as ampicillin plus gentamicin for treating Enterococcus faecalis infective endocarditis. Clinical infectious diseases. 2013;56(9):1261–1268. 10.1093/cid/cit052 [DOI] [PubMed] [Google Scholar]
53. Gavalda J, Len O, Miró JM, Munoz P, Montejo M, Alarcón A, et al. Brief communication: treatment of Enterococcus faecalis endocarditis with ampicillin plus ceftriaxone. Annals of internal medicine. 2007;146(8):574–579. 10.7326/0003-4819-146-8-200704170-00008 [DOI] [PubMed] [Google Scholar]
54. Loewe S. The problem of synergism and antagonism of combined drugs. Arzneimittelforschung. 1953;3:285–290. [PubMed] [Google Scholar]
55. Kristich CJ, Rice LB, Arias CA. Enterococcal infection—treatment and antibiotic resistance In: Enterococci: From commensals to leading causes of drug resistant infection [Internet]. Massachusetts Eye and Ear Infirmary; 2014. [PubMed] [Google Scholar]
56. Tripodi M, Locatelli A, Adinolfi L, Andreana A, Utili R. Successful treatment with ampicillin and fluoroquinolones of human endocarditis due to high-level gentamicin-resistant enterococci. European Journal of Clinical Microbiology and Infectious Diseases. 1998;17(10):734–736. 10.1007/s100960050171 [DOI] [PubMed] [Google Scholar]
57. Holmberg A, Mörgelin M, Rasmussen M. Effectiveness of ciprofloxacin or linezolid in combination with rifampicin against Enterococcus faecalis in biofilms. Journal of antimicrobial chemotherapy. 2011;67(2):433–439. 10.1093/jac/dkr477 [DOI] [PubMed] [Google Scholar]
58. Arias CA, Contreras GA, Murray BE. Management of multidrug-resistant enterococcal infections. Clinical microbiology and infection. 2010;16(6):555–562. 10.1111/j.1469-0691.2010.03214.x [DOI] [PMC free article] [PubMed] [Google Scholar]
59. Tang HJ, Chen CC, Zhang CC, Su BA, Li CM, Weng TC, et al. In vitro efficacy of fosfomycin-based combinations against clinical vancomycin-resistant Enterococcus isolates. Diagnostic microbiology and infectious disease. 2013;77(3):254–257. 10.1016/j.diagmicrobio.2013.07.012 [DOI] [PubMed] [Google Scholar]
60. Silvestri C, Cirioni O, Arzeni D, Ghiselli R, Simonetti O, Orlando F, et al. In vitro activity and in vivo efficacy of tigecycline alone and in combination with daptomycin and rifampin against Gram-positive cocci isolated from surgical wound infection. European journal of clinical microbiology & infectious diseases. 2012;31(8):1759–1764. 10.1007/s10096-011-1498-1 [DOI] [PubMed] [Google Scholar]
61. Bollenbach T, Quan S, Chait R, Kishony R. Nonoptimal microbial response to antibiotics underlies suppressive drug interactions. Cell. 2009;139(4):707–718. 10.1016/j.cell.2009.10.025 [DOI] [PMC free article] [PubMed] [Google Scholar]
62. Yu W, Hallinen KM, Wood KB. Interplay between antibiotic efficacy and drug-induced lysis underlies enhanced biofilm formation at subinhibitory drug concentrations. Antimicrobial agents and chemotherapy. 2018;62(1):e01603–17. 10.1128/AAC.01603-17 [DOI] [PMC free article] [PubMed] [Google Scholar]
63.Santos-Lopez A, Marshall CW, Scribner MR, Snyder D, Cooper VS. Biofilm-dependent evolutionary pathways to antibiotic resistance. bioRxiv. 2019; p. 581611. [DOI] [PMC free article] [PubMed]
64. Martin M, Hölscher T, Dragoš A, Cooper VS, Kovács ÁT. Laboratory evolution of microbial interactions in bacterial biofilms. Journal of bacteriology. 2016;198(19):2564–2571. 10.1128/JB.01018-15 [DOI] [PMC free article] [PubMed] [Google Scholar]
65. Reller LB, Weinstein M, Jorgensen JH, Ferraro MJ. Antimicrobial susceptibility testing: a review of general principles and contemporary practices. Clinical infectious diseases. 2009;49(11):1749–1755. 10.1086/647952 [DOI] [PubMed] [Google Scholar]
66.Dean Z, Maltas J, Wood KB. Antibiotic interactions shape short-term evolution of resistance in Enterococcus faecalis, Dryad, Dataset; 2020. Available from: 10.5061/dryad.j3tx95x92. [DOI] [PMC free article] [PubMed]

PLoS Pathog. doi: 10.1371/journal.ppat.1008278.r001

Decision Letter 0

François Balloux, Michael Otto

Transfer Alert

This paper was transferred from another journal. As a result, its full editorial history (including decision letters, peer reviews and author responses) may not be present.

5 Aug 2019

Dear Dr. Wood,

Thank you very much for submitting your manuscript "Antibiotic interactions shape short-term evolution of resistance in Enterococcus faecalis" (PPATHOGENS-D-19-01118) for review by PLOS Pathogens. Your manuscript was fully evaluated at the editorial level and by independent peer reviewers. The reviewers appreciated the attention to an important problem, but raised some substantial concerns about the manuscript as it currently stands. These issues must be addressed before we would be willing to consider a revised version of your study. We cannot, of course, promise publication at that time.

We therefore ask you to modify the manuscript according to the review recommendations before we can consider your manuscript for acceptance. Your revisions should address the specific points made by each reviewer.

In addition, when you are ready to resubmit, please be prepared to provide the following:

(1) A letter containing a detailed list of your responses to the review comments and a description of the changes you have made in the manuscript. Please note while forming your response, if your article is accepted, you may have the opportunity to make the peer review history publicly available. The record will include editor decision letters (with reviews) and your responses to reviewer comments. If eligible, we will contact you to opt in or out.

(2) Two versions of the manuscript: one with either highlights or tracked changes denoting where the text has been changed; the other a clean version (uploaded as the manuscript file).

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email us at figures@plos.org.

Additionally, to enhance the reproducibility of your results, PLOS recommends that you deposit your laboratory protocols in protocols.io, where a protocol can be assigned its own identifier (DOI) such that it can be cited independently in the future. For instructions see http://journals.plos.org/plospathogens/s/submission-guidelines#loc-materials-and-methods

We hope to receive your revised manuscript within 60 days. If you anticipate any delay in its return, we ask that you let us know the expected resubmission date by replying to this email. Revised manuscripts received beyond 60 days may require evaluation and peer review similar to that applied to newly submitted manuscripts.

[LINK]

We are sorry that we cannot be more positive about your manuscript at this stage, but if you have any concerns or questions, please do not hesitate to contact us.

Sincerely,

Michael Otto

Associate Editor

PLOS Pathogens

François Balloux

Section Editor

PLOS Pathogens

Kasturi Haldar

Editor-in-Chief

PLOS Pathogens

orcid.org/0000-0001-5065-158X

Grant McFadden

Editor-in-Chief

PLOS Pathogens

orcid.org/0000-0002-2556-3526

***********************

Reviewer's Responses to Questions

Part I - Summary

Please use this section to discuss strengths/weaknesses of study, novelty/significance, general execution and scholarship.

