As woolly mammoths and scimitar-toothed cats roamed what is now Canada 30,000 years ago, something with a surprising ability lurked beneath their feet. Frozen in the soil were bacteria that – defrosted millennia later – could fight off vancomycin, one of today’s antibiotics of last resort.
Antibiotic resistance is often portrayed as a modern crisis driven by antibiotic misuse. Yet its origins are much older. Many antibiotics weren’t dreamed up in the lab, but evolved in nature long ago to help microbes battle one another in the struggle for survival. Some resistance genes are equally as old, meaning the environment – even in places untouched by recent human activity – is both a sink and source for such genes.
Worryingly, that source is becoming more potent. A study published in April concluded that global warming is driving the spread of antibiotic resistance among bacteria in grassland soils. Factor in the possibility that our best drugs might not work against some of the resistance genes held by microbes now emerging from melting permafrost, and the future looks bleak.
But there is a silver lining. With modern gene-sequencing tools, we can analyse environmental samples in unprecedented detail, then use the information to identify reawakened resistance genes long before they can establish themselves in the bacteria responsible for some human diseases. “Instead of waiting until a resistance gene causes treatment failure in hospitals, we can begin to identify warning signals earlier,” says Kawther Zaher, a virologist at King Abdulaziz University in Saudi Arabia.
The approach could even herald a new era of medicine – one in which we future-proof antibiotics so they can overcome resistance that human pathogens have yet to evolve.
Antibiotic resistance is one of medicine’s greatest challenges. In 2019 alone, drug-resistant bacterial infections were directly responsible for an estimated 1.27 million deaths worldwide, and this figure has been forecast to rise to around 1.9 million deaths annually by 2050. It is no mystery how resistance arises. Every time we use an antibiotic, we impose a powerful selective pressure on microbial communities. Although most bacteria die from exposure to the drug, a few may survive because they have a natural genetic resistance. The survivors will multiply and the resistance genes they carry will become more common. Resistance will then spread.
Worse still, bacteria don’t have to rely on the standard process of inheritance to gain resistance genes. In the microbial world, snippets of DNA can be parcelled up and passed between neighbouring bacteria through a process called horizontal gene transfer. Bacteria can also leave the little parcels of DNA, known as plasmids, lying around in the environment to be picked up at a later date by a passing microbe. The donor and recipient don’t even have to belong to the same species. “It’s like a human being able to pass on some of their genes to an ant,” says Michael Bottery, an evolutionary microbiologist at the University of Manchester, UK.
This exchange has produced what scientists call the resistome: a vast reservoir of antibiotic-resistance genes that exists throughout the natural world. In 2006, Vanessa D’Costa, a microbiologist now at the University of Ottawa in Canada, led one of the first studies to examine the resistome in urban, agricultural and forest soils. Her team found high levels of resistance to established antibiotics, as well as newly introduced drugs. “The level and diversity of resistance we observed was incredibly shocking to us,” she says.
Many of us might assume the resistome is a product of the antibiotic era. But D’Costa and her colleagues have convincingly demonstrated that it has a much deeper history. In a 2011 study, they reported the discovery of genetic elements that encoded resistance to vancomycin in 30,000-year-old permafrost samples collected in the Yukon territory of Canada.
Since then, similar discoveries have been made in other frozen environments, from glaciers across the Himalayas to Romania’s 5000-year-old Scărișoara ice cave. These findings suggest that glaciers, permafrost and ice caves harbour rich, previously isolated reservoirs of antibiotic resistance.
The 2011 study made headlines around the world, but to many researchers, the discovery wasn’t wholly unexpected. They were already aware that microbes began evolving and using small molecules to fight their rivals long before Alexander Fleming discovered penicillin in 1928. “The gene clusters that program the biosynthesis of these complex, small molecules are thought to have evolved in nature several hundred thousand to over hundreds of millions of years ago,” says D’Costa.
