Collecting and Analyzing Breath from Intubated Mouse Models

Introduction

Exhaled breath contains volatile organic compounds (VOCs) that originate from biological processes in the body and therefore have the potential to be utilized as biomarkers for clinical use. Breath biomarkers have been identified across a broad range of disease areas (1,2), meaning they are promising for the broader adoption of next-generation non-invasive diagnostic and monitoring tools. However, the translation of these biomarkers into breath tests for clinical use remains limited, partly due to the lack of consistent methodologies and quality controls across the breath research literature. A reliable animal model breath analysis method for pre-clinical studies would be advantageous to address these challenges.

Establishing a breath analysis method using mice in a controlled lab setting as a substitute for human breath can reduce the variability and challenges seen in human studies and allow for a better understanding of breath as a sampling medium. This ultimately expedites the identification and validation of breath biomarkers for clinical use.

No studies previously have directly compared baseline breath VOC profiles between mice and humans using the same analytical method. Given the number of VOCs detectable in human breath, it is likely that there are compounds contained in mouse breath that could be detected in studies for better translatability to human biology. It is essential to develop sampling and analytical methods capable of providing a sufficient signal-to-noise ratio in mouse breath analysis. A recent study by Taylor et al (3) developed a method for accurately characterizing the VOCs in the breath of healthy intubated mice and compared the VOCs contained in mouse breath to system blank samples and VOCs found in human exhaled breath.

Methods

Mouse breath sampling

Male C57BL/6JRj mice were used in this study. The mouse breath sampling system was developed by modifying the commercial flexiVent® FX1 small animal ventilator. The study connected an ambient filter and a flexiVent filter to the ventilator with flexible tubing and a sorbet tube for breath collection. 

Figure 1. A schematic showing the mouse breath sampling system used in this study.

A total of 15 breath samples were collected from the intubated mice over four days. The mice were ventilated for 45 minutes for VOC breath collection, with a deep lung inflation procedure performed every two minutes during this period, resulting in approximately 1.8 L of breath obtained. For optimized untargeted analysis, a clean background is necessary to maximize the signal-to-noise ratio, so that VOCs of interest can be identified more easily. A total of 15 system blank samples were collected immediately before mouse breath collection. System blank collection was performed using the same procedure used for mouse breath sampling. Ambient blanks were collected at the start and end of the day in the room where breath collection was performed, to track environmental contamination.

Human breath sampling

A total of 13 healthy volunteers were recruited and each participant provided a single breath sample which was collected using the ReCIVA® Breath Sampler, and a paired system blank sample was collected at the same time. This allows for on-breath compounds to be distinguished from those present in the system background.  

Breath analysis

Mouse and human breath samples were both analyzed using the Breath Biopsy® OMNI® analysis method, including high-resolution thermal desorption gas chromatography and accurate mass spectrometry (TD-GC-MS). 

Data analysis

Three metrics have been used previously to distinguish on-breath and background VOCs and were used in this study to evaluate the performance capabilities of the breath sampling method. These metrics are type 1 = standard deviation, type 2 = paired t-test, and type 3 = receiver operating characteristic area under the curve, ROC-AUC. 

Results

Identifying on-breath compounds

A total of 472 molecular features were identified in mouse breath. To distinguish between compounds that are on-breath from those originating from the background air, the three metrics were used. Using all three metrics 21 on-breath compounds were identified using Type 1, 30 using Type 2, and 56 using Type 3. 

There is a good degree of overlap between the compounds identified between the three metrics, leaving a range of 15-66 on-breath VOCs identified in mice depending on how stringently a definition is set. The combined metrics maximize the opportunity for discovering potentially informative VOCs that can be considered as on-breath, as well as provide higher confidence in the results. 

Comparing mouse data to human breath data

To assess the transferability of the mouse breath VOCs to human VOCs, the same three metrics were used to quantify the number of on-breath VOCs in human breath. A total of 49 on-breath VOCs were common to both human and mouse breath. Most of these compounds were associated with the gut microbiome or were plant-derived. Examples of these include trimethylamine (TMA) and dimethyl sulfone. Certain compounds appear to be specific to mouse or human breath, for example, methyl nitrate and 2-butanol appeared to be exclusively on-breath in mice.

Figure 2. (A) Trimethylamine (TMA) is an example of an identified on-breath compound in both humans and mice that is significantly different from the background signal. (B) Dimethyl sulfone is on-breath in both humans and mice (Note log axis). (C) Methyl nitrate and (D) 2-butanol are on-breath compounds in mice, but not in humans.

Discussion

Using animal models for breath biomarker discovery work allows a more controlled study, reducing the variability from factors such as individuals, diet, and other environmental factors. The reduced risk of false biomarker discovery accelerates the process of identifying potential biomarkers for validation in clinical trials.

Existing literature has demonstrated the feasibility of murine models for breath VOC research, though the number of studies is limited. Our findings indicate that three exogenous VOCs often associated with environmental sources, including xylene, toluene, and styrene, were not identified as on-breath compounds in this study. This suggests that the breath collection method and analytical approach used in this study enabled the exclusion of these environmental compounds.

We compared the potential translatability of the on-breath VOCs identified in mice with those found in human breath. More VOCs were classified as on-breath in humans across all metric types compared to mice. This is not unexpected due to the relative body size difference, the difference in lung volume, and the differences in environmental exposure between mice and humans.

Despite the differences, 49 on-breath VOCs were found to be common between humans and mice, with a portion of these compounds originating from the microbiome. This is important as these volatile compounds are targets for biomedical research into the intricate interplay between the microbiota and disease. The notable presence of certain common on-breath VOCs in mice and humans implies the potential to investigate disease-relevant volatile compounds for human clinical studies in mouse models. This method therefore supports a promising avenue for translating the findings from pre-clinical mouse models to clinical human studies.

References

1. Ratiu IA, Ligor T, Bocos-Bintintan V, Mayhew CA, Buszewski B. Volatile Organic Compounds in Exhaled Breath as Fingerprints of Lung Cancer, Asthma and COPD. J Clin Med. 2020 Dec 24;10(1):32. doi: 10.3390/jcm10010032
2. Malderen KV, Winter BYD, Man JGD, Schepper HUD, Lamote K. Volatomics in inflammatory bowel disease and irritable bowel syndrome. eBioMedicine [Internet]. 2020 Apr 1 [cited 2023 Oct 9];54. Available from: https://www.thelancet.com/journals/ebiom/article/PIIS2352-3964(20)30100-6/fulltextdoi: 10.1016/j.ebiom.2020.102725
3. Taylor A, Blum S, Ball M, Birch O, Chou H, Greenwood J, et al. Development of a new breath collection method for analyzing volatile organic compounds from intubated mouse models. Biology Methods and Protocols. 2024 Jan 1;9(1):bpae087. doi: 10.1093/biomethods/bpae087