Reviewer #1: This is a potentially interesting study of resistance evolution to drug combinations in E. faecalis. The authors acquired growth curves of serially passaged E. faecalis in the presence of four different drug combinations, and within each combination examined each single drug and then 2 or more mixtures of the two drugs, all expected to have the same initial impact on growth rate. They conclude that adaptation trajectories are different for different dosing combinations and that this can be explained by rescaling the drug concentration to an effective concentration, by using final resistant profiles, and predicting growth rate from ancestral two drug growth response surface.

Reviewer #2: Dean et al present a compelling study of how different antibiotic combination treatments affect the subsequent phenotypes evolved under such treatment. The choice of combinations is clever, demonstrating examples of synergistic and antagonistic stresses between the antibiotics. Most interesting is the tagacycline/ciprofloxacin combination that has a critical value of tagacycline concentration when there is a sudden loss of cipro resistance. Subsequently, they come up with a rescaling to normalize the various cultures to simulated constant growth. The result is quite striking: bi-antibiotic resistance profiles are essentially linear. This is a wonderful result, showing that the phenomenological result of experimental evolution follows a quite simple rule!

I have the following comments:

p. 6, Fig 1c: I am trying to understand how the adaptation rate can differ but the resistance is almost identical between conditions. The authors note that this is a surprising result, but I would appreciate slightly more exposition on why I should not feel so naive being surprised about it.

p. 6, note on intro paragraph to section titled: “Aminoglycoside/β-lactam and β-lactam/fluoroquinolone combinations slow growth adaptation and select for resistant profiles distinct from those evolved to the component drugs.” - some may note that in most papers this section would belong in the introduction, but I believe that it is well-suited here because there are multiple antibiotic combinations done, so this reminds the reader of the importance of this particular combination.

Rescaling - I personally find this to be compelling and interesting, but I fear that some in the pathogens community will not appreciate how important it is, based on my prior experience. My hope is that somewhat challenging concepts from other fields can be allowed to stand.

Reviewer #3: In this manuscript the authors study the effect of drug combinations on the evolution of resistance. They use test cases of different drug interactions to show how it drives the resistance evolution. They use the picture of geometric rescaling to explain the results of the evolution experiments. To my knowledge, this kind of detailed experiments and analysis is innovative and can help with the optimization problem of effectiveness vs resistance. Following some revisions/clarifications, I think the manuscript will be more than adequate for publication.

**********

Part II – Major Issues: Key Experiments Required for Acceptance

Please use this section to detail the key new experiments or modifications of existing experiments that should be absolutely required to validate study conclusions.

Generally, there should be no more than 3 such required experiments or major modifications for a "Major Revision" recommendation. If more than 3 experiments are necessary to validate the study conclusions, then you are encouraged to recommend "Reject".

Reviewer #1: 1) A critical missing piece to the manuscript is a description of how exactly the per capita growth rate is measured. Almost all results and analyses depend critically on this value, and there are several places where results appear to disagree with each other

In the methods it is stated that the early exponential phase was fit to an exponential function and then normalized to ancestral cells

- The growth curves of those ancestral cells are needed

Are all data normalized to one set of controls? Are they normalized to matched controls done alongside each experiment?

Why in Fig 1 and 2 do the evolved populations struggle to meet a growth rate of ‘1’, despite acquiring resistance, whereas in Fig 3 combination D exceeds 1 – did this condition really evolve an E. faecalis strain that grows faster than wildtype?

- A more precise description than “typically OD<0.4” is needed to describe how the early exponential phase of the growth curve was defined and how lag phase was avoided

Presumably the growth rate is related to the slope of the exponential portion of the OD600 over time curve. If so:

Fig 1a condition D the D0 and D3 slopes look identical, and yet the ‘growth rate’ is very different when plotted in Fig1b.

Fig 2a condition A, the D0 slope looks steeper than the D2 slope, and yet the ‘growth rate’ in Fig 2b is close to 0 for D0 and at 0.5 for D2

2) It is difficult to assess if the differences are real and significant

-Each drug combination appears to have been examined in one serial passaging experiment, sometimes for two days, sometimes for three days. Would the trajectories and resistance evolution be reproducible in a second experiment?

-the adaptation rate is measured by a linear fit to very non-linear data. For many conditions it seems clear that the growth rate increases and then plateaus, so the linear fit reduces the adaptation rate. Would a better measure be time to maximum growth rate?

-Resistance is reported as the change in IC50, and often the level of resistance is fairly low (a two to three fold change), and several conclusions rest on “differences” in the time to resistance or on the degree of resistance acquired. However; it seems a typical MIC assay was used to assess resistance where cultures are grown overnight in presence of drug and then OD600 measured. This usually leads to stationary cultures below the inhibitory concentrations, and therefore a steep delineation between growth and no growth, with often 2-fold differences in MIC values for day to day assay variation for a given strain and antibiotic. This makes me skeptical that small observed changes in the IC50 are significant. The dose-response curves used to calculate IC50 could also be given as supplemental figures.

Perhaps a better dynamic range and more reliable IC50 would be obtained by measuring the growth rate of the evolved populations to a series of dilutions of the single drugs, this measure might also better align with the other data.

- The combination where these issues are most clearly seen is CRO-AMP. The authors state that for the experiment in Fig 1 “the day 2 curves vary substantially between replicates and conditions”. The day 3 curves in Fig 1a all look fairly similar to my eye, especially when compared to the other figures in the paper. Secondly they state that the “adaptation is significantly faster for the two combinations (B and C) than for the single drug treatments (A and B)”. Was a test used to assess significance? If so which one?

Would the data be reproducible if the experiment was repeated again on a different day?

In Fig 3, condition A is the same as in Fig 1, and the growth rate curve in panel b is substantially different between days (In Fig 1 it plateaus at 0.5-0.6 on day 1, in Fig 2 it plateaus at 1 on day 2). The overall adaptation rate is similar, both close to 0.2, and so is the level of resistance measured; however the difference between the adaptation rate between experiments for condition A in Fig 1 and 3 looks similar to the difference in adaptation rate within experiment between conditions A and C in Fig 1.

3) A major conclusion of the paper is that rescaling concentrations based on resistance profiles predicts the growth rate in the presence of the combination. This is an attractive conclusion, most robust for the TGC-CIP combination. In this case it looks as though the predicted growth rate actually maps very closely to the achieved growth rates in Fig S4. Would be useful to plot the predicted versus actual to drive this point home?

However based on points listed above and below the specificity of predictions and conclusions is hard to agree with for several of the other combinations

- For combination CRO-AMP why is the predicted growth rate in Fig 5 a better than 1? Is a difference between a growth rate of 1 and 1.1 really real and significant? My conclusion would be much more general that concentration rescaling predicts in all of the combinations the adapted populations should have near wildtype growth, and that matches what was observed.

- For AMP-STR and CRO-CIP the authors claim the resistant profiles fall on a line segment.

For AMP-STR the resistant changes are so small that it is hard to agree with this and for CRO-CIP it actually looks more like a curve, which is what I would expect from the curved antagonistic growth contours in Fig 3 a.