Antibiotic resistance has similarly ancient roots, predating not just modern medicine, but humanity itself. Microbes have long been fending off ancient bacterial threats, for example by altering the molecular targets the small molecules latch on to.
A deeper appreciation that the resistome is ancient and can exist in places untouched by modern medicine has fundamentally changed how researchers think about antibiotic resistance. Traditionally, environmental surveillance focused on tracking the supposedly one-way spread of resistance genes from hospitals, farms and sewage into the wider environment. Now, we know it is a two-way street. “Researchers now look at resistance genes not only as clinical markers of treatment failure, but as part of a wider ecological and evolutionary gene pool,” says Zaher.
We are also learning that the resistome isn’t the same everywhere. Last year, a team led by Jabir Thajudeen, a polar microbial ecologist at India’s National Centre for Polar and Ocean Research, published a study comparing resistance genes in Arctic and Antarctic soils recently exposed by glacial retreat. “Sampling glacier foreland soils means going to the ice margin itself – freshly exposed ground that may have been under ice for centuries,” says Thajudeen.
Despite their extreme histories, the soils contained resistance genes. But the resistome in these polar soils had little in common with those found in non-polar soils or the resistance genes found in hospital settings. Even the resistome in Arctic soils differed from that in Antarctic soils.
Thajudeen says the frigid environments at high latitudes seem to do more than preserve resistance genes: they actively shape how resistance evolves. “[Resistance genes] appear to have been evolving in isolation, under unique selective pressures such as cold, low nutrients and UV exposure, for a very long time.”
As ice retreats due to climate change, more of these soils are being uncovered, presenting an opportunity for their microbial communities to reawaken and thrive – and for the unique and ancient resistance genes they carry to spread and multiply.
“These forelands are natural time capsules,” says Thajudeen. “Soil that’s been sealed under ice, sometimes for millennia, is now being exposed as glaciers retreat. Our data provide a baseline snapshot of what’s stored in that ancient soil before it’s further disturbed.”
But he and his colleagues made another, more concerning discovery: the genetic samples containing the greatest number of resistance genes also tended to contain more of the parcels of DNA that allows these genes to move between bacteria. As warming accelerates, “more of these reservoirs will be exposed and begin interacting with modern microbial communities, wildlife and, eventually, human systems”, he says.
To make matters worse, “bacteria tend to grow better at elevated temperatures”, says evolutionary biologist Mato Lagator, also at the University of Manchester. “They divide more, meaning they generate more mutations that can result in [antibiotic] resistance.”
What’s more, any resistance genes – either ancient varieties previously locked away in frozen soils or newly evolved ones – may be increasingly likely to spread. That is because the rate at which bacteria share DNA through horizontal gene transfer also rises with temperature. This was one of the key conclusions of the April study linking climate change to the spread of antibiotic resistance in grassland soils.
In brief, then, rising global temperatures may accelerate the two main routes by which bacteria acquire resistance – mutation and horizontal gene transfer – potentially making the already challenging problem of antibiotic resistance far harder to tackle.
But this concerning research contains a message of hope. With their growing understanding of the resistome, particularly in freshly exposed polar soils, researchers are developing new strategies for tackling antibiotic resistance. Doing so involves analysing bacterial DNA in the environment to understand which resistance genes may become a problem in the years ahead, essentially creating an early-warning system.
The tools for running this system have only recently become available. A few decades ago, environmental surveillance could detect just a handful of known resistance genes. More recently, techniques have been developed to analyse all the DNA in an environmental sample. But although these metagenomic tools can identify many more resistance genes, they don’t give researchers a clear sense of how the genes are moving through microbial communities.
The picture has been transformed with the development of increasingly sophisticated computational technology. Using these tools, researchers can now establish which environments favour particular genes, and which of those genes are most likely to ultimately pass into human pathogens. “This has changed the resistome from a static list of genes into a dynamic ecological and evolutionary network,” says Zaher. “That is a major advance.”
