Some scientists have suggested that if humans temporarily cease using antibiotics, bacterial populations will lose their resistance in one of two ways — by evolving back to being susceptible to antibiotics or by being out competed by the remaining sensitive bacteria that are no longer at a disadvantage — at which point humans could resume using antibiotics with much greater effect.
The thinking goes like this: Removing the antibiotic would result in the evolution of sensitive bacteria only if the sensitive bacteria have an advantage over the resistant bacteria. The most commonly envisioned advantage? That antibiotic resistance has a cost. Maybe the changes that make bacteria resistant (changes in bacterial membranes, for example, or DNA replication) also make them less able to perform other bacterial functions. By removing the antibiotic from the environment, mutations back to the original bacterial physiology would be favored, because sensitive bacteria are better than resistant bacteria at doing all the things bacteria have to do in the absence of antibiotics.
This hypothesis is great, but would it work? Several experiments have been conducted regarding this question, and they show, unfortunately, that the solution isn't as simple as instituting a moratorium on antibiotic use. Why? For the reasons explained in the following sections.
Although bacteria resistance comes at a cost (they tend to grow more slowly in the presence of antibiotics than the original sensitive bacteria do), the process of amelioration — secondary mutations decreasing the debilitating effect of the initial mutation — is favored by selection. If the initial antibiotic-resistant mutant had a high fitness cost, subsequent evolution selects for additional mutations that increase the fitness of the antibiotic-resistant strain. As time goes by, the antibiotic-resistant bacteria don't have as low a fitness compared with the original bacteria (in the absence of antibiotics) as scientists would wish them to have. Thus, removing the antibiotics won't have as much of an effect as we would all want.
Starting with a wild sensitive bacterial strain, Stephanie Schrag and cowork-ers grew the strain until they could find and isolate an antibiotic-resistant bacterial mutant whose fitness was lower because of its mutation (the old DNA sequence was better than the new DNA sequence for some key bacterial function). They then grew this antibiotic-resistant bacteria for many generations in the lab and let evolution happen. As the antibiotic-resistant bacteria evolved, their fitness improved because of compensatory mutations, additional mutations that returned their fitness to about the level of the original antibiotic-sensitive strain.
Think back to that antibiotic-resistant bacterium that's not so good at DNA replication. Because it has the resistant allele, it manages to spread to the environment after all the sensitive bacteria have expired; now it lives in a sea of bacteria that are all resistant to the antibiotic but that all have a little trouble replicating their DNA. In this population, mutations that restore the ability to replicate DNA easily will be favored, and after selection has proceeded for some time, the antibiotic-resistant bacteria gain back much of the function they initially lost.
Compensatory and back mutations aren't what they're cut out to be
So you have a strain of antibiotic-resistant bacteria that are now just as fit as the original antibiotic-sensitive strain. Now imagine what would happen if you take away all the antibiotics. Mutations still occur, and eventually one will appear that undoes what the original resistance mutation did (replaces the new sequence with the old sequence again), making this strain sensitive again. Will this new antibiotic sensitive strain take over so we can start using antibiotics again? Unfortunately no, because the old DNA sequence is not better than the new DNA sequence in the presence of the subsequent, post resistance, DNA changes. It was in the original strain, but not any more.
eSiv^*^. To determine that compensatory mutations could actually lower the proba-tj/ bility that the resistant bacteria would evolve back toward sensitivity to antibiotics if antibiotics were removed, Schrag and coworkers conducted a second study. Using genetic engineering, they replaced the DNA responsible for antibiotic resistance with the DNA sequence of the sensitive strain while leaving the compensatory mutations unchanged. In effect, they created the bacteria that would exist if the antibiotic-resistant strain mutated back to being antibiotic sensitive.
The surprising result was that the antibiotic-sensitive bacteria they'd created in the laboratory grew more slowly than the resistant strain in the absence of antibiotics. What this result means is that all the compensatory mutations that piled up in the antibiotic-resistant bacteria were advantageous only in the presence of the antibiotic-resistant gene; they were deleterious in the presence of the original sensitive gene.