The line segment analysis is unsatisfying. What was used to determine the end points X and Y of the line segments? Why do they not lie at the x and y intercepts of the dotted lines through 0 resistance, or why do they not lie at the points of A and D, the single drug conditions?

Reviewer #2: No new experiments required.

Reviewer #3: To my opinion there are no new experiments required, but part the analysis must be revised/explained.

Looking at Figure 1a (right) the OD curves (D3) of A and D conditions looks different (it looks like in condition A the adaptation is faster), but A and D has the same adaptation rate in Figure 1b.

Similar issue is with conditions A and D in Figure 2 - it looks like the adaptation is faster condition D, but the adaptation rate are equal in Figure 2b. Moreover, it looks like the growth on day zero are not equal for all conditions.

I guess this due to the criterion for the time interval of "early exponential phase". You mentioned that it is usually OD<0.4, please elaborate more on the exact definition and the its motivation. How does the results of the paper depend on the definition? Does the main results of the paper hold with a different definition?

You mentioned this issue as a limitation: "It is clear that all growth curves are not purely exponential, and in fact the per capita growth rate can change with time. Our growth rate estimates should therefore be thought of as an effective growth rate that reduces the population dynamics each day to a single number. "

It is great that you are aware of this, but the current figures are confusing for the reader because of this issue. Additionally to a more detailed explanation of this issue in the methods section, please add the time interval that was used to estimate the growth rate in the OD figures (figure 1a,2a,...).

**********

Part III – Minor Issues: Editorial and Data Presentation Modifications

Please use this section for editorial suggestions as well as relatively minor modifications of existing data that would enhance clarity.

Reviewer #1: 4) There are several mistakes in the manuscript, pg 3 bottom line the authors refer to day 2 curves in Fig 1 but in Fig1a the curves are labeled D3. In Fig 1b the average rate of growth adaptation point for D has tiny error bars on it, but I do not see error bars on any other such point. On Pg 5 the authors refer to single-drug conditions (A and B), but presumably mean A and D, and this is repeated on pg 6

5) Two of the antagonistic combinations examined, AMP-STR, and CRO-CIP, include drugs that the E. faecalis strain used is already resistant to. This makes the concentrations used to study streptomycin and ceftriaxone relatively high, and lessens the relevance of the study

6) In Fig 2 parts a and b show results up to day 2, but part c resistance profiles are shown up to day 3. The data should be consistent

7) In the discussion the authors speculate that OD measurements at shorter intervals would allow better growth rate estimates – it is hard to imagine that acquiring more frequently than once in 20 minutes would make a drastic difference.

8) The growth rate curve in Fig 1b for condition D is surprising given the population is already resistant to AMP by day 1 according to Fig 1c. I suspect something went awry experimentally or with the analysis of day 1 growth rate.

Reviewer #2: See summary in Part I.

Reviewer #3: In Figure 1a, the legend is D3, in the text you mentioned day 2.

Are the OD curves in log scale? please give some indication on the figure.

Why the D0 curves doesn't start at t=0?

**********

PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Reviewer #2: Yes: Christian Ray

Reviewer #3: No

PLoS Pathog. 2020 Mar 2;16(3):e1008278. doi: 10.1371/journal.ppat.1008278.r002

Author response to Decision Letter 0

5 Nov 2019

Attachment

Submitted filename: ResponseToReviews_Dean2019.pdf

Click here for additional data file.^{(144.9KB, pdf)}

PLoS Pathog. doi: 10.1371/journal.ppat.1008278.r003

Decision Letter 1

Michael Otto

11 Dec 2019

Dear Dr. Wood,

We are pleased to inform that your manuscript, "Antibiotic interactions shape short-term evolution of resistance in Enterococcus faecalis", has been editorially accepted for publication at PLOS Pathogens.

Before your manuscript can be formally accepted and sent to production, you will need to complete our formatting changes, which you will receive by email within a week. Please note that your manuscript will not be scheduled for publication until you have made the required changes.

IMPORTANT NOTES

(1) Please note, once your paper is accepted, an uncorrected proof of your manuscript will be published online ahead of the final version, unless you’ve already opted out via the online submission form. If, for any reason, you do not want an earlier version of your manuscript published online or are unsure if you have already indicated as such, please let the journal staff know immediately at plospathogens@plos.org.

(2) Copyediting and Proofreading: The corresponding author will receive a typeset proof for review, to ensure errors have not been introduced during production. Please review the PDF proof of your manuscript carefully, as this is the last chance to correct any errors. Please note that major changes, or those which affect the scientific understanding of the work, will likely cause delays to the publication date of your manuscript.

(3) Appropriate Figure Files: Please remove all name and figure # text from your figure files. Please also take this time to check that your figures are of high resolution, which will improve the readbility of your figures and help expedite your manuscript's publication. Please note that figures must have been originally created at 300dpi or higher. Do not manually increase the resolution of your files. For instructions on how to properly obtain high quality images, please review our Figure Guidelines, with examples at: http://journals.plos.org/plospathogens/s/figures.

(4) Striking Image: Please upload a striking still image to accompany your article if one is available (you can include a new image or an existing one from within your manuscript). Should your paper be accepted, this image will be considered for our monthly issue image and may also appear on our website to feature your article. Please upload this as a separate file, selecting "striking image" as the file type upon upload. Please also include a separate "Other" file with a caption, including credits and any potential copyright information. Please do not include the caption in the main article file. If your image is from someone other than yourself, please ensure that the artist has read and agreed to the terms and conditions of the Creative Commons Attribution License at http://journals.plos.org/plospathogens/s/content-license. Please note that PLOS cannot publish copyrighted images.

(5) Press Release or Related Media: If your institution or institutions have a press office, please notify them about your upcoming paper at this point, to enable them to help maximize its impact. If they will be preparing press materials for this manuscript, please inform our press team in advance at plospathogens@plos.org as soon as possible. We ask that you contact us within one week to plan ahead of our fast Production schedule. If you need to know your paper's publication date for related media purposes, you must coordinate with our press team, and your manuscript will remain under a strict press embargo until the publication date and time. This means an early version of your manuscript will not be published ahead of your final version.

(6) PLOS requires an ORCID iD for all corresponding authors on papers submitted after December 6th, 2016. Please ensure that you have an ORCID iD and that it is validated in Editorial Manager. To do this, go to ‘Update my Information’ (in the upper left-hand corner of the main menu), and click on the Fetch/Validate link next to the ORCID field. This will take you to the ORCID site and allow you to create a new iD or authenticate a pre-existing iD in Editorial Manager

(7) Update your Profile Information: Now that your manuscript has been provisionally accepted, please log into Editorial Manager and update your profile, if needed. Go to https://www.editorialmanager.com/ppathogens, log in, and click on the "Update My Information" link at the top of the page. Please update your user information to ensure an efficient production and billing process.

(8) LaTeX users only: Our staff will ask you to upload a TEX file in addition to the PDF before the paper can be sent to typesetting, so please carefully review our Latex Guidelines http://journals.plos.org/plospathogens/s/latex in the meantime.

(9) If you have associated protocols in protocols.io, please ensure that you make them public before publication to guarantee immediate access to the methodological details.