For instance, it is now possible to determine where in a microbial genome a particular resistance gene lies. If it sits near genes that are known to have a high chance of becoming mobile and moving between bacteria through horizontal gene transfer, the chances are that the resistance gene may move and spread too. If, instead, the gene is in a stable region of the genome, it may be less likely to spread. Used in this way, environmental monitoring is no longer simply about documenting resistance, but about identifying the resistance genes that are most likely to become future clinical threats.
Identifying new resistance genes isn’t always straightforward. One promising approach is to compare the DNA in samples against vast databases of known resistance genes. Deep-learning tools such as DeepARG, which was released by a team at Virginia Tech in 2018, can identify distant relatives of those known resistance genes. The relatives can then be carefully analysed to see if they, too, are resistance genes – ones that we have never encountered before. But useful though this is, such tools are still limited to identifying resistance genes that are broadly similar to those we have already discovered.
Another machine learning tool named DRAMMA, developed by a team including David Burstein at Tel Aviv University in Israel and unveiled last year, doesn’t have this limitation. “The advancement here is finding genuinely new antimicrobial-resistance genes,” says Burstein. DRAMMA identifies broader biological signatures of antibiotic resistance – for example, the properties of the protein a gene produces, whether the gene occurs in distantly related bacteria and whether it lies beside other resistance genes or mobile DNA. This allows DRAMMA to predict entirely new resistance genes hidden within environmental DNA.
“We expect that tools like DRAMMA will shift surveillance from reacting to threats already causing clinical problems to catching [new] resistance genes while they’re still confined to agricultural or environmental reservoirs, before they reach a pathogen that may risk human lives,” says Ella Rannon, also at Tel Aviv University, who co-led the tool’s development.
These methods might reshape how we develop new antibiotics, says D’Costa. The hope is that researchers could use environmental surveillance to guide the design of drugs that can combat newly discovered resistance genes so the drug continues to work, even if those genes eventually find their way into human pathogens. “Carefully selecting antibiotic candidates against the known and predicted resistance repertoire could widen the window in which a new treatment stays effective,” says Rannon.
Better drug design isn’t the only possible approach. Some wonder whether we can prevent resistance genes crossing into pathogens at all. “Interrupting the ability of bacteria to pass on these genes is really promising,” says Bottery. “It’s about stopping resistance from getting into places that you don’t want it to be.”
One way to do this is by blocking horizontal gene transfer between bacteria. Chemical compounds capable of inhibiting plasmid transfer are now being explored, says Bottery. “But there’s a really exciting option here,” he says. “Viruses.”
Viruses that infect bacteria, known as bacteriophages, can be used to target the tiny surface structures that bacteria rely on to exchange plasmids. “These are like needles or rods that allow the passage of a plasmid from one bacterium to another,” says Bottery. By infecting a bacterium carrying a potentially problematic resistance gene, a bacteriophage can stop the gene from being passed on.
Emerging evidence suggests the approach might be effective. For instance, in a recent study that has yet to be published in a scientific journal, researchers in China and the UK used the strategy to break the chain of resistance gene transmission through the gut bacteria of chickens, while leaving most of the surrounding microbial community unaffected.
“There really is a big push to try and prevent these transmission networks of resistance,” says Bottery. “Resistance will inevitably evolve. But if we can prevent it from spreading from one bacterium to another, we can preserve essential treatments for much longer.”
Promising though all this research is, there are no guarantees that new drugs will remain useful for decades to come. “We have essentially set up a global evolutionary experiment,” says Bottery. “By using vast amounts of antibiotics, we’ve massively increased the selective pressure for bacteria to acquire resistance genes that were already present in environmental reservoirs.”
But these reservoirs, particularly those in the frozen, ancient soils found towards the poles, could also be our salvation, offering an evolutionary playbook that might help us predict antibiotic resistance and interrupt its spread. It is a resource that could be invaluable for medicine. And it’s one we can’t afford to squander.