Limits...
The genetic basis of the fitness costs of antimicrobial resistance: a meta-analysis approach.

Vogwill T, MacLean RC - Evol Appl (2014)

Bottom Line: We also find that epistasis can significantly alter the cost of mutational resistance.Overall, our study shows that the cost of antimicrobial resistance can be partially explained by its genetic basis.It also highlights both the danger associated with plasmidborne resistance and the need to understand why resistance plasmids carry a relatively low cost.

View Article: PubMed Central - PubMed

Affiliation: Department of Zoology, University of Oxford Oxford, UK.

ABSTRACT
The evolution of antibiotic resistance carries a fitness cost, expressed in terms of reduced competitive ability in the absence of antibiotics. This cost plays a key role in the dynamics of resistance by generating selection against resistance when bacteria encounter an antibiotic-free environment. Previous work has shown that the cost of resistance is highly variable, but the underlying causes remain poorly understood. Here, we use a meta-analysis of the published resistance literature to determine how the genetic basis of resistance influences its cost. We find that on average chromosomal resistance mutations carry a larger cost than acquiring resistance via a plasmid. This may explain why resistance often evolves by plasmid acquisition. Second, we find that the cost of plasmid acquisition increases with the breadth of its resistance range. This suggests a potentially important limit on the evolution of extensive multidrug resistance via plasmids. We also find that epistasis can significantly alter the cost of mutational resistance. Overall, our study shows that the cost of antimicrobial resistance can be partially explained by its genetic basis. It also highlights both the danger associated with plasmidborne resistance and the need to understand why resistance plasmids carry a relatively low cost.

No MeSH data available.


Related in: MedlinePlus

(A): There is no significant difference between the cost of resistance measured by a proxy (such as growth rate, density at a set time, etc.) (grey bars, mean ± SEM) or by direct competition assays (white bars, mean ± SEM). Each pair of bars is a separate published paper. 3(B): Competitive fitness correlates with growth rate. Each set of symbols represents a different paper, with each point an independent mutation.
© Copyright Policy - open-access
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4380922&req=5

fig03: (A): There is no significant difference between the cost of resistance measured by a proxy (such as growth rate, density at a set time, etc.) (grey bars, mean ± SEM) or by direct competition assays (white bars, mean ± SEM). Each pair of bars is a separate published paper. 3(B): Competitive fitness correlates with growth rate. Each set of symbols represents a different paper, with each point an independent mutation.

Mentions: After employing our criteria for inclusion, our search yielded 77 papers reporting the fitness cost of newly acquired antimicrobial resistance (Appendix S1), which represents a total of 822 independent resistant mutants. These papers used one or more of three broad methodologies to assess the fitness costs of resistance. Firstly, there are direct competition experiments against an ancestral strain performed in vitro (455 isolates). Secondly, there are in vitro proxy measures of fitness such as growth rates, doubling times, maximum yields, etc. (367 isolates). These measures are then standardized against the ancestral strain and used to infer the outcome of direct competitive interactions. Thirdly, there are competition experiments performed in vivo, or strictly speaking inside a mouse (23 isolates). Conveniently, several manuscripts have multiple methods on the same isolate, which thereby allows a comparison of the various methods. Specifically, 23 isolates (five papers) by both an in vitro and in vivo method, and 55 isolates (12 papers) had been assayed by both in vitro methods. If we use each paper as an independent data point, there is no significant difference in the mean cost of resistance between the in vitro and in murine fitness assays (Fig.2A; paired t-test on mean cost per manuscript, in vitro versus in murine, t = 1.17, df = 4, P = 0.307), nor between the two types of in-vitro assays (Fig.3A; paired t-test on mean cost per manuscript, growth rate versus competition, t = −0.394, df = 11, P = 0.703). This suggests that if a particular resistance mechanism is found to be either high or low cost by one assay, it is likely to be found to have the same relative cost by another methodology. Similarly, there are also significant correlations in fitness for individual resistance isolates which have been assayed in more than one way (Fig.2B: in vitro vs In murine, df = 21, r = 0.814, P < 0.001; Fig.3B: competition vs growth rate, df = 53, r = 0.763, P < 0.001). Both of these correlations remain significant if we control for the differences in the means of different papers (partial correlation between in vitro vs in murine controlling for differences between manuscripts, df = 20, r = 0.775, P < 0.001; partial correlation between competition versus growth rate controlling for differences between manuscripts, df = 52, r = 0.752, P < 0.001).


