Home
Search
Collections
Journals
About
Contact us
My IOPscience
Comparative analyses of volatile organic compounds (VOCs) from patients, tumors and transformed cell lines for the validation of lung cancer-derived breath markers
This content has been downloaded from IOPscience. Please scroll down to see the full text. 2014 J. Breath Res. 8 027111 (http://iopscience.iop.org/1752-7163/8/2/027111) View the table of contents for this issue, or go to the journal homepage for more
Download details: IP Address: 132.203.227.61 This content was downloaded on 11/06/2014 at 10:27
Please note that terms and conditions apply.
Journal of Breath Research J. Breath Res. 8 (2014) 027111 (13pp)
doi:10.1088/1752-7155/8/2/027111
Comparative analyses of volatile organic compounds (VOCs) from patients, tumors and transformed cell lines for the validation of lung cancer-derived breath markers Wojciech Filipiak 1,2 , Anna Filipiak 1,2 , Andreas Sponring 1,2 , Thomas Schmid 3 , Bettina Zelger 4 , Clemens Ager 1,2 , Ewa Klodzinska 5 , Hubert Denz 1,6 , Alex Pizzini 6 , Paolo Lucciarini 3 , Herbert Jamnig 6 , Jakob Troppmair 7,8 and Anton Amann 1,2,8 1
Breath Research Institute of the University of Innsbruck, A-6850 Dornbirn, Austria Univ.-Clinic for Anesthesia and Intensive Care, Innsbruck Medical University, A-6020 Innsbruck, Austria 3 Department of Visceral, Transplant and Thoracic Surgery, University Hospital Innsbruck, A-6020 Innsbruck, Austria 4 Institute of Pathology, Innsbruck Medical University, A-6020 Innsbruck, Austria 5 Institute for Engineering of Polymer Materials and Dyes, Torun, Poland 6 Department of Pneumology, University Hospital Innsbruck, A-6161 Natters, Austria 7 Daniel Swarovski Research Laboratory, Department of Visceral, Transplant and Thoracic Surgery, Innsbruck Medical University, A-6020 Innsbruck, Austria 2
E-mail: [email protected] and [email protected] Received 12 December 2013, revised 14 March 2014 Accepted for publication 25 March 2014 Published 27 May 2014 Abstract
Breath analysis for the purpose of non-invasive diagnosis of lung cancer has yielded numerous candidate compounds with still questionable clinical relevance. To arrive at suitable volatile organic compounds our approach combined the analysis of different sources: isolated tumor samples compared to healthy lung tissues, and exhaled breath from lung cancer patients and healthy controls. Candidate compounds were further compared to substances previously identified in the comparison of transformed and normal lung epithelial cell lines. For human studies, a breath sampling device was developed enabling automated and CO2-controlled collection of the end-tidal air. All samples were first preconcentrated on multibed sorption tubes and analyzed with gas chromatography mass spectrometry (GC-MS). Significantly (p < 0.05) higher concentrations in all three types of cancer samples studied were observed for ethanol and n-octane. Additional metabolites (inter alia 2-methylpentane, n-hexane) significantly released by lung cancer cells were observed at higher levels in cancer lung tissues and breath samples (compared to respective healthy controls) with statistical significance (p < 0.05) only in breath samples. The results obtained confirmed the cancer-related origin of volatile metabolites, e.g. ethanol and octane that were both detected at significantly (p < 0.05) elevated concentrations in all three kinds of cancer samples studied. This work is an important step towards identification of volatile breath markers of lung cancer through the demonstration of cancer-related origin of certain volatile metabolites.
8
Authors to whom any correspondence should be addressed.
1752-7155/14/027111+13$33.00
1
© 2014 IOP Publishing Ltd
Printed in the UK
W Filipiak et al
J. Breath Res. 8 (2014) 027111
Keywords: lung cancer, breath analysis, tumor tissue, cancer cell lines, breath markers, adsorptive preconcentration, GC-MS (Some figures may appear in colour only in the online journal) to obtain a survey of candidate compounds, which may detect a transformed phenotype, i.e. LC. For this purpose we analyzed headspace samples of lung tumors and healthy lung tissues as well as breath samples from LC patients and healthy controls for VOC production. The data generated were further compared to the results of our in vitro experiments with LC cell lines [41–44].
