Concept

Antibiotic resistance

T-042Home BU-203Threads systems · evolution

Statement

Evolution in action under drug selection.

Why it matters

bacterial-growth-curve established the exponential phase's population dynamics; antibiotic resistance is what happens when a strong evolutionary selective pressure — the drug — acts on that exponentially growing, genetically variable population. It is one of the clearest, most directly observable cases of natural selection acting within a human lifetime rather than over geological time, and it matters practically because resistance evolution now measurably outpaces the rate of new antibiotic development, making stewardship (limiting unnecessary antibiotic use) a central public-health strategy rather than an optional precaution.

Hypotheses

Standing genetic variation already exists in a bacterial population before drug exposure, via pre-existing mutation and via horizontal-gene-transfer of resistance genes from other, possibly unrelated organisms, rather than being newly and specifically induced by the drug itself.Without pre-existing variation, resistance could not be selected at all, because there would be nothing to select among; the drug is a selective agent, not a mutagen that creates resistance to order. A resistance mechanism carries a fitness benefit specifically in the presence of the drug, and this benefit typically comes with some cost, such as reduced growth rate, in the drug's absence.Without this trade-off structure, resistant strains might be expected to sweep to permanent, irreversible fixation even after drug use stops, which is not always what is observed; the trade-off is what makes reduced or targeted drug use a potentially effective strategy for reversing resistance.

Proof

\[ N(t)=N_0e^{kt},\quad\text{expected pre-existing resistant mutants}\approx N(t)\cdot\mu \]

Bacterial populations are large and grow exponentially (bacterial-growth-curve), and mutation occurs at each replication with some small per-generation probability \(\mu\); even a low per-generation mutation rate produces some resistant mutants already present in a sufficiently large population before any drug is ever applied. A

\[ \text{Resistance mechanism}\in\{\text{target modification, enzymatic inactivation, efflux/reduced uptake, pathway bypass}\} \]

A resistance mechanism generally falls into one of a small number of categories: modification of the drug's target (e.g. an altered penicillin-binding protein), enzymatic inactivation of the drug (e.g. a beta-lactamase hydrolysing the beta-lactam ring), reduced drug entry or active efflux, or bypass of the inhibited pathway. A

\[ \text{Selection coefficient }s>0\text{ for the resistance allele under drug exposure}\ \Rightarrow\ \text{its frequency }p\text{ rises each generation} \]

Applying the drug imposes a strong selective pressure: susceptible cells die or fail to reproduce, while pre-existing resistant cells continue dividing largely unimpeded, so the resistant sub-population's relative frequency rises sharply — differential survival and reproduction, the definition of natural selection, acting directly on a bacterial population within days. A

\[ \text{Resistance genes spread horizontally between unrelated bacteria (plasmids, transposons, transformation), not only vertically from parent to offspring} \]

horizontal-gene-transfer dramatically accelerates the spread of resistance beyond what mutation and vertical inheritance alone would achieve, allowing a single resistance gene to move between species that share no direct ancestor-descendant relationship. A

\[ s<0\ (\text{drug absent, fitness cost})\ \Rightarrow\ p\text{ may fall, but compensatory mutation can restore fitness without losing resistance} \]

If the resistance mechanism carries a fitness cost in the drug's absence, removing the drug can shift the selective balance back toward susceptible strains over subsequent generations, though not always to full reversal, since compensatory mutations can subsequently reduce or eliminate the original cost while resistance itself is retained. B

Result

\[ \text{Standing variation}+\text{selection under drug exposure}+\text{horizontal spread}\ \Rightarrow\ \text{resistant population dominates} \]

Reading. Resistance is not created by the drug; the drug selects among variation that already exists, and horizontal gene transfer lets that selected variation move well beyond the population, or even species, in which it first arose.

Scope. Applies generally to microbial populations under selective antimicrobial pressure of any kind; viral-replication-cycles' rapidly replicating viruses show a directly analogous drug-resistance phenomenon. The relative importance of mutation versus horizontal transfer varies by organism and drug class.

Corollaries & converses

microbial-metabolic-diversity already establishes how varied bacterial metabolism is between species; that same diversity underlies why a resistance mechanism useful to one species can, once transferred (Step 4), become clinically important in an entirely unrelated pathogen.
Completing a full prescribed antibiotic course, rather than stopping once symptoms improve, is recommended specifically to minimise the number of generations a partially-suppressed, more selection-prone bacterial population is exposed to sub-lethal drug concentrations, a regime that favours intermediate, partially resistant genotypes.
Converse: given a rapid rise in a resistance gene's frequency in a bacterial population following a clinical antibiotic course, one can infer a fitness advantage (Step 3) was already present in that population, without needing to observe the causative mutation event directly.

Fails without

Drop pre-existing standing genetic variation (assume resistance mutations could only arise on demand, after drug exposure begins): population-genetic theory and direct measurement of pre-exposure mutation rates both indicate resistant variants are already present at low frequency before treatment starts; without this standing variation, resistance would emerge far more slowly and far less reliably than is actually observed clinically.
Drop the fitness-cost trade-off of resistance: a resistant strain would never be out-competed once the drug is withdrawn, and resistance would be expected to persist and spread indefinitely even without ongoing selection — contrary to the fitness cost that underlies antibiotic stewardship strategies such as restricted or rotated drug use.