Thank you again for supporting Open Access publishing; we are looking forward to publishing your work in PLOS Pathogens.

Best regards,

Michael Otto

Section Editor

PLOS Pathogens

François Balloux

Section Editor

PLOS Pathogens

Kasturi Haldar

Editor-in-Chief

PLOS Pathogens

orcid.org/0000-0001-5065-158X

Grant McFadden

Editor-in-Chief

PLOS Pathogens

orcid.org/0000-0002-2556-3526

***********************************************************

Reviewer Comments (if any, and for reference):

Reviewer's Responses to Questions

Part I - Summary

Please use this section to discuss strengths/weaknesses of study, novelty/significance, general execution and scholarship.

Reviewer #1: This is a revised and improved version of the manuscript on the adaptation of E. faecalis to four different combinations of antibiotics. Their conclusions remain that within a drug pair, adaptation trajectories differ depending on the dosing combination of the component drugs, and that this can be explained by using resistance profiles to rescale concentrations. The authors have changed and performed several extra analyses, corrected errors, and provided more information. These changes improve the manuscript. The more interesting part of the study lies in the results and discussion on how resistance measures used to rescale the growth-response curve of a particular combination correlate with experimental outcome.

Method for calculating that growth-rate response is still not really clear and very difficult to understant. After considerable puzzling over the figures and methods, I now take it that essentially a straight line was fitted through the OD curves (ie exponential function with the y-axis on a log scale) in the range of OD 0.1 – 0.4 across the whole time scale. This combines the aspects of exponential growth rate and saturation density into one number to approximate the inhibitory effect of the drugs on the population.

There are still several issues that are confusing:

1) Why do all the growth rates on day 0 start essentially at 0?

The extent of inhibition from the drug conditions on D0 vary between the four experiments, and yet all of them seem to have a growth rate very near 0 in the growth rate over time curves in the part b of each figure.

To illustrate this, Fig 2, condition A, Amp alone. In Fig 2a, the growth curve for D0 seems to increase rapidly to OD 0.4 then plateau – surely even with the plateau blunting the “growth rate” or slope of the exponential fit, and even with normalizing to ancestral growth rate, the growth rate should still be above 0?

Also in Fig 1a, the location of the circles in the heatmap indicate that a contour of constant inhibition of approximately 0.5 was chosen, however, in part b the growth rates of D0 are all close to 0.

2) The growth curves in Fig S16 indicate that in Fig 2 for condition A the growth rate should decrease from Day 0 to Day 1, but this is not reflected in the graphs in panel b.

3) The authors have very helpfully included Fig S2 and Fig S3, examining if changing the OD window for calculating growth rate impacts the conclusions of the CRO-AMP combination. However, it does not seem that the differences between single drugs and drug combinations remain significant with the changed OD window. Specifically, the

legend of Fig S2 states that in all four methods a t.test shows significance comparing single drug conditions to drug combinations. For the panels on the right, examining OD [0.05,0.3] and OD [0.15,0.5] do not seem different.

Reviewer #2: The authors have gone to great lengths to address the first round of reviews. The strengths of the study largely remain intact, while at the same time the authors laudably discuss the weaknesses and limitations in scope of the study design. The combination of experimental evolution and sensitivity profiles to multiple drugs is significant both as a basic study of pathogen evolution that may emulate some in-host conditions and as a conceptual approach to see how multiple selective pressures change bacterial phenotypes.

Reviewer #3: In this manuscript, the authors study the effect of drug combinations on the evolution of

resistance. They use test cases of different drug interactions to show how it drives the resistance evolution. They use the picture of geometric rescaling to explain the results of the evolution experiments. To my knowledge, this kind of detailed experiments and analysis is innovative and can help with the optimization problem of effectiveness vs resistance. The author made a significant improvement in the MS but one issue is still unclear to me.

**********

Part II – Major Issues: Key Experiments Required for Acceptance

Please use this section to detail the key new experiments or modifications of existing experiments that should be absolutely required to validate study conclusions.

There are still several issues that are confusing:

1) Why do all the growth rates on day 0 start essentially at 0?

2) The growth curves in Fig S16 indicate that in Fig 2 for condition A the growth rate should decrease from Day 0 to Day 1, but this is not reflected in the graphs in panel b.

Reviewer #2: None at this stage.

Reviewer #3: (No Response)

**********

Part III – Minor Issues: Editorial and Data Presentation Modifications

Please use this section for editorial suggestions as well as relatively minor modifications of existing data that would enhance clarity.

There are still several issues that are confusing:

1) Why do all the growth rates on day 0 start essentially at 0?

2) The growth curves in Fig S16 indicate that in Fig 2 for condition A the growth rate should decrease from Day 0 to Day 1, but this is not reflected in the graphs in panel b.

Reviewer #2: None.

Reviewer #3: I still do not understand how did the authors evaluated the growth rate in a particular case. They mentioned OD in the range 0.1-0.4. I don’t understand what is the time interval for the fit in the case that the OD does not reach 0.4, is it 10 hours? What about the cases in which the OD reaches value larger than 0.4 and then declines? I am not sure that the definition of 0.1 to 0.4 is a good definition, because in the case of reaching saturation before 0.4 the effective growth rate depends on the time interval after saturation. A better definition could be the slope in the range where the log(OD) is linear. I leave it to author to decide whether to change the definition, but the clarification of the time interval (not just OD interval) is needed. Please provide the fit to exponent for some cases in the SI.

**********

PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Reviewer #2: No

Reviewer #3: No

PLoS Pathog. doi: 10.1371/journal.ppat.1008278.r004

Acceptance letter

Michael Otto

20 Feb 2020

Dear Dr. Wood,

We are delighted to inform you that your manuscript, "Antibiotic interactions shape short-term evolution of resistance in Enterococcus faecalis," has been formally accepted for publication in PLOS Pathogens.

We have now passed your article onto the PLOS Production Department who will complete the rest of the pre-publication process. All authors will receive a confirmation email upon publication.

The corresponding author will soon be receiving a typeset proof for review, to ensure errors have not been introduced during production. Please review the PDF proof of your manuscript carefully, as this is the last chance to correct any scientific or type-setting errors. Please note that major changes, or those which affect the scientific understanding of the work, will likely cause delays to the publication date of your manuscript. Note: Proofs for Front Matter articles (Pearls, Reviews, Opinions, etc...) are generated on a different schedule and may not be made available as quickly.

Soon after your final files are uploaded, the early version of your manuscript, if you opted to have an early version of your article, will be published online. The date of the early version will be your article's publication date. The final article will be published to the same URL, and all versions of the paper will be accessible to readers.

Thank you again for supporting open-access publishing; we are looking forward to publishing your work in PLOS Pathogens.

Best regards,

Kasturi Haldar

Editor-in-Chief

PLOS Pathogens

orcid.org/0000-0001-5065-158X

Michael Malim

Editor-in-Chief

PLOS Pathogens

orcid.org/0000-0002-7699-2064

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

S1 Fig. Growth curves for adaptation in AMP-CRO combinations.

(EPS)

Click here for additional data file.^{(1.1MB, eps)}

S2 Fig. Adaptation rate trends do not depend sensitively on choice of OD windows for growth rate fits.

