When a patient is suffering respiratory failure, such as can occur with COVID-19, breathing oxygen-enriched air may save their life, but there’s a chance it may also damage their lungs. The cause of the injury, at least in part, is an oxygen-induced shift in the balance of bacterial species in the lung, according to the results of a study in Science Translational Medicine today (August 12).
“It’s an important paper in establishing that there is a role of the microbiome in hyperoxia-induced lung injury,” says pulmonologist Alison Morris of the University of Pittsburg who was not involved in the research. The study “opens the door to looking more closely at the impact [of the microbiome] and how we can modulate it” for lung therapies, she continues.
“It’s a really neat set of experiments. They’ve taken a clinical conundrum, delved into human samples and then they’ve [followed up with] really elegant mouse models and mechanistic studies,” adds lung disease researcher Philip Molyneaux of Imperial College London.
Oxygen, an element so vital for life, is well known to be toxic and potentially lethal to animals when given at high concentrations. Concerns exist about oxygen’s widespread medical use too. The gas is given to many patients with low blood oxygen levels but it has been linked to lung injury and pneumonia.
“We’ve known about the damage caused by hyperoxia for a long time,” says Leopoldo Segal, a lung microbiome researcher and critical care specialist at New York University who did not participate in the research. “It is a major issue . . . particularly pertinent during this COVID pandemic, because [the patients] are extremely refractory to oxygen and so they stay on high doses of oxygen for a very long time.” It’s a balancing act between the known danger, Segal says, and the reality that “the patients would be in trouble if they were at lower oxygen concentrations.”
How oxygen causes damage and why some patients fair worse than others was unclear. Robert Dickson of the University of Michigan and colleagues suspected bacteria might be to blame. The lungs are populated with a community of microbes, Dickson explains, and “oxygen is very important for bacteria” in all sorts of ecosystems, not just the body. Oxygen affects which ones will grow well and which won’t. In the lungs, he continues, “you have aerobes, you have anaerobes, so we had a strong intuition that oxygen would be a main driver of the lung ecology.”
You can easily imagine that you can start intervening . . . where you try to modify the existing microbiota in the setting of hyperoxia to decrease the harm.—Leopoldo Segal, New York University
To test their hypothesis the researchers examined the bacterial content of lung fluid specimens procured from more than 1,500 patients between 2015 and 2018 who had been placed on a ventilator for at least 24 hours and exposed to low (21–43 percent) medium (43–55 percent) or high (over 55 percent) oxygen. They found that, while the likelihood of detecting Staphylococcus aureus—the bacterium most often linked with ventilator-associated pneumonia—and Pseudomonas aeruginosa—another lung bacterium—was comparable in the low oxygen samples, S. aureus was detected twice as much as P. aeruginosa in the high-oxygen specimens.
In healthy mice, exposure to hyperoxia also resulted in altered microbiome compositions, including increases in the proportions of Staphylococcus species, the team showed. These microbial population shifts preceded the occurrence of inflammation (shown by increases in detectable cytokines) and of lung injury (shown by elevated levels of leaked protein in lung fluids).
While these results hinted that oxygen-induced microbiome changes cause injury, the team’s next experiment was the kicker, says Dickson. It showed “that if you take the microbiome out, you use germ-free mice that have absolutely no bacteria on or in their body, and you give them hyperoxia, they are protected. They don’t get leaky lungs. . . . Under a microscope the lung tissue looks normal.”
This finding, Dickson says, “argues that one way or another, the microbiome is actually playing a causal role in this lung injury.”
“Up till now, we knew that we were causing damage, but there was nothing from a therapeutic standpoint that we have discovered that can prevent the damage,” Segal says. Now that there is a mechanism, he says, “you can easily imagine that you can start intervening . . . where you try to modify the existing microbiota in the setting of hyperoxia to decrease the harm.”
More experiments would be necessary to figure out if and how such microbial manipulations would work, says Segal, but antibiotics might not be sufficient. The team showed that two regimens of antibiotics given to the hyperoxia-treated animals did nothing to protect them from lung damage, while a third drug worsened the lung damage.
While clinical practice is unlikely to change on the strength of the paper’s results—after all, patients in respiratory failure still need oxygen and often need antibiotics—they do reinforce the message that “oxygen is a drug and the administration of any drug needs to be done appropriately,” says Molyneaux. “When you overdose someone with a drug, there are problems.”
S.L. Ashley et al., “Lung and gut microbiota are altered by hyperoxia and contribute to oxygen-induced lung injury in mice,” Sci Transl Med, 12:eaau9959, 2020.