The genetic basis of the fitness costs of antimicrobial resistance: a meta-analysis approach.

Vogwill T, MacLean RC - Evol Appl (2014)

(A): There is no significant difference between the cost of resistance measured by a proxy (such as growth rate, density at a set time, etc.) (grey bars, mean ± SEM) or by direct competition assays (white bars, mean ± SEM). Each pair of bars is a separate published paper. 3(B): Competitive fitness correlates with growth rate. Each set of symbols represents a different paper, with each point an independent mutation.
© Copyright Policy - open-access
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4380922&req=5

fig03: (A): There is no significant difference between the cost of resistance measured by a proxy (such as growth rate, density at a set time, etc.) (grey bars, mean ± SEM) or by direct competition assays (white bars, mean ± SEM). Each pair of bars is a separate published paper. 3(B): Competitive fitness correlates with growth rate. Each set of symbols represents a different paper, with each point an independent mutation.
Mentions: After employing our criteria for inclusion, our search yielded 77 papers reporting the fitness cost of newly acquired antimicrobial resistance (Appendix S1), which represents a total of 822 independent resistant mutants. These papers used one or more of three broad methodologies to assess the fitness costs of resistance. Firstly, there are direct competition experiments against an ancestral strain performed in vitro (455 isolates). Secondly, there are in vitro proxy measures of fitness such as growth rates, doubling times, maximum yields, etc. (367 isolates). These measures are then standardized against the ancestral strain and used to infer the outcome of direct competitive interactions. Thirdly, there are competition experiments performed in vivo, or strictly speaking inside a mouse (23 isolates). Conveniently, several manuscripts have multiple methods on the same isolate, which thereby allows a comparison of the various methods. Specifically, 23 isolates (five papers) by both an in vitro and in vivo method, and 55 isolates (12 papers) had been assayed by both in vitro methods. If we use each paper as an independent data point, there is no significant difference in the mean cost of resistance between the in vitro and in murine fitness assays (Fig.2A; paired t-test on mean cost per manuscript, in vitro versus in murine, t = 1.17, df = 4, P = 0.307), nor between the two types of in-vitro assays (Fig.3A; paired t-test on mean cost per manuscript, growth rate versus competition, t = −0.394, df = 11, P = 0.703). This suggests that if a particular resistance mechanism is found to be either high or low cost by one assay, it is likely to be found to have the same relative cost by another methodology. Similarly, there are also significant correlations in fitness for individual resistance isolates which have been assayed in more than one way (Fig.2B: in vitro vs In murine, df = 21, r = 0.814, P < 0.001; Fig.3B: competition vs growth rate, df = 53, r = 0.763, P < 0.001). Both of these correlations remain significant if we control for the differences in the means of different papers (partial correlation between in vitro vs in murine controlling for differences between manuscripts, df = 20, r = 0.775, P < 0.001; partial correlation between competition versus growth rate controlling for differences between manuscripts, df = 52, r = 0.752, P < 0.001).

Bottom Line: We also find that epistasis can significantly alter the cost of mutational resistance.Overall, our study shows that the cost of antimicrobial resistance can be partially explained by its genetic basis.It also highlights both the danger associated with plasmidborne resistance and the need to understand why resistance plasmids carry a relatively low cost.

View Article: PubMed Central - PubMed

Affiliation: Department of Zoology, University of Oxford Oxford, UK.

ABSTRACT
The evolution of antibiotic resistance carries a fitness cost, expressed in terms of reduced competitive ability in the absence of antibiotics. This cost plays a key role in the dynamics of resistance by generating selection against resistance when bacteria encounter an antibiotic-free environment. Previous work has shown that the cost of resistance is highly variable, but the underlying causes remain poorly understood. Here, we use a meta-analysis of the published resistance literature to determine how the genetic basis of resistance influences its cost. We find that on average chromosomal resistance mutations carry a larger cost than acquiring resistance via a plasmid. This may explain why resistance often evolves by plasmid acquisition. Second, we find that the cost of plasmid acquisition increases with the breadth of its resistance range. This suggests a potentially important limit on the evolution of extensive multidrug resistance via plasmids. We also find that epistasis can significantly alter the cost of mutational resistance. Overall, our study shows that the cost of antimicrobial resistance can be partially explained by its genetic basis. It also highlights both the danger associated with plasmidborne resistance and the need to understand why resistance plasmids carry a relatively low cost.

No MeSH data available.


Related in: MedlinePlus