(EPS)

Click here for additional data file.^{(371.3KB, eps)}

S3 Fig. Growth rate time series do not depend sensitively on choice of OD windows for growth rate fits.

(EPS)

Click here for additional data file.^{(553.8KB, eps)}

S4 Fig. Growth rate time series and fits to linear and saturating functions.

(EPS)

Click here for additional data file.^{(759KB, eps)}

S5 Fig. Adaptation rate trends do not depend sensitively on linear or saturating fits to growth rate time series.

(EPS)

Click here for additional data file.^{(166.3KB, eps)}

S6 Fig. Median growth curves within each condition capture qualitative trends of growth rate adaptation.

(EPS)

Click here for additional data file.^{(183.4KB, eps)}

S7 Fig. Dose response curves to estimate AMP resistance following CRO adaptation.

(EPS)

Click here for additional data file.^{(809.7KB, eps)}

S8 Fig. Dose response curves to estimate AMP resistance following AMP-CRO adaptation.

(EPS)

Click here for additional data file.^{(771.2KB, eps)}

S9 Fig. Dose response curves to estimate AMP resistance following AMP-CRO adaptation.

(EPS)

Click here for additional data file.^{(708.4KB, eps)}

S10 Fig. Dose response curves to estimate AMP resistance following AMP adaptation.

(EPS)

Click here for additional data file.^{(798KB, eps)}

S11 Fig. Dose response curves to estimate CRO resistance following CRO adaptation.

(EPS)

Click here for additional data file.^{(801.8KB, eps)}

S12 Fig. Dose response curves to estimate CRO resistance following AMP-CRO adaptation.

(EPS)

Click here for additional data file.^{(747.9KB, eps)}

S13 Fig. Dose response curves to estimate CRO resistance following AMP-CRO adaptation.

(EPS)

Click here for additional data file.^{(714.7KB, eps)}

S14 Fig. Dose response curves to estimate CRO resistance following AMP adaptation.

(EPS)

Click here for additional data file.^{(787KB, eps)}

S15 Fig. Relative error in resistance measurements for AMP and CRO.

(EPS)

Click here for additional data file.^{(52.5KB, eps)}

S16 Fig. Growth curves for adaptation in AMP-STR combinations.

(EPS)

Click here for additional data file.^{(971.5KB, eps)}

S17 Fig. Growth rate time series for adaptation in AMP-STR combinations.

(EPS)

Click here for additional data file.^{(344.2KB, eps)}

S18 Fig. Growth curves for adaptation in CRO-CIP combinations.

(EPS)

Click here for additional data file.^{(1.1MB, eps)}

S19 Fig. Growth rate time series for adaptation in CRO-CIP combinations.

(EPS)

Click here for additional data file.^{(434.8KB, eps)}

S20 Fig. Growth curves for adaptation in TGC-CIP combinations.

(EPS)

Click here for additional data file.^{(843.2KB, eps)}

S21 Fig. Growth rate time series for adaptation in TGC-CIP combinations.

Growth rate time series for each population (light lines) and the mean across populations for a given condition (A-K) for combinations of tigecycline (TGC) and ciprofloxacin (CIP).

(EPS)

Click here for additional data file.^{(377.4KB, eps)}

S22 Fig. Example exponential fits to growth curves.

(EPS)

Click here for additional data file.^{(585.6KB, eps)}

Attachment

Submitted filename: ResponseToReviews_Dean2019.pdf

Click here for additional data file.^{(144.9KB, pdf)}

Data Availability Statement

We have posted all data to Data Dryad (https://datadryad.org/) where it is freely available without restriction.

[ppat.1008278.ref001] 1. Davies J, Davies D. Origins and evolution of antibiotic resistance. Microbiol Mol Biol Rev. 2010;74(3):417–433. 10.1128/MMBR.00016-10 [DOI] [PMC free article] [PubMed] [Google Scholar]

[ppat.1008278.ref002] 2. Levy SB, Marshall B. Antibacterial resistance worldwide: causes, challenges and responses. Nature medicine. 2004;10(12s):S122 10.1038/nm1145 [DOI] [PubMed] [Google Scholar]

[ppat.1008278.ref003] 3. Read AF, Day T, Huijben S. The evolution of drug resistance and the curious orthodoxy of aggressive chemotherapy. Proceedings of the National Academy of Sciences. 2011;108(Supplement 2):10871–10877. 10.1073/pnas.1100299108 [DOI] [PMC free article] [PubMed] [Google Scholar]

[ppat.1008278.ref004] 4. Hansen E, Woods RJ, Read AF. How to use a chemotherapeutic agent when resistance to it threatens the patient. PLoS biology. 2017;15(2):e2001110 10.1371/journal.pbio.2001110 [DOI] [PMC free article] [PubMed] [Google Scholar]

[ppat.1008278.ref005] 5. Meredith HR, Srimani JK, Lee AJ, Lopatkin AJ, You L. Collective antibiotic tolerance: mechanisms, dynamics and intervention. Nature chemical biology. 2015;11(3):182–188. 10.1038/nchembio.1754 [DOI] [PMC free article] [PubMed] [Google Scholar]

[ppat.1008278.ref006] 6. Sorg RA, Lin L, Van Doorn GS, Sorg M, Olson J, Nizet V, et al. Collective resistance in microbial communities by intracellular antibiotic deactivation. PLoS biology. 2016;14(12):e2000631 10.1371/journal.pbio.2000631 [DOI] [PMC free article] [PubMed] [Google Scholar]

[ppat.1008278.ref007] 7. Tan C, Smith RP, Srimani JK, Riccione KA, Prasada S, Kuehn M, et al. The inoculum effect and band-pass bacterial response to periodic antibiotic treatment. Molecular systems biology. 2012;8(1):617 10.1038/msb.2012.49 [DOI] [PMC free article] [PubMed] [Google Scholar]

[ppat.1008278.ref008] 8. Karslake J, Maltas J, Brumm P, Wood KB. Population density modulates drug inhibition and gives rise to potential bistability of treatment outcomes for bacterial infections. PLoS computational biology. 2016;12(10):e1005098 10.1371/journal.pcbi.1005098 [DOI] [PMC free article] [PubMed] [Google Scholar]

[ppat.1008278.ref009] 9. Gjini E, Brito PH. Integrating antimicrobial therapy with host immunity to fight drug-resistant infections: classical vs. adaptive treatment. PLoS computational biology. 2016;12(4):e1004857 10.1371/journal.pcbi.1004857 [DOI] [PMC free article] [PubMed] [Google Scholar]

[ppat.1008278.ref010] 10. Zhang Q, Lambert G, Liao D, Kim H, Robin K, Tung Ck, et al. Acceleration of emergence of bacterial antibiotic resistance in connected microenvironments. Science. 2011;333(6050):1764–1767. 10.1126/science.1208747 [DOI] [PubMed] [Google Scholar]

[ppat.1008278.ref011] 11. Baym M, Lieberman TD, Kelsic ED, Chait R, Gross R, Yelin I, et al. Spatiotemporal microbial evolution on antibiotic landscapes. Science. 2016;353(6304):1147–1151. 10.1126/science.aag0822 [DOI] [PMC free article] [PubMed] [Google Scholar]

