New weapons.

Drug-resistant bacteria are killing more and more humans. We need new weapons.

Scientists are testing futuristic tools to dismantle antibiotic resistance.

by Cecilia Butini

The first antibiotics ushered in a medical revolution a century ago, saving millions of lives every year from deadly infections and allowing doctors to safely perform many medical procedures we now take for granted — simple and complex surgeries, chemotherapy, intensive care. People are living decades longer in the age of antibiotics than they did at the beginning of the 20th century.

But bacteria that evade the current treatments are fueling a silent pandemic, contributing to nearly 5 million deaths worldwide every year, a toll expected to double in the next few decades unless humanity finds new ways to fight one of its oldest foes.

The World Health Organization (WHO) reported in late April that the “extensive” overuse of antibiotics during the pandemic may have exacerbated antimicrobial resistance, which was already on the rise. A group of US-based researchers also concluded that hospital-based infections resistant to last-resort antibiotics remain higher today than before Covid-19.

In the face of that threat, scientists are developing innovative techniques for attacking antibiotic-resistant bacteria. These can be as straightforward as finding faster ways to combine existing antibiotics or as pioneering as trying to synthesize antimicrobial molecules that were present in ancient animals like the woolly mammoth. Some investigators are studying ways to replace or supplement antibiotics altogether with microbe-eating viruses or nanosponges that act as vacuum cleaners for toxins. 

But beating antibiotic resistance everywhere requires more than technological advances. Experts say a global coordinated strategy, similar to the international efforts needed to address climate change, is necessary. Later this year, the United Nations General Assembly is expected to approve a declaration to mobilize the world against antimicrobial resistance and set targets for progress. 

A group of world leaders working on recommendations ahead of the September UN meeting recently predicted that the average lifespan worldwide could start dropping within the next decade as a result of antimicrobial resistance. A 2016 estimate by the Review on Antimicrobial Resistance said that 10 million people could die annually from resistant infections by 2050. 

Antimicrobial resistance, the group warned, “threatens a century of progress in human medicine.”

Bacteria are smarter than us, and we’re allowing them to be

Among the thousands of bacteria species circulating worldwide, seven are particularly deadly, including E. coli and the bacteria that cause tuberculosis and staph infections. These and others, like the bacterium behind the Black Death, have killed tens of millions of people over the centuries.

“Bacteria are smarter than us, in many ways,” Sujith Chandy, executive director of the International Centre for Antimicrobial Resistance Solutions (ICARS), told me. “We have been given more ability than them, but somehow we don’t use it.”

Their power lies in their ability to continuously evolve. Vanessa Carter, chair of the WHO Task Force of AMR (antimicrobial resistance) Survivors and founder of a patient-advocacy organization called AMR Narrative, experienced firsthand how dangerous these invisible organisms have become.

In 2004, she was in a car crash in Johannesburg, South Africa, that took her right eye and left her with a broken nose, damaged sinuses, and a shattered cheekbone, which had imploded when she violently hit her face on the car’s dashboard.

After a month in the hospital, she was discharged. Carter underwent a series of surgeries to reconstruct her face, including implants. But years after the accident, her face “still didn’t look right,” she said. 

In 2010, a surgeon inserted a new type of implant that would support her cheekbone and prevent her right orbital bone from collapsing. Carter realized soon after the procedure that something was wrong. There was pus oozing from the surgery site, an unmistakable sign of infection. 

Standard antibiotics failed to stop the infection, as did surgeries to clean up the prosthetic. After 11 months of ineffective treatment, her maxillofacial surgeon decided to remove the prosthetic entirely. An analysis revealed Carter had developed a MRSA (or methicillin-resistant Staphylococcus aureus) infection. 

Had the prosthetic not been removed, “I probably would have died of sepsis,” Carter told me. “I wouldn’t be here.”  

Once the prosthetic was taken out, the infection subsided with the help of vancomycin, an antibiotic usually given through an intravenous line to treat complicated infections. She was declared infection-free in 2013.

Staphylococcus aureus is one of three pathogens that cause the most antibiotic-resistant infections around the world, according to the Institute for Health Metrics and Evaluation, a population health research group based at the University of Washington. Almost 750,000 people died globally in 2019 from resistant staph infections, with South Asia and sub-Saharan Africa among the hardest-hit regions.

The misuse and abuse of antimicrobial drugs are driving the evolution of these drug-resistant pathogens, according to the WHO. But diagnostic capabilities outside of rich countries are frequently insufficient. Health care workers struggle to distinguish between viral, fungal, and bacterial infections, making it more difficult to prescribe the right antimicrobial if needed. Many people live hours away from the nearest doctor and medicines are often sold in informal settings without prescriptions.

As Chandy of ICARS told me, identifying bacteria requires laboratory infrastructure, which is costly and logistically challenging in nonurban or underserved areas. Existing field diagnostics are imperfect: They cannot always specify which bacteria are present, can be expensive to obtain, and aren’t always easy to use. What is needed are more precise testing tools that are also cheap, portable, and don’t require expertise to handle.