Common errors

Believing bacteria "develop" resistance in direct response to the drug's presence, as if the drug itself caused the relevant mutation; the drug selects among variation already present beforehand (Step 1), rather than creating resistance to order.
Assuming resistance always carries an unavoidable fitness cost that reliably disappears once drug pressure is removed; compensatory mutations frequently restore fitness while resistance persists (Step 5), which is why resistant strains can remain common long after a particular drug falls out of use.
Treating antibiotic resistance as spreading only vertically, from a resistant bacterium to its direct descendants, and overlooking horizontal-gene-transfer's much faster, cross-species route (Step 4).
Assuming a single antibiotic's resistance mechanism generalises to every drug in its class; target modification in particular is often specific to structural details of one drug or a narrow group of chemically related drugs.

Discussion

Alexander Fleming warned, in his 1945 Nobel lecture, of the risk of under-dosing penicillin, anticipating that resistant strains could be selected for by exposure to sub-lethal concentrations — a remarkably early anticipation of what is now one of the most pressing problems in clinical medicine.

Because horizontal gene transfer (Step 4) allows resistance genes to move between species, resistance genes selected for by antibiotic use in agriculture can, in principle, ultimately reach human pathogens even without any direct human antibiotic exposure at all — one of the key arguments behind restricting non-clinical antibiotic use, not only clinical stewardship.

Common misconception: that an individual person or animal "becomes resistant" to an antibiotic. Resistance is a property of the microbial population's genotype, not of the host; a person can be infected by a resistant strain of bacteria, but a host organism does not itself become resistant to a drug the way it might become immune to a pathogen.

Worked examples

\[ N_0=10^6\text{ cells},\ \mu\sim10^{-8}\text{ per cell per generation (order-of-magnitude, commonly cited)},\ 20\text{ generations before drug exposure} \]

Twenty generations of exponential growth take the population from \(10^6\) to roughly \(10^6\times2^{20}\approx10^{12}\) cells; multiplying by the order-of-magnitude mutation-rate estimate \(\mu\sim10^{-8}\) gives an expected pre-existing resistant-mutant count on the order of \(10^4\) cells, entirely before any drug is applied. B

\[ \text{Applying the drug removes susceptible cells; the}\sim10^4\text{ pre-existing resistant cells continue dividing} \]

Once the drug is applied, the resistant sub-population, though initially a small minority, has \(s>0\) (Step 3) and grows to dominate the surviving population within a comparatively small number of further generations, since susceptible cells are simultaneously being removed. A

\[ \text{Expected resistant mutants}\sim10^4\text{ cells, present before any drug is applied} \]

Reading. Even an extremely rare per-generation mutation event becomes essentially certain to have already occurred at least once, given a sufficiently large and rapidly dividing population — precisely the situation created by bacterial-growth-curve's exponential phase.

Scope. Order-of-magnitude reasoning only; actual resistant-mutant counts depend on the true locus-specific mutation rate, population size and generation number, all of which vary widely between species, genes and growth conditions.

Further example. A clinical isolate of Escherichia coli is tested against ciprofloxacin. The Minimum Inhibitory Concentration (MIC) is determined to be 2 \(\mu\text{g/mL}\). The pharmacokinetic profile of the patient indicates that the peak serum concentration (\(C_{\max}\)) achieved after a standard dose is 15 \(\mu\text{g/mL}\), but the trough concentration (\(C_{\min}\)) just before the next dose is 0.5 \(\mu\text{g/mL}\). Calculate the ratio of the peak concentration to the MIC (\(C_{\max}/\text{MIC}\)) and the ratio of the trough concentration to the MIC (\(C_{\min}/\text{MIC}\)). Based on these ratios, determine if the dosing regimen maintains concentrations above the MIC for the entire dosing interval, and explain the evolutionary consequence of the period where concentrations fall below the MIC.

\[ \text{Given: } C_{\max}=15\ \mu\text{g/mL},\ \text{MIC}=2\ \mu\text{g/mL},\ C_{\min}=0.5\ \mu\text{g/mL} \]

Identify the critical pharmacokinetic and microbiological parameters. The MIC represents the threshold concentration required to inhibit visible growth, while \(C_{\max}\) and \(C_{\min}\) define the upper and lower bounds of drug exposure during a dosing interval. A

\[ \frac{C_{\max}}{\text{MIC}} = \frac{15}{2} = 7.5 \]

Calculate the peak exposure ratio. A ratio of 7.5 indicates that at the peak of the dosing interval, the drug concentration is 7.5 times higher than the minimum concentration needed to inhibit the bacterium, providing a strong initial bactericidal or bacteriostatic effect. A

\[ \frac{C_{\min}}{\text{MIC}} = \frac{0.5}{2} = 0.25 \]