1. Introduction The six most common cancers in humans are those of lung, breast, colon, rectum, stomach and prostate, with the highest mortality seen in lung cancer (LC) [1]. Although benign tumors can be cured by surgical resection [2], tumor size, type, location and associated co-morbidities predispose only 20– 30% of patients suffering from lung cancer to such treatment [3, 4]. An extensive study confirmed that most patients with stage III or IV lung cancer at the time of diagnosis time did not undergo surgery but instead were more likely to receive radiation treatment or even no treatment [5]. Worldwide, the most common method for lung cancer screening is chest x-ray (CXR) alone or in combination with cytology (sputum, bronchoscopy or biopsy). More recently, computed tomography (CT) [6–9] and positron emission tomography (PET) [10, 11] are also used. Though both imaging techniques have impressive sensitivity, which translates into a reduction in lung cancer deaths in the screened population [12–14], they also lead to the detection of many benign lung nodules and false positive results [15]. Furthermore, CT and PET techniques are expensive and require experienced operating personnel, and additional concerns about radiation exposure associated with frequent CT imaging persist [16]. Therefore, the development of an early diagnostic method which, due to its simplicity, non-invasiveness and low cost, eventually could be used by a regular physician is of utmost importance. Exhaled breath analysis has great potential for noninvasive diagnosis [17–19] as exemplified by the urea test for detection of Helicobacter pylori infection [20, 21] or the nitric oxide test for asthma monitoring [22, 23]. Particular effort has been made to develop breath analysis for the diagnosis of LC [24–32]. Numerous substances have been proposed as potential breath markers of LC [33, 34], but the clinical relevance for many of them is questionable [35]. Apart from strictly technical reasons, such as differences in breath sampling procedures or analytical techniques, an important issue is the limited knowledge about the origin of these compounds with regard to cellular sources (cancer cell, immune cells, infectious agents, etc) and biochemical pathways. Due to metabolic activity of organs like the liver, the kidneys or the lungs themselves, substances once released from an organ or a tumor site into the bloodstream may be further converted, resulting in an altered profile, when measured in exhaled air or other body emanations. It has been demonstrated recently that comparing volatile organic compounds (VOCs) from various sources yields complementary information with regard to the human volatolome [36, 37]. In our research approach we combine the comparative breath analysis of LC patients and normal subjects [37], ex vivo experiments with human lung tissues, in vitro experiments with normal and transformed cells as well as microorganisms frequently causing lung infections [38–40]. Here we describe our efforts
2. Materials and methods 2.1. Patient cohort and sample collection in ex vivo study with lung tissues
The lung tissues were collected from 14 patients: 8 women (amongst which 4 were active smokers) and 6 men (all nonsmokers) with different stages of LC (table 1) [45]. In all cases, tissue specimens were taken during segmentectomy and the sampling procedure is depicted in figure 1. From each patient undergoing surgery, tumor and healthy tissues were collected. Unfixed gross specimens were carefully examined by palpation and serial laminations. Grayish white, indurate tissue, macroscopically representing malignant neoplasia was taken in duplicates for our studies, and at least half of the tumor tissue was formalin fixed and paraffin embedded for diagnostic hematoxylin/eosin (H&E) staining and immunohistochemical evaluation according to an established classification scheme [45]. The same procedure was done with unaffected lung tissue for control samples whereby the amount of both tissue specimens (tumor and healthy) were collected in preferably the same amounts, mostly 2–3 g (dependent on tumor size). Immediately after surgical excision, specimens were rinsed in Ca2+ and Mg2+ free Dulbecco PBS buffer (Sigma-Aldrich, Steinheim, Germany) to remove blood, and carefully dried with a sterile swab prior to enclosing them in a gas-tight 20 ml glass vial. To minimize tissue necrosis, all glass vials containing specimens were kept in an ice-water bath at 0 ◦ C during transport while the adsorption of tissue headspace was done at 37 ◦ C and the entire procedure was done within approximately 30 min after the surgical resection of the specimen. Informed consent was obtained from all participants and the study was approved by the Ethics Committee of the Innsbruck Medical University. 2.2. Adsorptive preconcentration of headspace of lung tissues
For the dynamic headspace sampling a continuous flow (5 ml min−1) of synthetic air of purity 5.5 (Linde, Stadl-Paura, Austria) additionally purified on SupelcarbTM hydrocarbon trap (Bellefonte, PA, USA) was purging through the vial containing the specimen. Usage of air instead of nitrogen reduces the tissue necrosis, hence additionally minimizes the release of VOCs from dying cells. Apparently 100 ml of tissues’ headspace was adsorbed on the multibed sorption tube. To prevent excessive water uptake on carbon molecular sieves, 2
W Filipiak et al
J. Breath Res. 8 (2014) 027111
(a)
(b)
(c)
(d)
(f)
(e)
Figure 1. Procedure of lung tissues preparation and headspace sampling. All lung specimens were collected during segmentectomy (a). From the large tumor mass resected (b), healthy and tumor tissue were carefully separated by a pathologist (c), tissues were rinsed in buffer to remove blood (d), subsequently transferred to a tightly closed glass vial (e) and transported to the laboratory on ice. A custom-made system for VOC preconcentration ( f ) used a continuous flow (5 ml min−1) of purified synthetic air through the glass vial containing lung tissue with simultaneous sample dilution (1:6) with purified nitrogen to decrease the sample’s relative humidity to prevent excessive adsorption of water on multibed tube. VOC from the tissues’ headspace adsorbed on multibed sorption tubes were ultimately analyzed by GC-MS. Table 1. Lung tumor staging [45] and smoking habits of lung tissue donors.