[ppat.1008278.ref012] 12. Greulich P, Waclaw B, Allen RJ. Mutational pathway determines whether drug gradients accelerate evolution of drug-resistant cells. Physical Review Letters. 2012;109(8):088101 10.1103/PhysRevLett.109.088101 [DOI] [PubMed] [Google Scholar]

[ppat.1008278.ref013] 13. Hermsen R, Deris JB, Hwa T. On the rapidity of antibiotic resistance evolution facilitated by a concentration gradient. Proceedings of the National Academy of Sciences. 2012;109(27):10775–10780. 10.1073/pnas.1117716109 [DOI] [PMC free article] [PubMed] [Google Scholar]

[ppat.1008278.ref014] 14. Moreno-Gamez S, Hill AL, Rosenbloom DI, Petrov DA, Nowak MA, Pennings PS. Imperfect drug penetration leads to spatial monotherapy and rapid evolution of multidrug resistance. Proceedings of the National Academy of Sciences. 2015;112(22):E2874–E2883. 10.1073/pnas.1424184112 [DOI] [PMC free article] [PubMed] [Google Scholar]

[ppat.1008278.ref015] 15. De Jong MG, Wood KB. Tuning spatial profiles of selection pressure to modulate the evolution of drug resistance. Physical review letters. 2018;120(23):238102 10.1103/PhysRevLett.120.238102 [DOI] [PMC free article] [PubMed] [Google Scholar]

[ppat.1008278.ref016] 16. Trindade S, Sousa A, Xavier KB, Dionisio F, Ferreira MG, Gordo I. Positive epistasis drives the acquisition of multidrug resistance. PLoS genetics. 2009;5(7):e1000578 10.1371/journal.pgen.1000578 [DOI] [PMC free article] [PubMed] [Google Scholar]

[ppat.1008278.ref017] 17. Borrell S, Teo Y, Giardina F, Streicher EM, Klopper M, Feldmann J, et al. Epistasis between antibiotic resistance mutations drives the evolution of extensively drug-resistant tuberculosis. Evolution, medicine, and public health. 2013;2013(1):65–74. 10.1093/emph/eot003 [DOI] [PMC free article] [PubMed] [Google Scholar]

[ppat.1008278.ref018] 18. Yoshida M, Reyes SG, Tsudo S, Horinouchi T, Furusawa C, Cronin L. Time-programmable dosing allows the manipulation, suppression and reversal of antibiotic drug resistance in vitro. Nat Commun. 2017;8 10.1038/ncomms15589 [DOI] [PMC free article] [PubMed] [Google Scholar]

[ppat.1008278.ref019] 19. Meredith HR, Lopatkin AJ, Anderson DJ, You L. Bacterial temporal dynamics enable optimal design of antibiotic treatment. PLoS computational biology. 2015;11(4):e1004201 10.1371/journal.pcbi.1004201 [DOI] [PMC free article] [PubMed] [Google Scholar]

[ppat.1008278.ref020] 20. Nichol D, Jeavons P, Fletcher AG, Bonomo RA, Maini PK, Paul JL, et al. Steering evolution with sequential therapy to prevent the emergence of bacterial antibiotic resistance. PLoS computational biology. 2015;11(9):e1004493 10.1371/journal.pcbi.1004493 [DOI] [PMC free article] [PubMed] [Google Scholar]

[ppat.1008278.ref021] 21. Fuentes-Hernandez A, Plucain J, Gori F, Pena-Miller R, Reding C, Jansen G, et al. Using a sequential regimen to eliminate bacteria at sublethal antibiotic dosages. PLoS biology. 2015;13(4):e1002104 10.1371/journal.pbio.1002104 [DOI] [PMC free article] [PubMed] [Google Scholar]

[ppat.1008278.ref022] 22. Imamovic L, Sommer MOA. Use of collateral sensitivity networks to design drug cycling protocols that avoid resistance development. Sci Transl Med. 2013;5:204ra132 10.1126/scitranslmed.3006609 [DOI] [PubMed] [Google Scholar]

[ppat.1008278.ref023] 23. Kim S, Lieberman TD, Kishony R. Alternating antibiotic treatments constrain evolutionary paths to multidrug resistance. Proc Natl Acad Sci USA. 2014;111:14494–14499. 10.1073/pnas.1409800111 [DOI] [PMC free article] [PubMed] [Google Scholar]

[ppat.1008278.ref024] 24. Pál C, Papp B, Lázár V. Collateral sensitivity of antibiotic-resistant microbes. Trends in microbiology. 2015;23(7):401–407. 10.1016/j.tim.2015.02.009 [DOI] [PMC free article] [PubMed] [Google Scholar]

[ppat.1008278.ref025] 25. Barbosa C, Trebosc V, Kemmer C, Rosenstiel P, Beardmore R, Schulenburg H, et al. Alternative evolutionary paths to bacterial antibiotic resistance cause distinct collateral effects. Molecular biology and evolution. 2017;34(9):2229–2244. 10.1093/molbev/msx158 [DOI] [PMC free article] [PubMed] [Google Scholar]

[ppat.1008278.ref026] 26. Barbosa C, Beardmore R, Schulenburg H, Jansen G. Antibiotic combination efficacy (ACE) networks for a Pseudomonas aeruginosa model. PLoS biology. 2018;16(4):e2004356 10.1371/journal.pbio.2004356 [DOI] [PMC free article] [PubMed] [Google Scholar]

[ppat.1008278.ref027] 27. Rodriguez de Evgrafov M, Gumpert H, Munck C, Thomsen TT, Sommer MO. Collateral resistance and sensitivity modulate evolution of high-level resistance to drug combination treatment in Staphylococcus aureus. Molecular biology and evolution. 2015;32(5):1175–1185. 10.1093/molbev/msv006 [DOI] [PubMed] [Google Scholar]

[ppat.1008278.ref028] 28. Nichol D, Rutter J, Bryant C, Hujer AM, Lek S, Adams MD, et al. Antibiotic collateral sensitivity is contingent on the repeatability of evolution. Nature communications. 2019;10(1):334 10.1038/s41467-018-08098-6 [DOI] [PMC free article] [PubMed] [Google Scholar]

[ppat.1008278.ref029] 29.Maltas J, Wood KB. Pervasive and diverse collateral sensitivity profiles inform optimal strategies to limit antibiotic resistance. bioRxiv. 2019. [DOI] [PMC free article] [PubMed]

[ppat.1008278.ref030] 30. Podnecky NL, Fredheim EGA, Kloos J, Sorum V, Primicerio R, Roberts AP, et al. Conserved collateral antibiotic susceptibility networks in diverse clinical strains of Escherichia coli. Nature Communications. 2018;9 10.1038/s41467-018-06143-y [DOI] [PMC free article] [PubMed] [Google Scholar]

[ppat.1008278.ref031] 31. Imamovic L, Ellabaan MMH, Machado AMD, Citterio L, Wulff T, Molin S, et al. Drug-driven phenotypic convergence supports rational treatment strategies of chronic infections. Cell. 2018;172(1-2):121–134. 10.1016/j.cell.2017.12.012 [DOI] [PMC free article] [PubMed] [Google Scholar]