“The model is to have improved diagnostics that will tell you whether you do or do not need antibiotics, but at the moment, those do not exist; we have to recognize that,” said Michael Sharland, professor of pediatric infectious diseases at St. George’s Hospital Medical School at the University of London, and a global expert in antimicrobial prescribing. 

When people take antibiotics they don’t need, the drugs kill many but not all of the bacteria, good and bad, in their body. The result is natural selection in overdrive: The bacteria that survive because of antimicrobial-resistant properties multiply and spread; they can share those properties with other bacteria in the wild that have not been exposed to antimicrobial treatments. 

If those bugs end up infecting someone, the antibiotics they’re given may not work. 

Vaccines have been explored as a treatment for bacterial infections, but they have limits. They are currently available for only two of the seven deadliest drug-resistant bacteria. Pfizer tested a potential vaccine against Staphylococcus aureus but discontinued the study when it became clear that the shot wouldn’t be effective. Vaccines targeting the deadly Pseudomonas aeruginosa bacterium fizzled out in clinical trials. There are no vaccines available against E. coli or the bacterium that causes gonorrhea, which infects more than 80 million people a year.

With new treatments still years away, experts say the world should focus in the short term on restricting antibiotic use as much as possible, both to minimize the risk of further evolution and ensure that currently available antibiotics can be more equitably distributed. In realistic terms, “99 percent of the global population are never going to see a new antibiotic,” Sharland said. 

“In the long run, what we need to do is to match the amount of antibiotics that are used in [a] country to the amount of antibiotics that are needed in the country,” Sharland told me.  

Koen Pouwels, a senior researcher at Oxford Population Health’s Health Economics Research Centre, is building models to make that possible, but they are a work in progress. His team is developing statistical models to predict the effect of interventions like vaccines and diagnostics on deaths, hospitalizations, health care costs, and the overall economy.

But models are imperfect, Pouwels told me, because they inevitably rely on assumptions. His team’s model is one of the few in development, and their work is expected to be completed in a few years. 

Even with more responsible antibiotic use, experts say we need an active pipeline of new treatments that can act against difficult-to-treat bacteria. Scientists are utilizing artificial intelligence, nanotechnologies, and novel biotechnology in a bid to reinvigorate the antibiotic pipeline.

César de la Fuente, a professor of chemistry and founder of a lab at the University of Pennsylvania called Machine Biology Group, is one of the pioneers of two burgeoning fields: AI-based antibiotic discovery and molecular de-extinction. The existential threat to humanity posed by antimicrobial resistance and the lack of investment in developing new drugs inspired de la Fuente to explore how machines and biology could combine in the quest to discover new antibiotics. 

At first, the scientists trained a computer to recognize molecules from the natural world and predict which ones would have antimicrobial properties. This led to the discovery of an antimicrobial molecule from guava plants that was then physically synthesized in a lab and tested against different microbes. The molecule proved effective in mice, but it hasn’t yet been tested on humans. De la Fuente’s group is considering creating a company to move things forward and set up trials, and he said that different stakeholders have reached out with interest in the technology.

After their breakthrough with the guava plant, de la Fuente and his colleagues theorized AI could further accelerate antibiotic discovery. “This was motivated by the fact that it takes forever to even identify preclinical candidates, and it’s very expensive,” de la Fuente told me. The cost of developing an antibiotic can reach $1 billion and it usually takes 10 to 15 years. 

De la Fuente’s team decided to speed up their search for antibiotic candidates by looking beyond the human proteome, the full set of proteins expressed by an organism’s genome, to those of extinct organisms. They trained a deep-learning model to sift through a dataset of molecules and find those that most resembled an antibiotic, de la Fuente said.

Other scientists had already sequenced ancient creatures’ DNA and made that genetic data publicly available. In a new study, de la Fuente’s team developed a deep-learning approach that, in a matter of hours, can sift through the proteomes of every extinct organism known to science and look for peptides — or strings of amino acids that constitute the building blocks of proteins — with potential antimicrobial properties. The AI identified peptides from the woolly mammoth and the ancient sea cow, among other ancient animals, as promising candidates. According to the study, after being tested on mice, these peptides showed anti-infective activity.

Other groups of scientists are also betting that the rapid evolution of AI can match the rapid evolution of bacteria, serving as a catalyst for antibiotic discovery. Researchers from Stanford Medicine have used generative AI to develop recipes for antimicrobial molecules which can then be made in a lab and tested.   

Nevertheless, this work is still in its early stages. While it can identify promising candidates more quickly, those molecules still need to prove themselves in clinical trials that will take years. 

Drug discovery and approval is always a long process, but antibiotic development has slowed down in recent decades because major companies have dropped it to focus on more lucrative drug classes. Antibiotics are relatively cheap to make compared to other drugs. They also aren’t usually sold in large volumes, given that people will only take them occasionally — unlike with long-term treatments such as those for hypertension or diabetes. There isn’t a strong economic incentive for most companies to continue funding research and development for antibiotics, and those who are active in the space are usually small biotechnology enterprises. No new antibiotic classes have been discovered since the 1980s.

