Our study - "Enrichment of the lung microbiome with gut bacteria in sepsis and the acute respiratory distress syndrome" - is the product of three fortunate convergences.
The first convergence was of my two jobs. When I'm not in the lab researching the lung microbiome, I'm in the intensive care unit caring for patients with critical illness. I've often thought about my patients in the ICU - bombarded by antibiotics and so vulnerable to infection - and wondered how their microbiota are altered by their disease and by our therapies, and in turn how their disturbed microbiota may be making them sicker. Though the molecular techniques we use are relatively new, the question our team is asking is an old one: what role do the body's bacteria play in a patient's progression from infection to sepsis to multi-organ failure? The field has known for more than half a century that if we suppress gut bacteria with antibiotics, we can protect critically ill animals and patients from sepsis, organ dysfunction and death. But the mechanism behind this effect - translocation of gut bacteria via the blood or lymphatics, a change in the virulence of gut bacteria, or suppression of secondary infections - is unsettled. Our group, along with others, has spent the past half decade sorting out the ecological determinants of the lung microbiome in health and chronic lung disease, so we felt well-equipped to ask how these ecological forces change acutely in critical illness.
The second fortunate convergence was a collaboration with my friend and colleague, Ben Singer. Ben, another critical care physician at the University of Michigan, studies why many of our ICU patients suffer from cognitive dysfunction after recovering from their acute illness. Several years ago, when Ben and I were both still in training as clinical fellows, we starting talking about how our research approaches could complement each other for a project like this: his animal modeling of sepsis, my ecological analysis of bacterial communities in the lung. With the trusting support of our mentors (Ted Standiford and Gary Huffnagle, respectively), Ben and I launched the series of experiments that ultimately comprised this study. From the start of this project, I've felt that we need to do more than merely describe the microbiome in sepsis and ARDS: we need to use the microbiome as a tool to interrogate the pathophysiology of critical illness. Though this study is only our first step in that direction, I think we benefited from anchoring our experimental approach on well-established animal models with decades of characterization.
The third and final convergence was between our team's bench science and our institution's clinical research tradition. Mice aren't humans, and the mouse microbiome sure isn't the human microbiome. As excited by I was by the results of our mouse experiments, I'm humbled by the long list of promising animal observations in sepsis and ARDS that have ultimately proven irrelevant to human disease or its treatment. Thankfully, my clinical research colleagues at Michigan have been studying humans with ARDS for decades, and they've had the foresight to collect and characterize specimens from patients along the way. The human lung lavage and blood specimens we used in this study were obtained in a human ARDS trial performed a decade ago, years before the lung microbiome was on anyone's mind. It's going to take years for us to prospectively validate and explain our experimental findings in humans, but I think that our inclusion of these banked clinical specimens - even given the limitations of retrospective analysis and clinical confounding - gave our findings a strong slug of human plausibility. I'm lucky to work daily with investigators who - individually and collectively - straddle the bench-bedside divide.