[ppat.1008278.ref032] 32. Baym M, Stone LK, Kishony R. Multidrug evolutionary strategies to reverse antibiotic resistance. Science. 2016;351(6268):aad3292 10.1126/science.aad3292 [DOI] [PMC free article] [PubMed] [Google Scholar]

[ppat.1008278.ref033] 33. Greco WR, Bravo G, Parsons JC. The search for synergy: a critical review from a response surface perspective. Pharmacological reviews. 1995;47(2):331–385. [PubMed] [Google Scholar]

[ppat.1008278.ref034] 34. Zhao XM, Iskar M, Zeller G, Kuhn M, Van Noort V, Bork P. Prediction of drug combinations by integrating molecular and pharmacological data. PLoS computational biology. 2011;7(12):e1002323 10.1371/journal.pcbi.1002323 [DOI] [PMC free article] [PubMed] [Google Scholar]

[ppat.1008278.ref035] 35. Wood K, Nishida S, Sontag ED, Cluzel P. Mechanism-independent method for predicting response to multidrug combinations in bacteria. Proceedings of the National Academy of Sciences. 2012;109(30):12254–12259. 10.1073/pnas.1201281109 [DOI] [PMC free article] [PubMed] [Google Scholar]

[ppat.1008278.ref036] 36. Zimmer A, Katzir I, Dekel E, Mayo AE, Alon U. Prediction of multidimensional drug dose responses based on measurements of drug pairs. Proceedings of the National Academy of Sciences. 2016;113(37):10442–10447. 10.1073/pnas.1606301113 [DOI] [PMC free article] [PubMed] [Google Scholar]

[ppat.1008278.ref037] 37. Zimmer A, Tendler A, Katzir I, Mayo A, Alon U. Prediction of drug cocktail effects when the number of measurements is limited. PLoS biology. 2017;15(10):e2002518 10.1371/journal.pbio.2002518 [DOI] [PMC free article] [PubMed] [Google Scholar]

[ppat.1008278.ref038] 38. Tendler A, Zimmer A, Mayo A, Alon U. Noise-precision tradeoff in predicting combinations of mutations and drugs. PLoS computational biology. 2019;15(5):e1006956 10.1371/journal.pcbi.1006956 [DOI] [PMC free article] [PubMed] [Google Scholar]

[ppat.1008278.ref039] 39. Russ D, Kishony R. Additivity of inhibitory effects in multidrug combinations. Nature microbiology. 2018;3(12):1339 10.1038/s41564-018-0252-1 [DOI] [PMC free article] [PubMed] [Google Scholar]

[ppat.1008278.ref040] 40. Michel JB, Yeh PJ, Chait R, Moellering RC, Kishony R. Drug interactions modulate the potential for evolution of resistance. Proceedings of the National Academy of Sciences. 2008;105(39):14918–14923. 10.1073/pnas.0800944105 [DOI] [PMC free article] [PubMed] [Google Scholar]

[ppat.1008278.ref041] 41. Hegreness M, Shoresh N, Damian D, Hartl D, Kishony R. Accelerated evolution of resistance in multidrug environments. Proceedings of the National Academy of Sciences. 2008;105(37):13977–13981. 10.1073/pnas.0805965105 [DOI] [PMC free article] [PubMed] [Google Scholar]

[ppat.1008278.ref042] 42. Pena-Miller R, Laehnemann D, Jansen G, Fuentes-Hernandez A, Rosenstiel P, Schulenburg H, et al. When the most potent combination of antibiotics selects for the greatest bacterial load: the smile-frown transition. PLoS biology. 2013;11(4):e1001540 10.1371/journal.pbio.1001540 [DOI] [PMC free article] [PubMed] [Google Scholar]

[ppat.1008278.ref043] 43. Chait R, Craney A, Kishony R. Antibiotic interactions that select against resistance. Nature. 2007;446(7136):668 10.1038/nature05685 [DOI] [PubMed] [Google Scholar]

[ppat.1008278.ref044] 44. Torella JP, Chait R, Kishony R. Optimal drug synergy in antimicrobial treatments. PLoS computational biology. 2010;6(6):e1000796 10.1371/journal.pcbi.1000796 [DOI] [PMC free article] [PubMed] [Google Scholar]

[ppat.1008278.ref045] 45. Munck C, Gumpert HK, Wallin AIN, Wang HH, Sommer MO. Prediction of resistance development against drug combinations by collateral responses to component drugs. Science translational medicine. 2014;6(262):262ra156–262ra156. 10.1126/scitranslmed.3009940 [DOI] [PMC free article] [PubMed] [Google Scholar]

[ppat.1008278.ref046] 46. Beganovic M, Luther MK, Rice LB, Arias CA, Rybak MJ, LaPlante KL. A review of combination antimicrobial therapy for Enterococcus faecalis bloodstream infections and infective endocarditis. Clinical Infectious Diseases. 2018;67(2):303–309. 10.1093/cid/ciy064 [DOI] [PMC free article] [PubMed] [Google Scholar]

[ppat.1008278.ref047] 47.Clewell DB, Gilmore MS, Ike Y, Shankar N. Enterococci: from commensals to leading causes of drug resistant infection. Massachusetts Eye and Ear Infirmary; 2014. [PubMed]

[ppat.1008278.ref048] 48. Baddour LM, Wilson WR, Bayer AS, Fowler VG Jr, Tleyjeh IM, Rybak MJ, et al. Infective endocarditis in adults: diagnosis, antimicrobial therapy, and management of complications: a scientific statement for healthcare professionals from the American Heart Association. Circulation. 2015;132(15):1435–1486. 10.1161/CIR.0000000000000296 [DOI] [PubMed] [Google Scholar]

[ppat.1008278.ref049] 49. Chirouze C, Athan E, Alla F, Chu VH, Corey GR, Selton-Suty C, et al. Enterococcal endocarditis in the beginning of the 21st century: analysis from the International Collaboration on Endocarditis-Prospective Cohort Study. Clinical microbiology and infection. 2013;19(12):1140–1147. 10.1111/1469-0691.12166 [DOI] [PubMed] [Google Scholar]

[ppat.1008278.ref050] 50. Huycke MM, Sahm DF, Gilmore MS. Multiple-drug resistant enterococci: the nature of the problem and an agenda for the future. Emerging infectious diseases. 1998;4(2):239 10.3201/eid0402.980211 [DOI] [PMC free article] [PubMed] [Google Scholar]

[ppat.1008278.ref051] 51. Hancock LE, Gilmore MS. Pathogenicity of enterococci In: Gram-Positive Pathogens, Second Edition American Society of Microbiology; 2006. p. 299–311. [Google Scholar]

[ppat.1008278.ref052] 52. Fernández-Hidalgo N, Almirante B, Gavaldà J, Gurgui M, Peña C, De Alarcón A, et al. Ampicillin plus ceftriaxone is as effective as ampicillin plus gentamicin for treating Enterococcus faecalis infective endocarditis. Clinical infectious diseases. 2013;56(9):1261–1268. 10.1093/cid/cit052 [DOI] [PubMed] [Google Scholar]

