Application of PTR-TOF-MS to investigate metabolites in exhaled breath of patients affected by coeliac disease under gluten free diet

Highlights

• First direct breath analysis of coeliac disease patients and controls by PTR-ToF-MS.

• Two precursor ions (H₃O⁺ or NO⁺) were used to generate breath fingerprint.

• Time resolution of 1 s and a mass range up to 371m/z are achieved.

• More than 500 peaks have been found in the patients’ breath.

• Coeliac disease patients under gluten free diet have same breath profile as controls.

Abstract

Coeliac disease (CD) is a common chronic inflammatory disorder of the small bowel induced in genetically susceptible people by the exposure to gliadin gluten. Even though several tests are available to assist the diagnosis, CD remains a biopsy-defined disorder, thus any non-invasive or less invasive diagnostic tool may be beneficial. The analysis of volatile metabolites in exhaled breath, given its non-invasive nature, is particularly promising as a screening tool of disease in symptomatic or non-symptomatic patients.

In this preliminary study the proton transfer reaction time of flight mass spectrometry coupled to a buffered end-tidal on-line sampler to investigate metabolites in the exhaled breath of patients affected by coeliac disease under a gluten free diet was applied. Both H₃O⁺ or NO⁺ were used as precursor ions. In our investigation no differences were found in the exhaled breath of CD patients compared to healthy controls. In this study, 33 subjects were enrolled: 16 patients with CD, all adhering a gluten free diet, and 17 healthy controls. CD patients did not show any symptom of the disease at the time of breath analysis; thus the absence of discrimination from healthy controls was not surprising.

Introduction

Coeliac disease (CD) is a chronic inflammatory disorder of the small bowel induced in genetically susceptible people by the irritant gluten and possibly other environmental cofactors [1]. The clinical manifestation of CD is very variable, mainly depending on the age at diagnosis. In fact affected individuals may show gastrointestinal symptoms, extra-intestinal symptoms or no signs of disease [2], making diagnosis difficult. The prevalence of CD has been estimated at approximately 0.5–1% globally [2]. CD is strongly associated with human leukocyte antigen (HLA) DQ2 and DQ8 haplotypes and is characterised by small bowel tissue damage with villous atrophy, cryptic hyperplasia and chronic inflammatory (mainly lymphocytic) infiltrates. Although serological screening is now widely used and the symptomatological response to gluten free diet (GFD) represent an important element to confirm a diagnosis, CD remains a biopsy-defined disorder; multiple duodenal biopsies are needed for a correct diagnosis [3], [4] and thus any non-invasive or less invasive diagnostic tool may be beneficial.

Comprehensive metabolomic approaches are typically based on hyphenated methods coupling chromatographic separation and mass spectrometry [5]. The need of high-throughput techniques, which is intrinsic to metabolomics, is however driving the development and application of more rapid, chromatography-free techniques that, on one hand, allow screening larger sample-sets and, on the other hand, reduce the possible artefacts caused by the extraction and concentration procedures [6].

Direct injection mass spectrometry (DIMS) for VOCs detection and quantification has recently being investigated and several methods have been proposed and applied in different fields such as environmental monitoring, food science and technology, health sciences [7] but little has been done in metabolomics and this only very recently [8], [9], [10]. A very promising DIMS method for VOC detection is proton transfer reaction-mass spectrometry (PTR-MS) [11] and in particular its recent version based on a time-of-flight (ToF) mass analyser (PTR-ToF-MS) that, while maintaining ultrahigh sensitivity and low limits of detection (parts per trillion by volume), improves rapidity and analytical information: a single spectrum can be obtained in a split second and in most cases the sum formula of the observed ion peaks can be determined [12].

The main idea at the basis of PTR-MS is the chemical ionization of VOCs having proton affinity higher than water by means of the reaction with primary hydronium ions (H₃O⁺). PTR-ToF-MS is characterized by a large dynamic range, since it is sensitive from the low pptv region (parts per trillion by volume) up to several ppmv: this aspect is very important for metabolomic applications involving metabolites whose abundance can span many orders of magnitude [5]. Furthermore the higher mass resolution of the ToF and the mass accuracy achieved helps the determination of the exact elemental composition of the measured peaks [12]. Analytical potential of PTR-MS increased notably by the recent coupling of PTR-MS instruments with systems that allow switching from H₃O⁺ to, for instance, NO⁺, O₂⁺, Kr⁺ and Xe⁺ [13], [14]. The different ionisation induced by different precursor ions allows, in some cases, the separation of isomeric compounds [14]. For example, ion products derived from reactions of NO⁺ and H₃O⁺ with aldehydes and ketones provide a possible means of distinguishing between isobaric molecules allowing their quantification as well [15].

PTR-MS has already been used for breath analysis, the first application in this direction goes back to 1994 [16] when PTR-MS was used to measure methanol, ethanol and acetone in human breath. Since then the possibility to use PTR-MS to measure and monitor the composition of breath as an indicator of the disease state of an individual has been increasingly explored. Several studies used this technique to study the exhaled breath of patients with lung cancer [17], [18], [19]. It was also used to show that elevated levels of exhaled nitric acid and hydrogen cyanide were present in the breath of patients with confirmed H. pylori infection [18]. Further examples can be found in dedicated reviews [20], [21], [22], [23].

Only one study explored the possibility of monitoring VOCs in exhaled breath to screen people for the presence of CD [24]. Based on the hypotheses “that malabsorption of dietary carbohydrates would lead to over production of alcohol fermentation products in the large intestine”, the level of several alcohols in the breath of 10 patients affected by CD with those of 10 healthy controls (Ctrl) was compared, but no significant difference was found [24]. In a recent study the VOC profiles in the exhaled breath of healthy subjects during 4 weeks GFD and after switching to a normal diet were compared, and 12 VOCs could be distinguished between the dietary regimes [25]. Among the volatile compounds, 2 alcohols (2-butanol and 2-propyl-1-pentanol) were reported. This finding contrasts Hryniuk and Ross [24] who report no significant differences for exhaled alcohols between CD patients on GFD and Ctrl.

PTR-ToF-MS has been used for breath analysis in a limited number of studies. Previously, PTR-ToF-MS was used for profiling exhaled breath of rats and found that VOCs were related to the diet and the physiological state of the animals [26]. Another study profiled exhaled breath of cirrhotic patients and found that VOCs were related to disease severity and correlated with some serum blood parameters [27].

Based on the results of previous investigations using PTR-ToF-MS for exhaled breath profiling in animal models [26] and in humans affected by liver cirrhosis [27], the aim of this study was to extend the screening of possible CD markers to VOCs other than alcohols (all those measurable with PTR-ToF-MS using H₃O⁺ or NO⁺, see material and methods for more details) and to verify the apparent incongruence between the two studies considering a GFD.