Antibiotics: selective toxicity and how resistance defeats it

Killing bacteria is easy. Bleach does it, so does fire. The hard requirement is selective toxicity: damaging the microbe at a concentration the patient tolerates. Every useful antibiotic exploits a structure or a pathway that bacteria have and human cells either lack or build differently.

The folate case is the most elegant. You get folate from food. A bacterium has no transporter for it and must synthesise it, so blocking that pathway starves the microbe and leaves your diet untouched.

Each path runs from a drug class to the molecule it binds, then to the failure that follows. Note that two of the four block a process rather than destroy a structure, which is why they stop growth rather than immediately killing.

Whether a drug works on a particular isolate is a measurement, not a property of the species.

Definition: The minimum inhibitory concentration, or MIC, is the lowest concentration of an antimicrobial that prevents visible growth of an organism after overnight incubation, usually reported in milligrams per litre.

The standard method is broth microdilution. A plate of wells holds a doubling series of drug concentrations, every well is inoculated with the same number of cells, and after incubation you read the lowest clear well.

So if wells at 0.5, 1 and 2 mg/L are cloudy and the 4 mg/L well is clear, the MIC is 4 mg/L. Because the series doubles, MIC values are always powers of two, and a genuine one-step change is a doubling rather than a percentage.

The single most common misunderstanding about resistance is about who changes.

Gotcha: People do not become resistant to antibiotics. Bacterial populations do. Your body does not adapt to penicillin; the bacteria in and around you are selected for surviving it, and those survivors are what a later infection may be seeded from.

The distinction matters because it changes what an intervention has to target. If patients acquired resistance, the problem would stay with the patient. Because populations of bacteria acquire it, resistance spreads between people, between farms and hospitals, and across species boundaries through mechanisms covered later in this lesson. A resistant strain that arose in one person's gut can end up causing an infection in someone who has never taken the drug.

Resistance is not one trick. There are four broad strategies, and a single cell can run several at once.

Efflux and permeability are the broad, unfussy defences: they act on whole chemical classes at once rather than one molecule, which is how a single change can raise the MIC for several unrelated drugs.

A cell can acquire a resistance gene in two fundamentally different ways.

Vertical: a mutation arises during replication and is inherited by that cell's descendants. Slow, and confined to one lineage.

Horizontal: the gene arrives ready-made from another cell, often a different species, by one of three routes.

• Conjugation: direct contact, and a plasmid copy is transferred. Considered the dominant route for resistance genes.
• Transformation: the cell takes up naked DNA released by dead neighbours.
• Transduction: a bacteriophage packages host DNA and injects it into the next cell it infects.

Horizontal transfer is why resistance does not have to be reinvented. A plasmid carrying several resistance genes can convert a susceptible cell into a multidrug-resistant one in a single event.

The recipient does not merely gain the gene. It gains a copy of the mobile element that carries it, and can pass it on again, which is what makes conjugative plasmids spread through a population rather than merely persist in it.

There is a subtle question underneath all of this, and getting it wrong leads to bad conclusions about how to use antibiotics.

The practical consequence is that any exposure which kills the susceptible majority and spares a resistant minority enriches for resistance. Under-dosing and stopping early are dangerous for exactly this reason.

The Global Burden of Disease bacterial AMR study, published in The Lancet in 2024, estimated the toll across 204 countries, 22 pathogens and 84 pathogen-drug combinations.

Definition: Attributable counts deaths that would have been avoided had the infection been drug-susceptible. Associated counts deaths in people who had a resistant infection, whatever the immediate cause. The first is the stricter number, and the four-fold gap between them is why headline figures on this topic vary so much.

The WHO Bacterial Priority Pathogens List, updated in 2024, covers 24 pathogens across 15 families sorted into critical, high and medium priority. The critical tier is dominated by carbapenem-resistant Gram-negatives: carbapenem-resistant Acinetobacter baumannii, carbapenem-resistant Enterobacterales, and third-generation cephalosporin-resistant Enterobacterales.

This follows directly from the envelope you met in lesson one. A Gram-negative cell has an outer membrane the Gram-positives lack, and that membrane is a permeability barrier before any specific resistance gene is involved. A drug must cross it, usually through porin channels, and survive efflux pumps sitting immediately behind.

In practice: Losing a porin and upregulating a pump are cheap changes that need no new gene from outside, and together they can lift the MIC for several drug classes at once.

Three levers act on the mechanisms in this lesson rather than on the symptom.

1. Infection prevention removes transmission events, so a resistant strain that arises has fewer chances to reach a new host.
2. Diagnostics shorten the window of broad-spectrum therapy. A culture and MIC take a day or two; every hour of unnecessarily broad coverage is selection pressure on the whole flora.
3. Stewardship matches drug, dose and duration to the isolate, keeping concentrations above the MIC for long enough that the resistant minority is not simply handed the field.

Key idea: The same GBD analysis estimated that improving access to healthcare and to antibiotics could save around 92 million lives between 2025 and 2050. In much of the world the immediate problem is people not getting an effective antibiotic at all, so restriction alone is not the whole answer.