[ppat.1008278.ref053] 53. Gavalda J, Len O, Miró JM, Munoz P, Montejo M, Alarcón A, et al. Brief communication: treatment of Enterococcus faecalis endocarditis with ampicillin plus ceftriaxone. Annals of internal medicine. 2007;146(8):574–579. 10.7326/0003-4819-146-8-200704170-00008 [DOI] [PubMed] [Google Scholar]

[ppat.1008278.ref054] 54. Loewe S. The problem of synergism and antagonism of combined drugs. Arzneimittelforschung. 1953;3:285–290. [PubMed] [Google Scholar]

[ppat.1008278.ref055] 55. Kristich CJ, Rice LB, Arias CA. Enterococcal infection—treatment and antibiotic resistance In: Enterococci: From commensals to leading causes of drug resistant infection [Internet]. Massachusetts Eye and Ear Infirmary; 2014. [PubMed] [Google Scholar]

[ppat.1008278.ref056] 56. Tripodi M, Locatelli A, Adinolfi L, Andreana A, Utili R. Successful treatment with ampicillin and fluoroquinolones of human endocarditis due to high-level gentamicin-resistant enterococci. European Journal of Clinical Microbiology and Infectious Diseases. 1998;17(10):734–736. 10.1007/s100960050171 [DOI] [PubMed] [Google Scholar]

[ppat.1008278.ref057] 57. Holmberg A, Mörgelin M, Rasmussen M. Effectiveness of ciprofloxacin or linezolid in combination with rifampicin against Enterococcus faecalis in biofilms. Journal of antimicrobial chemotherapy. 2011;67(2):433–439. 10.1093/jac/dkr477 [DOI] [PubMed] [Google Scholar]

[ppat.1008278.ref058] 58. Arias CA, Contreras GA, Murray BE. Management of multidrug-resistant enterococcal infections. Clinical microbiology and infection. 2010;16(6):555–562. 10.1111/j.1469-0691.2010.03214.x [DOI] [PMC free article] [PubMed] [Google Scholar]

[ppat.1008278.ref059] 59. Tang HJ, Chen CC, Zhang CC, Su BA, Li CM, Weng TC, et al. In vitro efficacy of fosfomycin-based combinations against clinical vancomycin-resistant Enterococcus isolates. Diagnostic microbiology and infectious disease. 2013;77(3):254–257. 10.1016/j.diagmicrobio.2013.07.012 [DOI] [PubMed] [Google Scholar]

[ppat.1008278.ref060] 60. Silvestri C, Cirioni O, Arzeni D, Ghiselli R, Simonetti O, Orlando F, et al. In vitro activity and in vivo efficacy of tigecycline alone and in combination with daptomycin and rifampin against Gram-positive cocci isolated from surgical wound infection. European journal of clinical microbiology & infectious diseases. 2012;31(8):1759–1764. 10.1007/s10096-011-1498-1 [DOI] [PubMed] [Google Scholar]

[ppat.1008278.ref061] 61. Bollenbach T, Quan S, Chait R, Kishony R. Nonoptimal microbial response to antibiotics underlies suppressive drug interactions. Cell. 2009;139(4):707–718. 10.1016/j.cell.2009.10.025 [DOI] [PMC free article] [PubMed] [Google Scholar]

[ppat.1008278.ref062] 62. Yu W, Hallinen KM, Wood KB. Interplay between antibiotic efficacy and drug-induced lysis underlies enhanced biofilm formation at subinhibitory drug concentrations. Antimicrobial agents and chemotherapy. 2018;62(1):e01603–17. 10.1128/AAC.01603-17 [DOI] [PMC free article] [PubMed] [Google Scholar]

[ppat.1008278.ref063] 63.Santos-Lopez A, Marshall CW, Scribner MR, Snyder D, Cooper VS. Biofilm-dependent evolutionary pathways to antibiotic resistance. bioRxiv. 2019; p. 581611. [DOI] [PMC free article] [PubMed]

[ppat.1008278.ref064] 64. Martin M, Hölscher T, Dragoš A, Cooper VS, Kovács ÁT. Laboratory evolution of microbial interactions in bacterial biofilms. Journal of bacteriology. 2016;198(19):2564–2571. 10.1128/JB.01018-15 [DOI] [PMC free article] [PubMed] [Google Scholar]

[ppat.1008278.ref065] 65. Reller LB, Weinstein M, Jorgensen JH, Ferraro MJ. Antimicrobial susceptibility testing: a review of general principles and contemporary practices. Clinical infectious diseases. 2009;49(11):1749–1755. 10.1086/647952 [DOI] [PubMed] [Google Scholar]

[ppat.1008278.ref066] 66.Dean Z, Maltas J, Wood KB. Antibiotic interactions shape short-term evolution of resistance in Enterococcus faecalis, Dryad, Dataset; 2020. Available from: 10.5061/dryad.j3tx95x92. [DOI] [PMC free article] [PubMed]

PERMALINK

Antibiotic interactions shape short-term evolution of resistance in E. faecalis

Ziah Dean

Jeff Maltas

Kevin B Wood

Roles

Abstract

Author summary

Introduction

Results

Selection of antibiotic pairs with different interactions

Laboratory evolution across iso-inhibitory dosage combinations

Dual β-lactam combination accelerates growth adaption but selects for similar resistance profiles as adaption to component drugs

Fig 1. Dual β-lactam combination accelerates growth adaption but selects for similar resistance levels to component drugs.

Aminoglycoside/ β-lactam and β-lactam/fluoroquinolone combinations slow growth adaptation and select for resistant profiles distinct from those evolved to the component drugs

Fig 2. Antagonistic β-lactam and aminoglycoside combination slows adaptation with little cross-resistance.

Fig 3. Antagonistic β-lactam and fluoroquinolone combination slows adaptation and selects for populations lacking observed collateral sensitivity of single-drug isolates.

Tigecycline suppresses growth adaptation and eliminates evolution of fluoroquinolone resistance

Fig 4. Tigecycline eliminates fluoroquinolone resistance above a critical concentration.

Geometric rescaling of ancestral growth surface explains condition-dependent growth adaptation when resistant profiles are unchanged

Fig 5. Geometric rescaling of growth surface in ancestral strain explains growth differences between different conditions even when resistance profiles are identical.

Resistance profiles selected in different AMP-STR and CRO-CIP combinations are (nearly) growth-optimized along contours connecting the profiles selected by component drugs

Fig 6. Resistance profiles observed in different AMP-STR and CRO-CIP dosage combinations are (nearly) growth-optimized along the contour separating profiles selected by component drugs.

Discussion

Materials and methods

Strains, media, and growth conditions

Table 1. Antibiotics used in this study.

Laboratory evolution and growth measurement

Estimating per capita growth rate and drug response surfaces

Phenotypic resistance profiling

Supporting information

Acknowledgments

Data Availability

Funding Statement

References

Decision Letter 0

François Balloux

Michael Otto

Roles

Transfer Alert

Author response to Decision Letter 0

Decision Letter 1

Michael Otto

Roles

Acceptance letter

Michael Otto

Roles

Associated Data

Supplementary Materials

Data Availability Statement

ACTIONS

PERMALINK

RESOURCES

Similar articles

Cited by other articles

Links to NCBI Databases