When penicillin came into widespread use in the 1940s, many doctors believed that bacterial infections would soon be a problem of the past. Within a few years, however, hospitals were reporting strains of Staphylococcus aureus that the drug could no longer kill. Alexander Fleming, who had discovered penicillin, had warned of exactly this in 1945: a bacterium exposed to too little of the drug, he said, could be "educated" to resist it. Eight decades later, the World Health Organization counts antimicrobial resistance among the leading threats to global health.
Resistance is not something an individual bacterium learns. It arises from the ordinary variation that exists in any large population. Bacteria divide rapidly, and each time a cell copies its DNA there is a small chance of error. Most of these mutations are harmful or have no effect, but occasionally one alters a protein in a way that protects the cell against a particular drug. In the absence of the antibiotic, such a cell has no special advantage. Once the drug is present, it is among the few that survive, and its descendants soon make up most of the population. The antibiotic does not cause the mutation; it selects for it.
Bacteria can also acquire resistance in a way that animals cannot. Small circular molecules of DNA called plasmids can pass from one bacterium to another, even between different species, carrying genes that confer resistance. A single plasmid may carry several such genes at once, so that a bacterium which receives it becomes resistant to several drugs in one step. This horizontal transfer helps to explain how resistance can spread through a hospital far faster than mutation alone would allow.
The mechanisms themselves are varied. Some bacteria produce enzymes that break the drug down before it can act; the β-lactamases that destroy penicillin are the best-known example. Others alter the target the drug binds to, so that it no longer fits. Still others pump the drug out of the cell as quickly as it enters, or reduce the number of channels through which it can get in.
Because resistance is driven by selection, every use of an antibiotic carries a cost, including uses that are unnecessary. Antibiotics have no effect on viruses, yet they are still often taken for colds and influenza. A dose that is too low, as Fleming warned, may kill the most sensitive bacteria and leave the hardiest alive. Much of the response to resistance therefore depends less on discovering new drugs than on using existing ones more carefully.
