Determining the presence of asthma-related molecules and salivary contamination in exhaled breath condensate

Abstract

Background: Researchers investigating lung diseases, such as asthma, have questioned whether certain compounds previously reported in exhaled breath condensate (EBC) originate from saliva contamination. Moreover, despite its increasing use in 'omics profiling studies, the constituents of EBC remain largely uncharacterized. The present study aims to define the usefulness of EBC in investigating lung disease by comparing EBC, saliva, and saliva-contaminated EBC using targeted and untargeted mass spectrometry and the potential of metabolite loss from adsorption to EBC sample collection tubes.

Methods: Liquid chromatography mass spectrometry (LC-MS) was used to analyze samples from 133 individuals from three different cohorts. Levels of amino acids and eicosanoids, two classes of molecules previously reported in EBC and saliva, were measured using targeted LC-MS. Cohort 1 was used to examine contamination of EBC by saliva. Samples from Cohort 1 consisted of clean EBC, saliva-contaminated EBC, and clean saliva from 13 healthy volunteers; samples were analyzed using untargeted LC-MS. Cohort 2 was used to compare eicosanoid levels from matched EBC and saliva collected from 107 asthmatic subjects. Samples were analyzed using both targeted and untargeted LC-MS. Cohort 3 samples consisted of clean-EBC collected from 13 subjects, including smokers and non-smokers, and were used to independently confirm findings; samples were analyzed using targeted LC-MS, untargeted LC-MS, and proteomics. In addition to human samples, an in-house developed nebulizing system was used to determine the potential for EBC samples to be contaminated by saliva.

Results: Out of the 400 metabolites detected in both EBC and saliva, 77 were specific to EBC; however, EBC samples were concentrated 20-fold to achieve this level of sensitivity. Amino acid concentrations ranged from 196 pg/mL - 4 μg/mL (clean EBC), 1.98 ng/mL - 6 μg/mL (saliva-contaminated EBC), and 13.84 ng/mL - 1256 mg/mL (saliva). Eicosanoid concentration ranges were an order of magnitude lower; 10 pg/mL - 76.5 ng/mL (clean EBC), 10 pg/mL - 898 ng/mL (saliva-contaminated EBC), and 2.54 ng/mL - 272.9 mg/mL (saliva). Although the sample size of the replication cohort (Cohort 3) did not allow for statistical comparisons, two proteins and 19 eicosanoids were detected in smoker vs. non-smoker clean-EBC.

Conclusions: We conclude that metabolites are present and detectable in EBC using LC-MS; however, a large starting volume of sample is required.

Keywords: Amino acids; Asthma; EBC; Eicosanoids; LC-MS; Leukotriene; Lung; Metabolomics; Proteomics; Saliva.

Fig. 1

Experimental setup and total ion chromatograms (TIC) of exhaled breath condensate (EBC) controls. a Setup of the control experiment using a syringe and syringe pump to administer control unspiked water and a control spiked water sample through the TURBO-DECCS EBC collection apparatus; b TIC of unspiked water. * indicates a contaminant peak at 0.933 min which was putatively identified as propiolic acid (8.69 ppm error, 82.8 score), c TIC of spiked water; d TIC of a ‘clean’ EBC sample; e Visualization of EBC samples collected from volunteers showing the clean EBC from non-droolers, compared to EBC collected from droolers which show the presence of saliva in the samples

Fig. 2

Experimental setup and recoveries of leukotrienes using various exhaled breath condensate (EBC) collection devices. a EBC simulation device connected to a syringe pump, showing the plastic tube and condenser; b Glass (left) versus plastic (right) EBC tube; c RTube spike recovery comparison for cysteinyl leukotrienes; d TURBO-DECCS spike recovery comparison for leukotrienes. PET: polyethylene terephthalate

Fig. 3

Overlap of metabolites detected in clean EBC, saliva-contaminated EBC, and saliva samples of healthy volunteers in Cohort 1. Untargeted metabolomics was performed on EBC and saliva from healthy volunteers. Metabolite peaks were extracted using MassHunter Profinder software (Agilent). Samples were filtered using a 3000 abundance cutoff and a presence in at least two of the three sample groups. A total of 77 metabolites were determined to be unique to EBC

Fig. 4

Concentrations of ten eicosanoids in saliva and EBC samples of matched asthmatics subjects in Cohort 2. Quantitative analysis was performed on an Agilent triple quadrupole (QQQ) 6410 mass spectrometer using targeted multiple reaction monitoring (MRM); concentration units in pg/mL; blue circles are saliva samples (n = 106), red triangles are EBC samples (n = 107); black line is sample mean

Fig. 5

Distribution of compounds across EBC groups in Cohort 3. a Venn diagram depicting the overlap of metabolites in four categories based on subject from the untargeted metabolomics analysis. Metabolites were filtered for presence in at least two out of the four groups; b Hierarchical clustering of 172 metabolites present in at least two EBC groups. Blue sections indicate low metabolite abundances and red sections indicate high abundance levels. The healthy EBC subjects appear to have a majority of lower abundance metabolites compared to the other three groups; c Concentration levels of four eicosanoids detected in all four sample groups using targeted LC-MS. Samples were analyzed on a triple quadruple mass spectrometer