Because new drug discovery is so slow, combining existing antimicrobials offers more hope in the near term. Jeff Wang, professor of mechanical engineering at Johns Hopkins Whiting School of Engineering, said his team has designed a device to automate and scale the process of combining antibiotics and testing them against bacteria.

They built a mechanical platform that can pump and arrange thousands of antibiotic droplets. The platform is connected to a computer-controlled robotic arm that will pick up the droplets and quickly place them in a petri dish so they can interact with bacteria. A camera captures images of the petri dish, which is analyzed by an algorithm to determine whether certain bacteria stop growing when attacked by a specific antibiotic. In initial testing, the system, called RoboDrop, identified a new combination of three antibiotics that worked against an E. coli strain. 

Wang believes the invention has the potential to be easily scalable and affordable: The robotic arm is as big as a printer and can be purchased online. The camera work could even be done by a phone, he said. But according to Wang, it might take up to five years before RoboDrop becomes the kind of FDA-approved device that would be available to health care facilities. 

Hoping for something more effective than antibiotics

More than a century ago, scientists discovered the existence of viruses that could infect bacteria and, in effect, eat them. These remarkable entities were named bacteriophages — literally bacteria-eaters — or phages for short. 

The development of antibiotics sidelined bacteriophage use for decades, except in a handful of countries in the former Soviet Union and Eastern Europe. But the increase in antimicrobial resistance has prompted scientists to revisit phages as a potential alternative to traditional antibiotics.

Phages are currently used only as a last resort for serious resistant infections where doctors have identified the responsible bacterium and tested phages to see which one could neutralize the infection. They don’t yet exist as approved drugs that doctors can easily prescribe. 

Some companies and researchers believe, however, that phages could be commercially viable in the future if they prove effective at combating otherwise antibiotic-resistant bugs. 

Locus Biosciences, a North Carolina-based company, is developing phage cocktails that can be used against most existing bacteria, rather than relying on one phage at a time. Like so many other scientists working on antibiotic resistance, they are relying on robots and AI. 

Phages are naturally present in the environment and can be found in wastewater. Locus collects wastewater samples and runs them through a robot that searches for phages, Paul Garofolo, the company’s CEO, told me. Once phages are isolated, a predictive algorithm simulates interactions between millions of phages and pathogens to understand which can eat which. 

The company’s first phage cocktail to enter human trials targets E. coli in urinary tract infections, which affect millions of people each year and are becoming increasingly difficult to treat as bacteria have developed resistance to the existing drugs. The cocktail has already been tested on people in early-stage trials, and Locus began enrolling patients for a phase 2 trial this spring. The company anticipates FDA approval to come in 2028. 

Garofolo says the phage cocktails are going to be inexpensive enough that even less wealthy countries should be able to buy small batches to protect people from the most dangerous infections. Phage therapies can be priced reasonably low due to the wide availability of phages in nature and their ability to replicate to high concentrations, he said. One course of Locus phage therapy is expected to initially be priced in the range of $1,500 to $3,500, but the company plans to decrease the price as scale and volume increase. 

Investment in phage development could remain limited, however, until medical authorities develop clear clinical protocols for using them as an alternative to antibiotics. There isn’t a regulatory framework for phage treatment, more clinical data is still needed, and there isn’t certainty around that market’s size, said Jonathan Iredell, a University of Sydney infectious disease expert who is working with colleagues on phages.

Other scientists are drawing lessons from the human body itself to design new treatments. Whenever pathogens attack a body, the immune system will normally try and fight them by sending out substances to neutralize them. These substances produce an inflammatory response that is especially strong in cases of serious infections. A group of scientists from UC San Diego has developed a microscopic biological device called a nanosponge, meant to reduce inflammation and absorb the toxins freed into the body when an infection takes hold. 

Nanosponges act as decoys to mop up the harmful toxins and inflammation proteins that could otherwise wreak havoc, said Jessica Field, senior director of nonclinical development at Cellics Therapeutics, one company that is developing nanosponges. They are made from human cell membranes plus a nanoparticle and are meant to be used in conjunction with antibiotics in critically ill patients, Field explained. 

None of Cellics' products are available on the market yet and probably won’t be until clinical trials can prove their effectiveness to the satisfaction of federal regulators, which could take years. The most advanced nanosponge, currently in the first phase of human trials, could be used in patients who have pneumonia as a result of MRSA, the same infection Vanessa Carter developed when her face implant became infected. 

According to Field, nanosponges are an “agnostic” treatment, meaning their effectiveness at fighting infections isn’t tied to the presence of a specific bacteria. That could make them especially versatile. 

“I don't think that we can just keep producing different iterations of antibiotics, because bacteria are smart and are going to keep becoming resistant to whatever,” Field said. “So I think we need to take a step back and really have more of a broad approach to what we're doing.”

The excitement about all of these up-and-coming innovations should be kept in check, because most drugs that go into the initial phase of clinical testing do not move forward, Anthony McDonnell, a health economist and senior policy analyst at the Center for Global Development, said. 

Instead, “there needs to be global targets or some global system to reduce unnecessary [antibiotic] use, we need to do more infection control, because that stops people getting sick in the first place,” he told me.

The future holds promise. But people around the world need help today.