Feasibility of Early Detection of Acute Pulmonary Exacerbations by Exhaled Breath Condensate Metabolomics

XIAOLING ZANG; MARÍA EUGENIA MONGE; NAEL A. MCCARTY; ARLENE A. STECENKO; FACUNDO M. FERNANDEZ

St. Louis

Conferencia; 63rd ASMS Conference on Mass Spectrometry and Allied Topics; 2015

American Society for Mass Spectrometry

IntroductionProgressive lung function decline in cystic fibrosis (CF) often does not proceed in a linear fashion, rather is punctuated by acute pulmonary exacerbations (APEs). The frequency of APEs severe enough to require hospitalization is a crucial factor that may contribute to death from CF. APE diagnosis remains challenging due to changes in patient symptoms over time, as observed by the attending clinician. Reliable measurements to predict oncoming APEs would enable early clinical intervention and potentially prevent loss of lung function.Herein we address this issue using ultra performance liquid chromatography-quadrupole-time-of-flight mass spectrometry and a supervised classification model to profile and identify metabolites in exhaled breath condensate (EBC) samples, thereby facilitating discrimination of pre-APE patient samples from stable CF patient samples.MethodsEBC samples from 4 pre-APE patients (CF patients 1 to 3 months before an APE) and 19 stable CF patients (EBC samples from CF subjects who are clinically stable without an APE for ≥3 months) were profiled using a Waters ACQUITY Ultra Performance liquid chromatograph coupled to a Xevo G2 quadrupole-time-of-flight mass spectrometer (QTof-MS). Chromatography was performed using a Waters ACQUITY UPLC BEH C18 column. Samples were lyophilized over 24 hours and then concentrated by a factor of 20 via reconstitution in methanol and water (10:90 v/v).Mass spectra were acquired in negative ionization mode in the 50-1500 m/z range. Metabolic features were extracted using Progenesis QI Version 2.0. Orthogonal partial least square-discriminant analysis (OPLS-DA) was used for sample classification.Preliminary DataA total of 389 spectral features (Rt, m/z pairs) were quantitatively detected in EBC metabolome profiles from the cohort. Out of the 389 spectral features, 174 were present in at least 50% of pre-APE or stable CF patient samples. Twenty-seven out of these 174 spectral features had tentative endogenous identities after searching through databases. Reverse interval PLSDA feature selection was performed on the 27 tentatively identified spectral features, which resulted in the selection of 8 discriminant features. The 8 discriminant features tentatively identified comprised a metabolic profile that distinguished the 4 pre-APE EBC samples from the 19 stable EBC samples with 100% accuracy, specificity and sensitivity using the OPLS-DA model. Three of the key differentiating metabolites were tentatively identified as lactic acid, pterins and their derivatives, and eicosanoids and their metabolites. These 3 metabolites were all elevated in pre-APE patient EBC samples, compared to stable CF patient EBC samples. Interestingly, lactic acid has been reported to be elevated in the bronchoalveolar lavage fluid (BALF) of CF patients, relative to control subjects. Furthermore, there is evidence to suggest that the pH of EBC collected from children with CF and asthma is more acidic than the pH of EBC collected from control subjects. Novel AspectHigh classification accuracy suggests this approach is a promising tool for the reliable prediction of the onset of an APE.