Antibiotic resistance

Antibiotics are medicines that kill the bacteria that make us sick, and for about eighty years they have turned once-deadly infections into minor problems. But bacteria are fighting back, and some have grown so tough that our strongest medicines barely touch them. We call these survivors superbugs.

Here is the surprising part: the antibiotic does not turn a bacterium into a superbug. Bacteria live in enormous numbers, and when they copy themselves they make tiny random mistakes, like typos. Most do nothing, but every so often, by pure luck, one happens to make a bacterium able to survive a particular antibiotic. That lucky bacterium was already there before the medicine arrived.

Now take the antibiotic. It wipes out almost all the ordinary bacteria, but the lucky survivors are left untouched, and with their neighbours gone they multiply like crazy. Bacteria can make a new generation in twenty minutes, so within a day there are billions of them, nearly all carrying the survival trick. The drug did not create the superbug. It cleared away everything except the superbug. That is evolution, fast enough to watch. Worse, bacteria can pass these tricks to each other like trading cards, even to completely different kinds of bacteria, so one bug's trick can spread far and wide.

This is why doctors say not to use antibiotics unless we truly need them. And here is the twist. Bacteria did not start resisting because of us. They have been making poisons to fight each other, and surviving each other's poisons, for billions of years. Scientists have found resistance inside bacteria frozen in the ground for tens of thousands of years, long before any human made a medicine. We did not start this war. We walked into an ancient one and handed the toughest fighters a reason to take over.

In 1928 Alexander Fleming noticed a mould had killed the bacteria on one of his plates, and the antibiotic era began. The mould was making penicillin, a chemical weapon against bacteria, and within decades antibiotics had transformed medicine. They work by attacking features bacterial cells have and ours do not, like the tough wall around a bacterium or the machinery it uses to build proteins, which lets them kill the invader without poisoning us.

Almost immediately the bacteria resisted, and the reason is evolution. A colony holds billions of cells that reproduce astonishingly fast, sometimes doubling every twenty minutes, and each time a bacterium copies its DNA it makes occasional random errors. In a population that large, essentially every possible mutation is generated somewhere all the time. A few, by chance, happen to let a bacterium survive a drug, and crucially they appear at random whether or not the antibiotic is present. The drug does not cause them. It acts as a filter: the vast majority without the lucky mutation die, the rare resistant ones survive and multiply to fill the gap, and within a few generations the population is dominated by resistant cells. This is natural selection in its fastest form, quick enough to watch in days.

A resistant bacterium has several possible tricks. It can make an enzyme that chops the antibiotic apart, which is how penicillin resistance often works. It can change the shape of the molecule the drug attacks so the drug no longer fits. Or it can pump the drug out, or seal the openings it uses to get in. Each can be the payoff of a single mutation, or a gene picked up from elsewhere.

That last point is what makes resistance so dangerous. Bacteria do not only pass genes to their offspring; they also swap them sideways, even between species, by trading small loops of DNA called plasmids. A single plasmid can carry resistance to several antibiotics at once, so one exchange can create a multidrug-resistant bug. This horizontal gene transfer lets a trick that evolves in one place leap across the bacterial world far faster than ordinary inheritance.

Every use of an antibiotic strengthens this selection. Prescribing them for colds and flu, which are viral and untouched by antibiotics, only breeds resistance, as does the huge use of antibiotics in farm animals. The result is superbugs like MRSA that shrug off standard treatment. Drug-resistant infections were directly responsible for an estimated 1.27 million deaths worldwide in 2019, and on current trends one analysis projects around 39 million deaths between 2025 and 2050.

The deepest surprise is that none of this began with us. Microbes have produced antibiotics to compete for billions of years, and resisted one another's weapons for just as long. Bacterial DNA preserved in permafrost for thirty thousand years already carries genes for resistance to several modern antibiotic types. By flooding the world with antibiotics we did not create resistance. We turned an ancient, slow-burning arms race into a global emergency.

State the problem precisely and it is evolutionary biology on a compressed timescale. A clinical bacterial population spans \(10^9\) cells or more, reproduces in tens of minutes, and mutates at roughly \(10^{-9}\) to \(10^{-10}\) per base pair per division, so across the whole population essentially every single-nucleotide variant is generated continuously. Antibiotic therapy imposes intense directional selection on this standing variation, and resistance follows.

Resistance is selection on pre-existing variation, not an induced response. The Luria-Delbrück fluctuation test of 1943 showed that resistance mutations arise spontaneously and at random before exposure to the selective agent. The signature was the variance: independent cultures yielded wildly different numbers of resistant cells, the fingerprint of rare mutations occurring at random times during growth, not a directed reaction to the agent. The drug is a sieve, not a teacher. The picture is complicated but not overturned by stress-induced mutagenesis, where sub-lethal exposure raises mutation rates and hypermutator lineages supply selection with more raw material.

The molecular routes are few in kind and endlessly varied in detail. Enzymatic inactivation: beta-lactamases hydrolyze the beta-lactam ring, with carbapenemases such as NDM-1 and KPC defeating ever broader classes. Target modification: MRSA's mecA gene encodes a penicillin-binding protein with low drug affinity, and gyrase mutations confer fluoroquinolone resistance. Reduced accumulation: efflux pumps expel drugs and porin loss limits entry, especially in Gram-negatives. Target bypass: a redundant, drug-insensitive enzyme. Most follow from a single mutation or one acquired gene, which is why they appear so readily.

Horizontal gene transfer mobilizes resistance across the bacterial world. Conjugation transfers plasmids, often carrying several resistance genes, directly between cells and across species; transformation takes up free DNA, and transduction moves it inside bacteriophages. These genes sit in mobile elements, transposons and integrons, that assemble multidrug cassettes. The community's shared pool of resistance genes, the resistome, is therefore highly mobile, which is why multidrug resistance can emerge in a single transfer rather than requiring several independent mutations.

Selection operates below the killing dose, and resistance is not freely reversible. Resistant variants are favored across a wide window of sub-inhibitory concentrations, conditions common in residual drug levels and environmental contamination. Resistance often carries a fitness cost, so resistant cells grow more slowly without the drug, which should let susceptible strains rebound when use stops. In practice reversal is slow or absent, because compensatory mutations restore fitness while keeping resistance, locking the trait in. Stewardship can slow resistance but cannot reliably undo it.

The pipeline has not kept pace, and the economics work against it. No major new antibiotic class has reached the clinic in decades; most recent drugs are derivatives to which resistance often emerges quickly. The incentives are perverse, since a valuable new antibiotic should be used as little as possible to preserve it, undercutting a development effort costing on the order of a billion dollars. The 2015 discovery of the plasmid-borne mcr-1 gene, conferring transferable resistance to the last-resort drug colistin, showed even the final tier is mobile. Resistance to any given antibiotic is best understood as an eventual certainty, not a risk.

And here is what the crisis really is. We picture resistance as something our drugs created, but antibiotics are ecological molecules that fungi and bacteria have used against each other for hundreds of millions of years, and resistance is the equally ancient countermeasure. Genes for resistance to several modern drug classes have been recovered from thirty-thousand-year-old permafrost, and bacteria sealed in an isolated cave for some four million years proved resistant to drugs they could never have met. Fleming warned of it in his 1945 Nobel lecture. The clinical crisis is not a new phenomenon but the mobilization of an old one: by saturating the planet with antibiotics, we have run an evolutionary experiment that rewards every organism able to tap a vast pre-existing reservoir. The bacteria were ready. We are the variable that changed.