There was a reminder in a recent edition of Nature (Dance, vol 632, 494) that pointed out that the current crisis with antibiotics leads to the requirement familiar to those who follow Lewis Carol – we have to run increasingly fast just to stay in the same place. There is plenty of blame to go around for why we are in this predicament, but that is pointless. We are here: live with it.

It is nearly a century ago the penicillin was discovered, and shortly after, a number of antibiotics were discovered from soil microbes, particularly from Actinomyces bacteria. For some time we were winning the war; microbial infections could easily be cured. But then the inevitable happened: unable to resist the “miracle drugs” they were badly misused. When I was a chemistry undergrad, I recall the lecturer being rather pleased to point out that most of the antibiotics then in use were basically lactams, and chemistry could play a myriad of substitution tricks so the bacteria would never win. That was the sort of hubris that played out. What eventuated was that the microbes developed resistance to the new variants relatively quickly, which made researching new variants a financial loser for the pharmaceutical companies. They claimed they could never recover the development costs before the drug became no longer desirable. Apparently in 2019, 1.27 million deaths world-wide could be attributed to drug-resistant infections. This is a serious problem. But there are possible routes for development.

The first is somewhat obvious. In the early days, the search was for broad-spectrum antibiotics: the wonder drug that would cure all infections. During the screening of natural products from the Actinomyces bacteria, some might have been rejected because they were too narrow spectrum, and only cured specific infections. These could be re-examined. In some ways, provided you know the bacterium, the specific cure has the obvious advantage that it will be much harder for bacteria to generate resistance. They are not exposed to antibiotics that are only partially successful against them.

A further route comes from the fact that previously antibiotics were developed from the few bacteria that are easily grown in the lab. The next strategy is to work out how to grow some of the other bacteria and hope we discover new antibiotics. A further possibility is a new method for testing whether bacteria are likely to contain antibiotics. This would accelerate the discovery process by a factor of about ten. But we still have to select bacteria that actually contain antibiotics that are useful.

Another possibility is to take a somewhat different approach. The question is, how did we overcome infections prior to antibiotics, and the answer is than many animal proteins and peptides have antimicrobial activity. Of course, the microbes have been exposed to those from humans as long as we have been here, so resistance is inbuilt. But what about extinct animals, such as the woolly mammoth? The current microbes may well have lost the ability to defend against these peptides. In answer to the point that we do not have a good supply of woolly mammoths, that is irrelevant. If we know the structure of the peptide, we can synthesise it, and we would always do that anyway. The reason we would need the mammoth flesh is to obtain the structure of desirable peptides. In trying to develop such peptides, AI is also useful. If it can isolate the features common to more than one peptide antibiotic, it might be able to design new smaller peptides that are easier to make and do the job. Apparently, a small number of such drugs are coming onto the market.

The next approach is to have a cocktail, hoping more than one would work synergistically. There are also non-antibiotics that might help, thus one of the compounds found in strawberries helps remove a protective film from bacteria. This won’t kill the bacterium but it might help other antibiotics to gain access. Finally, because many possibilities are somewhat bacterium specific, we need a rapid test to work out what causes the infection. If the physician has to wait a day or so for the lab to provide an answer, the patient may be in deep trouble.

So, overall, there are possibilities, but a lot more work needs to be done to convert them into practical uses.