Just the Right Amount: Microbial Nutrients Drive Success and Failure of Antibiotics

Antibiotics are medical marvels that have transformed once deadly bacterial infections into manageable conditions. But with a rise in antibiotic resistance that renders existing treatments ineffective, new agents are urgently needed. Scientists at Caltech and Princeton University have now shed fresh light on why antibiotics that work well in laboratory tests often fail against real infections in humans.

By studying antibiotic and bacterial interactions in environments resembling those found in the body, they have revealed that microbial nutrients, such as glucose, play a crucial role in antibiotic efficacy. Their findings provide a unique framework for developing novel medications and investigating antibiotic resistance.

"We know that antimicrobial resistance against antibiotics is one of the biggest health issues of our time," says Sujit Datta, professor of chemical engineering, bioengineering, and biophysics at Caltech. "Our experiments have revealed antibiotic 'death fronts' that kill cells progressively rather than uniformly—with metabolically active surface cells dying while nutrient-starved interior cells survive."

According to Datta, this nutrient bottleneck was predicted decades ago but never observed experimentally until now. Seeing how it works opens new directions for developing successful antibiotics therapies.

Datta and his colleagues describe their experiments and observations of these death fronts in a paper in the journal Nature Communications. The lead author of the study is Anna Hancock, a postdoctoral scholar in Datta's group.

Typically, researchers test antibiotics in the lab by growing bacteria in a well-mixed liquid culture and then adding different concentrations of antibiotics to see when the bacteria stop growing. This establishes what is known as the minimum inhibitory concentration, or MIC, of the drugs.

"If you can identify the MIC, that tells you to use a larger concentration of the antibiotic to kill the bacteria," Datta explains. "But that's not what happens in reality." Instead, in humans and other organisms, bacterial communities present as spatially structured, or organized, non-mixed populations, and these are what Datta's lab has been studying. "To investigate antibiotic and bacteria interactions from a different angle, Anna used the experimental tools that we've developed in my group to essentially create a model bacterial community spread out in space."

The setup looks like this: Imagine a dish that is half-filled with bacteria cells in a gel developed by the Datta lab that resembles the extracellular matrix—a network of proteins and other molecules that regulate cell behavior and give structure to tissues—they inhabit in real communities. The other half of the dish has just that gel with no cells, which is meant to simulate the outside world. There is a divider between the two sides.

"Anna loads the side without bacteria cells with whatever antibiotics she's going to test as well as nutrients for the bacterial cells," Datta explains; many antibiotics require metabolic activity to kill cells, which the nutrients provide. "To start the experiments, she raises the divider. Now the antibiotics go in, the nutrients go in, and that mimics exposing a real bacterial population to those elements."

Using optical microscopy, the researchers can watch fluorescent signals from the cells to tell if they are dead or alive.

When the team first tested the experimental setup, they introduced strong levels of antibiotics without any nutrients. Nothing happened.

"The bacteria are still just sitting there, not dying," Datta says. "But when doing the exact same experiment, now with nutrients, you actually see a wave of death going through the population. You can see them dying successively from the outside in this really cool sweeping way, and that suggests that consumption of nutrients somehow plays a role."

But further experiments revealed a double-edged sword: While increasing nutrient supply improves the killing of bacteria, excess nutrients can unexpectedly promote regrowth of resistant subpopulations.

"One interesting and confounding feature of microbial populations is that they're heterogeneous, so not all the cells have the same susceptibility to the antibiotic," Datta explains. "What we saw is when this death front goes through, the antibiotic kills a lot of the cells, but there are some still left behind. If you've given too much nutrient, those cells can continue to grow in the wake of the death front, meaning you have resistant subpopulations that have arisen in the wake of an antibiotic exposure."

In addition to running experiments, the team came up with a simple mathematical model that captures the dynamics of nutrients diffusing into the bacterial communities, waking up the bacteria cells, and making them an active target for antibiotics.

"The model did a beautiful job of capturing what we saw in the experiments, making it possible to make and confirm predictions about when antibiotics would be able to kill the population and at what doses, how much of the population the antibiotics would be able to kill and how quickly, and how much nutrient was needed," Datta says. "That helps to put all this into a quantitative framework to begin to predict and control these processes."

Datta and his team say that while their findings won't immediately lead to new and better antibiotics, the work provides an important starting point for predicting which antibiotics will succeed in the human body and for investigating antibiotic resistance using a different approach.

"We have a new way to think about these problems that gives us a foundation to try to develop strategies for avoiding issues like resistant regrowth," Datta says. "The answer, as with all biology, is not simple, but that's what makes it interesting."

Additional authors of the paper, "A nutrient bottleneck controls antibiotic efficacy in structured bacterial populations," are Arabella S. Dill-Macky, Jenna A. Moore, Catherine Day, and Mohamed S. Donia of Princeton University. The work was supported by the National Science Foundation, the Camille Dreyfus Teacher-Scholar Program, the Pew Scholars Program in the Biomedical Sciences, the Eric and Wendy Schmidt Transformative Technology Fund, and the Princeton Catalysis Initiative.