812 Talk: Examining microbial signal in pulmonary samples from a murine model of oxygen-induced lung injury

Seminar Details
Friday, November 13, 2020 - 12:00pm


Ph.D. Candidate, Dickson & Huffnagle Labs, Dept. of Microbiology & Immunology, Dept. of Internal Medicine- Pulmonary & Critical Care


Zoom Meeting:
Meeting ID: 944 6835 3155
Passcode: 340574
Mentors: Bob Dickson & Gary Huffnagle
Reviewers: A. Tai & C. Wobus

Background: In recent years, microbiome research has grown rapidly as the impact of microbial communities on human health and disease and associated mechanisms are increasingly understood. Using germ-free mice and antibiotic regimens in a murine model of oxygen-induced lung injury, our lab has previously shown that the microbiome plays a key role in the pathogenesis of acute lung injury. To determine the mechanisms by which resident microbiota influence acute lung injury pathogenesis, it will be crucial to characterize changes in local (lung) and distal (gut) microbial community structure. While methods for characterizing gut microbiota in mice are well-established, the optimal sampling strategy for characterizing murine lung microbiota has not been empirically determined.

Methods: We compared bacterial DNA from the lungs of healthy adult mice collected via two common sampling approaches: homogenized whole lung tissue and bronchoalveolar lavage (BAL) fluid. We quantified bacterial DNA using droplet digital PCR, characterized bacterial communities using 16S rRNA gene amplicon sequencing, and systematically assessed the quantity and identity of bacterial DNA in both specimen types, comparing bacteria detected in lung specimens to each other and to potential source communities: negative (background) control specimens and paired oral samples.

Results: By all measures, whole lung tissue in mice contained greater bacterial signal and less evidence of contamination than did BAL fluid. Relative to BAL fluid, whole lung tissue exhibited a greater quantity of bacterial DNA, distinct community composition, decreased sample-to-sample variation, and greater biological plausibility when compared to potential source communities. In contrast, bacteria detected in BAL fluid were minimally different from those of procedural, reagent, and sequencing controls.

Conclusions: Taken together, these results indicate that whole lung tissue is the preferred specimen type for murine lung microbiome studies. The scarcity of biologically plausible microbial signal in BAL fluid has important implications for the concurrent study of lung microbiota and injury in mice and necessitates careful experimental design moving forward. More broadly, these data highlight the necessity of including, sequencing, and analyzing ample negative controls in any low-biomass microbiome study, as they are crucial for biological interpretation regardless of specimen type.