^jjjABEft After an antibiotic-resistant gene has appeared and compensatory mutations of the sort that Schrag and coworkers found in their laboratory have arisen, the antibiotic-resistant gene can be more fit than the antibiotic-sensitive gene, even in the absence of antibiotics. Why? Because a back mutation (a mutation that undoes exactly what the first mutation did) rendering one of these antibiotic-resistant bacteria susceptible could have lower fitness even in the absence of antibiotics. That situation is troubling because when all the bacteria become resistant, they stay resistant even if the antibiotic is stopped for a while. While we don't know exactly how common this sort of result would be, it's not good news.
Mutations aren't necessarily as costly as you'd think
Although most antibiotic-resistant mutations have associated costs, some of them have very small costs, and some even have no cost. An awful lot of bacteria are around, and if even a few of them are capable of acquiring antibiotic-resistant mutations with no deleterious effects, this small class of mutations will be able to beat out the much larger number of antibiotic-resistant mutations that result in less healthy resistant bacteria. In a battle between the debilitated bacteria and the undamaged bacteria, the debilitated ones won't last long.
Just as genetic differences exist between different humans, genetic differences exist between different bacteria within a species. This situation raises the question of how important these different genetic backgrounds are in determining the associated costs of resistance. Fred Cohan and coworkers found that the genetic background in which an antibiotic mutation appeared was an important factor in the associated costs of that mutation. They determined that in some cases, acquiring resistance to an antibiotic was associated with few or no costs in other areas of bacterial physiology.
Cohan and his coworkers performed an experiment using the common soil bacteria Bacillus subtilis, which can take up DNA from the environment and incorporate it into its own genetic material. The researchers isolated a strain of Bacillus subtilis that was resistant to the antibiotic rifampin, made many copies of the gene that made the bacteria antibiotic resistant, and then introduced this gene into a collection of wild Bacillus subtilis strains simply by exposing samples of the different bacteria to the DNA responsible for antibiotic resistance. As they expected, the sample bacteria took up this gene.
For each of the wild bacteria, they compared the growth rate of the strain before the introduction of the antibiotic-resistant gene with the growth rate of the strain after the introduction of antibiotic resistance. In most, but not all, cases, they found a substantial decrease in growth rate with the introduction of the antibiotic-resistant gene.
Their conclusion? Antibiotic resistance indeed had a cost — but not always. Costs differed greatly among the different genetic backgrounds, and in a small number of cases, the antibiotic-resistant bacteria weren't significantly debilitated compared with the original sensitive strain.
To determine the clinical importance of antibiotic mutants with different fitness effects, Sebastien Gagneux, Clara Davis, Brendan Bohannan, and others conducted a series of experiments with the strain of bacteria that causes tuberculosis. They found that the mutations they could identify in the laboratory as having the lowest fitness costs were the same antibiotic-resistant mutations that were responsible for the resistant strains of tuberculosis present in clinical settings.
The researchers conducted laboratory experiments to measure the fitness iff costs of a series of antibiotic-resistant mutations. They examined a variety of mutations and found that some tended to have low fitness costs and that others tended to have high fitness costs. Then they examined the mutations present in clinical settings.
What they found was that antibiotic-resistant tuberculosis strains isolated from patients were very likely to contain antibiotic-resistance genes that had been shown in their laboratory experiments to have low fitness costs. They never found, in their patients, antibiotic-resistant tuberculosis strains with resistance alleles that had been shown to have high fitness costs. Both types of mutations occur, but the types with high fitness costs don't increase to the point where they are detectable. Furthermore, some of the clinical strains harboring the lowest-cost antibiotic-resistance mutation exhibited no decrease in fitness relative to the sensitive strain.
¿jjjABEft This is a finding of major importance, and unfortunately it's bad news. It had been suggested (and hoped) that if we reduced our use of antibiotics, the antibiotic-resistant bacteria would decline because they were supposed to be less fit than the original strains in the absence of antibiotics. They would just get overwhelmed by the few remaining resistant bacteria. Turns out that while many possible mutants are less fit, the ones that take over are the ones that aren't.
Was this article helpful?