The quest for new antibiotics is going back to the Stone Age.
The urgency to identify possible candidates has never been greater as the global population faces nearly 5 million deaths every year that are associated with microbial resistance, according to the World Health Organisation.
A research team led by bioengineering pioneer César de la Fuente is using artificial intelligence-based computational methods to mine genetic information from extinct human relatives such as Neanderthals and long-gone ice age creatures such as the woolly mammoth and giant sloth.
READ MORE: Matthew Perry's cause of death revealed
The scientists say some of these small protein, or peptide, molecules they have identified have bacteria-fighting powers that may inspire new drugs to fight infections in humans.
The innovative work also opens up a completely new way to think about drug discovery.
"It has enabled us to uncover new sequences, new types of molecules that we have not previously found in living organisms, expanding the way we think about molecular diversity," said de la Fuente, Presidential Assistant Professor at the University of Pennsylvania, where he heads the machine biology group.
"Bacteria from today have never faced those molecules so they may give us a better opportunity at targeting the pathogens that are problematic today."
The approach may seem to come out of left field, but experts say that new ways of looking at the problem of antimicrobial resistance to existing medicines, a deadly and pressing problem for global health, are sorely needed.
"The world is facing an antibiotic resistance crisis. My view is that a land, sea, and air approach is needed to solve the problem — and if we need to go to the past to provide potential solutions for the future — I am all for it," said Michael Mahan, a professor in the department of molecular, cellular and developmental biology at the University of California, Santa Barbara.
He wasn't involved in the research.
Antibiotics and where their alternatives may come from
Most antibiotics come from bacteria and fungi and have been discovered by screening microorganisms that live in soil.
But in recent decades, pathogens have become resistant to many of these drugs because of overuse.
Scientists engaged in the global fight against superbugs are exploring different potential weapons, including phages, or viruses created by nature to eat bacteria.
Another exciting avenue of research involves antimicrobial peptides, or AMPs, which are infection-fighting molecules produced by many different organisms — bacteria, fungi, plants and animals, including humans.
AMPs have a broad range of antimicrobial properties against different pathogens such as viruses, bacteria, yeast and fungi, Mahan said.
While most traditional antibiotics work by focusing on a single target in a cell, antimicrobial peptides bind to and disrupt a bacterial membrane at many places, he added.
It's a more complicated mechanism that consequently may make drug resistance less likely, but, because of the molecules' potential to disrupt cell membranes, it can also result in increased toxicity, according to Mahan.
There are a handful of peptide-based antibiotics in clinical use, such as colistin, which is made from a bacteria-based AMP. It's used as a drug of last resort to treat certain bacterial infection because it can be toxic, Mahan said.
One human AMP known as LL-37 has also shown potential.
Other promising AMPs have been found in unexpected places: pine needles and the blood of the Komodo dragon.
A 'Jurassic Park' moment
De la Fuente had been using computational methods for the past decade to assess the potential of a wide range of peptides as alternatives to antibiotics.
The idea to look at extinct molecules came up during a lab brainstorm when the blockbuster movie "Jurassic Park" was mentioned.
"The notion (in the film) was to bring back entire organisms, and obviously, they had a lot of issues," he said.
His team started thinking about a more feasible idea: "Why not bring back molecules from the past?"
Advances in the recovery of ancient DNA from fossils mean that detailed libraries of genetic information about extinct human relatives and long-lost animals are now publicly available.
To find previously unknown peptides, the research team trained an AI algorithm to recognize fragmented sites in human proteins that might have antimicrobial activity.
The scientists then applied it to publicly available protein sequences of modern humans (Homo sapiens), Neanderthals (Homo neanderthalensis) and Denisovans, another archaic human species closely related to Neanderthals.
The researchers then used the properties of previously described antimicrobial peptides to predict which of their newly identified ancient counterparts had the most potential to kill bacteria.
Next, the researchers synthesised and tested 69 of the most promising peptides to see whether they could kill bacteria in petri dishes.
The team selected the six most potent — four from Homo sapiens, one from Homo neanderthalensis and one from Denisovans — and gave them to mice infected with the bacterium Acinetobacter baumannii, a common cause of hospital-borne infections in humans.
"I think one of the most exciting moments was when we were resurrecting the molecules in the laboratory using chemistry and then we were bringing them back to life for the first time. And so it was really cool from a scientific perspective to have had that moment," de la Fuente said of the research that published in August in the scientific journal Cell Host & Microbe.
In infected mice that developed a skin abscess, the peptides actively killed the bacteria; in those that had a thigh infection, the treatment was less effective but still halted the growth of bacteria.
"The best (peptide) was what we call Neanderthalien 1, which comes from Neanderthals. And that was the one that was most effective in the mouse model," de la Fuente said.
He cautioned that none of the peptides were "ready to go antibiotics" and would require a lot of tweaking.
More important, he says, is the framework and tools his team has developed to identify promising antimicrobial molecules from the past.
In research expected to publish next year, de la Fuente and his colleagues have developed a new deep-learning model to explore what he describes as the "extinctome" — the protein sequences of 208 extinct organisms for which detailed genetic information is available.
The team found more than 11,000 previously unknown potential antimicrobial peptides unique to extinct organisms and synthesised promising candidates from the Siberian woolly mammoth, Steller's sea cow (a marine mammal that was wiped out in the 18th century by Arctic hunting), the 10-foot-long (3-meter) Darwin's ground sloth (Mylodon darwinii) and the giant Irish elk (Megaloceros giganteus).
He said that the peptides they discovered displayed "excellent anti-infective activity" in mice.
"Molecular de-extinction offers a unique opportunity to combat antibiotic resistance by resurrecting and tapping into the power of molecules from the past," he said.
A wacky but worthwhile approach
Dr. Dmitry Ghilarov, group leader at the John Innes Centre in the United Kingdom who studies peptide antibiotics, said the real bottleneck in the search for new antibiotics wasn't necessarily a lack of promising compounds, but getting pharmaceutical companies to develop and clinically test potential peptide antibiotics, which can be unstable and difficult to synthesise.
He was not involved in the research.
"I don't see an immediate reason to look at paleo proteomes. We have already … have a lot of these peptides," he said.
"What we really need in my view is deep understanding on the underlying … principles: what makes the peptide bioactive to be able to design them."
"There are a lot of these peptide antibiotics which were not developed and pursued by the industry because of difficulties like toxicity," Ghilarov said.
According to a paper published in May 2021, of 10,000 promising compounds identified by researchers, only one or two antibiotic drugs reached US Food and Drug Administration approval.
Dr. Monique van Hoek, a professor and associate director of research at George Mason University's School of Systems Biology in Fairfax, Virginia, said the idea of molecular de-extinction was "a really interesting approach."
She was not involved in either study.
Van Hoek said it was rare that a peptide found in nature — be it extinct or from a living organism — would directly lead to a new type of antibiotic or other drug.
More often, she said, the discovery of a new peptide will offer a starting point for researchers, who could then use computational techniques to tinker and optimize the peptide's potential as a drug candidate.
Van Hoek's research currently focuses on a synthetic peptide inspired by one found naturally in the American alligator. The peptide is currently undergoing preclinical testing.
"So far it's going really well. And that's exciting because many other peptides that I've worked on over the years fail for one reason or another," she said.
Van Hoek said that while it may appear wacky to look at alligators or extinct humans for a new source of antibiotics, the magnitude of the crisis makes the approach worthwhile.
De la Fuente agreed. "I think what we need is as many new and different approaches as possible, and that will increase our chances of being eventually successful," he said.
"I think we can find a lot of potential useful solutions by looking behind us."
