Snooze button. By Sean McGrath. CC BY 2.0
Snooze button. By Sean McGrath. CC BY 2.0

Almost all of our antibiotics are designed to kill growing bacteria. If the bacteria don’t grow, they don’t die. Think of the microbes as bicycles and antibiotics as sticks. If a bicycle’s wheels are turning, a well-placed stick can do catastrophic damage. If the wheels are still, the stick passes harmlessly through. This strategy for beating antibiotics is called tolerance or persistence or dormancy. It has many names but a common theme: keep your head down, wait for danger to pass, and start growing again.

That’s exactly what the common gut bacteria Escherichia coli did when Ofer Fridman from the Hebrew University in Jerusalem exposed them to the antibiotic ampicillin. Even though he used very high doses, around 1 in 1,000 cells didn’t die. Fridman took these survivors and grew them afresh before subjecting them to the drug again. After ten cycles, around 1 in 10 cells could withstand the drug.

These bacteria weren’t resistant. They turned out to be just as sensitive to ampicillin as their first-generation ancestors. Instead, they had evolved tolerance. They spent more time in a dormant state before starting to grow and divide again, allowing them to sleep through the ampicillin blitz. And they adjusted this dormant interval—the lag time—with astonishing precision.

Fridman always exposed his E.coli to ampicillin bouts of the same length. Whatever the duration, the bacteria quickly evolved a timer to match. The cells that were subjected to three-hour bouts of ampicillin evolved an average lag time of 3.5 hours. Those that experienced five-hour bouts evolved a lag time of 5.1 hours. Eight hours of drug? Ten hours of lag. They didn’t just shut down—they shut down for exactly the right amount of time!

No one knows how the bacteria manage to fine-tune their growth so precisely, but the team has at least found clues about how they became tolerant. They sequenced the genomes of all the evolved tolerant bacteria, and found that they all had common mutations in six genes. The team then switched these mutated genes back to their original versions and found that three of them were responsible for tolerance.

Two of these are involved in cellular programmes that can ‘convince’ bacteria that there is no food around, and trigger a starvation mode that makes them delay their growth. The third is a mystery.

“It wasn’t clear whether dormant bacteria just happen to be there within bacterial populations, or whether this strategy of staying dormant could evolve,” says Nathalie Balaban, who led the study. Her team found clear evidence that the latter is possible, with a speed that Balaban found surprising. “Within 10 days of antibiotic treatment, such a strategy can evolve. It suggests that similar phenomena may happen in patients.”

If the bacteria that infect us are changing their lag times with similar precision, that’s a huge problem. Unlike resistance mutations, which allow bacteria to foil specific classes of antibiotics, lag-time tolerance is an all-purpose defence. It could potentially allow microbes to withstand all of our drugs.

Indeed, Fridman’s team showed that once E.coli evolved to tolerate ampicillin, it could tolerate a different antibiotic—norfloxacin—as well. And since tolerant bacteria get repeatedly exposed to the same drug, there’s every chance that they could evolve to be resistant too.

Can tolerance be beaten? That’s a tough question, says Balaban. It might help to keep antibiotic levels high for as long as possible, but that might just boost the emergence of resistance. It might be possible to interfere with the genes that fuel the evolution of tolerance.

Or perhaps we might do best to develop drugs that can kill dormant bacteria that aren’t growing. A few such drugs are being developed, including some that can punch through the outer walls of bacteria, causing their insides to spill out. But none of these are close to clinical approval yet. When it comes to antibiotics, good news is thin on the ground, while the hits just keep on coming.

Reference: Fridman, Goldberg, Ronin, Shoresh & Balaban. 2014. Optimization of lag time underlies antibiotic tolerance in evolved bacterial populations. Nature

More on persistence and tolerance: