Our group is interested in bringing sequencing technology to the bedside to improve the rapid diagnosis of infections, and have previously demonstrated the potential of real-time metagenomics to identify respiratory pathogens in a clinically relevant timeframe. Among the barriers to this goal is that we can now generate metagenomic data faster than our bioinformatic tools can make sense of it. A problem with many specimens (e.g., respiratory) is the astronomical host:bug ratio. In metagenomic sequencing results from sputum or bronchoalveolar lavage fluid, human DNA overwhelms microbial DNA by a ratio of >99.9:00.1, and you burn all of your time and computational resources classifying human sequences.

For several years, we’ve been working with Jenna Wiens and Meera Krishnamoorthy (Computer Science & Engineering) to address this problem using machine learning. We’re excited to share AMAISE: A Machine Learning Approach to Index-Free Sequence Enrichment. This tool uses machine learning to perform "in silico host depletion," so you can jump faster to microbial classification. It quickly and accurately identifies and excludes host-derived sequences… no need for time-intensive alignment of human sequences.

AMAISE: a machine learning approach to index-free sequence enrichment

Manuscript (Communications Biology)

Download AMAISE (GitHub)

Meera’s blog post about AMAISE

Rishi Chanderraj MD

Lab member Rishi Chanderraj has accepted a faculty position in the Division of Infectious Diseases at the University of Michigan. He will be an Assistant Professor, and in July will start at the Ann Arbor Veterans Affairs Hospital. Rishi will continue to study the role of the gut microbiome in ICU infections and clinical outcomes. Congrats, Rishi!

Robert Dickson will deliver a “State of the Art” lecture at this year’s Aspen Lung Conference. His talk will be on June 8th and will be titled “The Role of Gut and Lung Microbiota in Calibrating Lung Immunity.”

Link to event website

At the American Thoracic Society International Conference, lab members won the following awards:

• Robert Dickson: Mid-Career Peer Recognition Award (Pulmonary Infections and Tuberculosis Assembly)

• Rishi Chanderraj: Abstract Scholarship

• Reid McCallister: Abstract Scholarship

Congrats!

Everyone did a spectacular job sharing their science at ATS. So proud of all of them!

Just published online in the Blue Journal: “Therapeutic Targeting of the Respiratory Microbiome,” a Perspective piece by Robert Dickson and other leaders in the study of the respiratory microbiome.

In the past decade, we’ve learned the the respiratory microbiome is 1) detectable in health, 2) altered in disease, 3) variable across patients, 4) correlated with lung immunity, 5) predictive of disease outcomes, and 6) modified by in-use therapies (e.g. antibiotics and oxygen). All of this makes the respiratory microbiome an attractive potential “treatable trait” for future clinical trials.

But what will it take in order for us to modulate the respiratory microbiome therapeutically? In this essay, we explore what is known and what would need to be known before we can achieve this goal.

Manuscript: Therapeutic Targeting of the Respiratory Microbiome (American Journal of Respiratory and Critical Care Medicine)

We said goodbye to Brooke and Danny, and debated whether beavers know what they’re doing.

Robert Dickson has been appointed Associate Chief for Research for the Division of Pulmonary and Critical Care Medicine at the University of Michigan. He will work closely with MeiLan Han (Division Chief) to advance the Division’s research mission, including team science initiatives, faculty recruitments, and collaborations with other investigators from around the University.

Just published in PLOS ONE: a new study by Kale Bongers and our group teasing out how much of the effects of antibiotics on murine metabolism are due 1) to changes in gut microbiota vs 2) behavioral changes (e.g. food and water aversion due to bad-tasting antibiotics.

A common approach in murine microbiome studies is to put a cocktail of antibiotics (often ampicillin, metronidazole, vancomycin, neomycin) into their drinking water and ascribe all biologic consequences to perturbation of the microbiome. But this overlooks important off-target effects of antibiotics, including food and water avoidance. Metronidazole especially is extremely bitter. Your mice are thus potentially dehydrated and starved at the time of your measurements or experimental exposures.

In this study, we systematically studied the effects of multiple antibiotic regimens (both oral and systemic) on mice: their food and water consumption, their body composition (via NMR), their organ- and tissue-specific metabolism, and their gut bacterial density and community composition.

We found that mice are even more avoidant of water when it contains metronidazole than when it contains a commercial bitterant (denatonium benzoate, the stuff they put antifreeze and animal deterrents and nail-polish). Much of the metabolic consequences previously reported in that common four-drug regimen (ampicillin, metronidazole, vancomycin, neomycin) may actually be due to dehydration and starvation. But we did find that other enteral regimens (cefoperazone, enrofloxacin/ampicillin) can effectively deplete gut bacteria without causing nearly as much food and water avoidance. Interestingly, systemic antibiotics (intraperitoneal ceftriaxone) also result in decreased food and water consumption, suggesting that the microbiome plays a role in behavior (food and water aversion) independent of the direct aversive effects of taste.

The gut microbiome is an important and overlooked player in systemic metabolism, both in health and critical illness. Kale’s study gives us a firm methodological footing for future work determining how the microbiome modulates tissue-specific and organismal metabolism.

Manuscript: Antibiotics cause metabolic changes in mice primarily through microbiome modulation rather than behavioral changes (PLOS ONE)

Kale Bongers MD PhD

Robert Dickson will be speaking on the gut microbiome’s role in critical illness in the Dudrick Symposium as part of the 2022 ASPEN Nutrition Science and Practice Conference in Seattle. The talk will be between 8am and 10am on Monday, March 28.

Link to event website

Oxygen is (arguably) the most commonly given drug in the ICU. It can be life-saving for patients with respiratory failure, but it can also be a poison: healthy mice exposed to >95% oxygen will uniformly die within 5 days. Our group previously showed that oxygen alters respiratory and gut bacteria, and the microbiome plays a role in oxygen-induced lung injury. Despite the ubiquity of oxygen in clinical practice, we still don’t know how to dose it.

In this just-published study, Sanjeev Tyagi (co-mentored by Robert Dickson and Mike Sjoding) set out to determine: 1) what level of arterial oxygenation is associated with harm, 2) how common is this harmful hyperoxemia in ICU patients, and 3) what are the clinical predictors of harmful hyperoxemia? Sanjeev studied >2,000 mechanically ventilated patients with >33,000 (!) arterial blood gases. He found that 1) there is no relationship between arterial oxygen and mortality below 200 mmHg, but a linear positive relationship above this threshold, 2) there is no duration threshold below which this relationship is not seen (i.e. there doesn’t appear to be a minimum duration below which PaO2 > 200 is safe), 3) more than half (55%) of mechanically ventilated patients have at least one PaO2 above this threshold, 4) more than 1 in 7 patients (13.1%) were exposed to PaO2 > 200 on multiple days of their ICU stay, and 5) the ICU of admission was the strongest predictor of severe hyperoxemia. ”If patients spent an entire day exposed to PaO2 > 200 mmHg, they had 2.19 (95% CI 1.33 – 3.60, p = 0.002) greater odds of 30-day mortality in an adjusted analysis.”

Several recent RCTs have looked at conservative vs liberal oxygen dosing, yet none of them include patients exposed to arterial hyperoxemia in the range that unambiguous causes harm in humans and animals (PaO2 > 200). Yet in real-world ICU patients, harmful hyperoxemia is incredibly common (>55% of all patients) and persistent (>13% on multiple days). There is no legitimate clinical indication for a PaO2 > 200 mmHg, and abundant evidence that it is harmful. Perhaps the field should shift from trying to finding an optimal “dose” of oxygen within the safe range and instead treat harmful, persistent hyperoxemia as an unacceptable “never event” that should be studied like other quality-improvement issues.

Manuscript: Outcomes and Predictors of Severe Hyperoxemia in Patients Receiving Mechanical Ventilation: A Single-Center Cohort Study (Annals of the American Thoracic Society)

Sanjeev Tyagi

Robert Dickson was a member of the American Thoracic Society workshop dedicated to defining acute lung injury in experimental models. The workshop report was just published, and will be presented and discussed at the American Thoracic Society International Conference in May.

Manuscript: Update on the Features and Measurements of Experimental Acute Lung Injury in Animals: An Official American Thoracic Society Workshop Report (American Journal of Respiratory Cell and Molecular Biology)

Just published in Microbiome: a new study from our group, led by Rishi Chanderraj, studying the methodological, biological, and clinical importance of bacterial density in rectal swabs obtained from hospitalized patients.

Imagine comparing two cities (say, Chicago and Ann Arbor), limiting your comparison to relative demographics (age, gender, race), and ignoring that one is >20 times the population size of the other. We do this all the time with gut microbiome studies: we compare community composition and diversity, and don’t bother to measure and report bacterial density.

In this new study, we quantified the bacterial density in rectal swabs collected from 118 hospitalized patients. We found that bacterial density matters methodologically (it correlates with background contamination in low-biomass swabs), biologically (it correlates with the relative abundance of bacteriocin-producing taxa, and clinically (it is predictably influenced by antibiotic exposure and is predictive of extra-intestinal infections in this at-risk population).

Population density is a fundamental ecologic features of communities (both microbial and otherwise). We should stop overlooking it in microbiome studies.

Manuscript: The bacterial density of clinical rectal swabs is highly variable, correlates with sequencing contamination, and predicts patient risk of extraintestinal infection (Microbiome)

Rishi Chanderraj