Calculate the trough exposure ratio. A ratio of 0.25 indicates that at the end of the dosing interval, the drug concentration has fallen to 25% of the MIC. This is below the threshold required to inhibit growth. A

\[ C_{\min} < \text{MIC} \implies \text{Sub-inhibitory exposure} \]

Since the trough concentration is less than the MIC, there is a portion of the dosing interval where the drug concentration is insufficient to inhibit bacterial growth. This period allows susceptible bacteria to resume replication and provides a window for selection of resistant mutants. B

\[ \text{Evolutionary Consequence: } \text{Selection for } R \text{ under sub-MIC pressure} \]

Sub-inhibitory concentrations can act as a selective pressure that favors resistant strains over susceptible ones, as the resistant strains can grow while susceptible strains are inhibited or killed. This promotes the emergence and spread of antibiotic resistance. B

\[ C_{\max}/\text{MIC} = 7.5,\ C_{\min}/\text{MIC} = 0.25;\ \text{Regimen fails to maintain } C > \text{MIC} \]

Reading. The dosing regimen does not maintain drug concentrations above the MIC for the entire interval. The period of sub-inhibitory exposure creates a selective environment that favors the survival and proliferation of resistant mutants, thereby promoting antibiotic resistance.

Scope. This analysis assumes that the MIC is a reliable predictor of inhibitory concentration and that pharmacokinetic parameters are constant. In reality, tissue penetration, protein binding, and individual patient variability can alter effective drug concentrations at the site of infection.

Problems

A patient stops a course of antibiotics early, once symptoms resolve, while some bacteria remain. Explain, using Steps 1 and 3, why this practice specifically favours the emergence of resistance compared with completing the full course.
Solution
Stopping early leaves surviving low-susceptibility or intermediate genotypes under continued, sub-lethal selective pressure for longer, effectively giving \(s>0\) more generations to act (Step 3) before the infection is fully cleared, raising the resistant sub-population's relative frequency and its odds of becoming established, compared with a course long enough to eliminate the infection entirely.
Explain, referencing Step 4, why a resistance gene first evolving in a harmless environmental bacterium could eventually appear in a clinically important human pathogen that never directly encountered the original antibiotic-producing organism.
Solution
Horizontal gene transfer via plasmids, transposons or transformation moves genetic material across species boundaries independent of reproductive lineage; a resistance gene can therefore move from the environmental organism into an intermediate species, and from there into a human pathogen, without the pathogen ever directly encountering the original source organism or antibiotic.
A resistant bacterial strain is found to grow more slowly than a susceptible strain when cultured without antibiotic present. Using Step 5, predict what happens to the frequency of the resistant strain in a mixed population if antibiotic use is discontinued, and explain one reason the resistant strain might nonetheless persist longer than naively expected.
Solution
The naive prediction is that the resistant strain's frequency declines over subsequent generations, since \(s\lt0\) in the drug's absence (Step 5). It may nonetheless persist longer than expected if a compensatory mutation arises that restores growth rate while the resistance mechanism itself is retained, removing or reducing the fitness cost that would otherwise drive the decline.
A patient carries \(N = 1.0\times10^{8}\) bacteria of a susceptible strain, and resistance to a drug arises by a single point mutation at rate \(\mu = 1.0\times10^{-9}\) per cell division. Treating the population as having arisen through about \(N\) divisions, estimate (a) the expected number of resistant mutants present before treatment starts and (b) the probability that at least one is present. (c) Explain why combination therapy with two drugs (each with \(\mu = 10^{-9}\), independent targets) changes the picture.
Solution
(a) Expected mutants \(m = N\mu = (1.0\times10^{8})(1.0\times10^{-9}) = 0.10\). (b) Mutations are rare and independent, so the number present is approximately Poisson with mean \(m\): \(P(\ge 1) = 1 - e^{-m} = 1 - e^{-0.10} = 0.095\), about a 10% chance that treatment selects a pre-existing resistant clone. (c) A cell must carry both mutations to survive, with probability per division \(\mu^2 = 10^{-18}\); then \(m = 10^{8}\times10^{-18} = 10^{-10}\), so pre-existing double mutants are essentially never present. This is the logic behind multidrug regimens for tuberculosis and HIV.
Explain why the use of broad-spectrum antibiotics in livestock feed, even at sub-therapeutic doses, can contribute to the emergence of antibiotic resistance in human pathogens. Reference the concepts of horizontal gene transfer and selective pressure.
Solution
Broad-spectrum antibiotics in livestock feed exert selective pressure on the gut microbiome of animals, favoring the survival and proliferation of bacteria that carry resistance genes. These resistant bacteria can then transfer their resistance genes to other bacteria, including potential human pathogens, via horizontal gene transfer mechanisms such as conjugation, transformation, or transduction. When these pathogens are subsequently transmitted to humans (e.g., through the food chain), they may already carry resistance genes, making infections harder to treat. This process is amplified by the large population sizes of bacteria in the gut and the continuous exposure to sub-therapeutic antibiotic levels, which provide a persistent selective advantage to resistant strains.