Nr
Gender
Age
Smoking status
Cancer group
Cancer type
Stage of lung cancer
N
M
Tumor staging
1 2 3 4 5 6 7 8 9 10 11 12 13 14
Female Male Female Female Female Male Female Female Male Male Male Female Female Male
68 65 47 59 69 52 63 61 78 72 80 78 62 59
Non Non Active Active Non Unknown Active Active Non Non Non Non Ex Non
NSCLC NSCLC NSCLC NSCLC NSCLC NSCLC NSCLC NSCLC NSCLC NSCLC NSCLC NSCLC SCLC Metastasis to the lungs
Adenocarcinoma Adenocarcinoma Adenocarcinoma Adenocarcinoma Adenocarcinoma Large-cell carcinoma Large-cell carcinoma Squamous-cell carcinoma Squamous-cell carcinoma Squamous-cell carcinoma Squamous-cell carcinoma Squamous-cell carcinoma Small-cell lung cancer Sarcoma
1 2 2 2 3 2 3 2 2 2 3 1 1 4
0 0 0 0 1 0 2 0 0 0 2 0 0 1
0 0 0 0 0 0 0 0 0 0 0 0 0 1
pT1 pT1 pT1 pT1 pT3 pT2 pT4 pT1 pT1 pT2 pT2 pT1 pT1 n.a.
at an elevated temperature of 40 ◦ C to avoid condensation of water and depletion of polar analytes.
the sample relative humidity was reduced by dilution with dry nitrogen, as proposed by others [46, 47]. For this purpose, the headspace sample was mixed in a ratio of 1:6 with a nitrogen of purity 6.0 (i.e. 999 999%, Air Liquide Germany GmbH, Krefeld, Germany) additionally purified on a Carboxen 1000 trap. All transfer lines within the sampling system were kept
2.3. Patient cohort and sample collection in breath analysis
A total of 36 LC patients and 28 healthy individuals were recruited for this study (table 2). All LC patients and healthy 3
W Filipiak et al
J. Breath Res. 8 (2014) 027111
Pump model M2K3, SCHEGO Schemel & Goetz GmbH & Co KG, Offenbach, Germany) creating an underpressure drawing breath gas into the sampling bag. Importantly, the pump is placed outside the path of sample flow, preventing contamination or loss. (iii) An alternative sampling mode offers breath collection when the expired CO2 exceeds 80% of its level in a previous exhalation. This is required for subjects with breathing difficulties (older persons, patients suffering from advanced COPD) with gradually decreasing CO2 levels in subsequent exhalations. (iv) Usage of inert materials (Teflon) for tubing and custom-made valves preventing sample contamination with non-patient-derived VOCs. (v) The possibility of indoor-air collection (blank samples for breath analysis) by an inbuilt small pump (M2K3, SCHEGO). This device has been designed and built in the laboratory of Breath Research Institute of the Austrian Academy of Sciences.
Table 2. Histological analyses of tumors from study participants. Only non-smokers and ex-smokers (with at least one year of non-smoking history) were recruited.
Lung cancer
Lung cancer + COPD
Cancer type
Male Female
Male Female
Total number
Small-cell LC Squamous-cell LC Adenocarcinoma Large-cell carcinoma Mesothelioma Carcinoid Sum
1 1 3 0 1 0 6
3 10 6 0 0 0 19
7 12 14 1 1 1 36
1 0 4 1 0 1 7
2 1 1 0 0 0 4
volunteers were non- or ex-smokers with at least one year of non-smoking history. The cancer patient cohort consisted of 25 males aged from 44 to 79 years and 11 females aged 42 to 74 years, all of them with LC confirmed by imaging techniques (CXR, CT or PET) and histology (sputum, bronchoscopy or biopsy). The healthy group included 12 males (22–87 years) and 16 females (22–78 years) without any lung diseases or injury. Although healthy controls were significantly younger (mean age healthy 52 ± 17 years versus mean age cancer 63 ± 7 years; p-value of Kruskal–Wallis test p = 0.016), no significant correlation between VOC concentrations and age could be found with Spearman’s correlation coefficient (ρ < 0.5, p > 0.05) for any substance significant for this study. The exclusion criteria for both groups were the previous surgical excision of a tumor (segmentectomy, lobectomy) and active smoking. To prevent fluctuations in breath gas composition caused by physical activity [48–51], all individuals rested for ∼10 min before breath collection and food consumption was not less than 2 h prior to sampling. All breath and indoor-air samples were collected in the morning and were stored in R bag for a maximum of 5 h, after which they the Tedlar were adsorbed on multibed sorption tubes and analyzed with GC-MS. Two breath sampling techniques were compared: (i) use of a plastic straw for collection of mixed-expiratory R R , Kynar or air (into sampling bags, such as Tedlar R Flexfilm bags, the usability of which has been discussed in other studies [52–54]) and (ii) use of a custom-made apparatus for collection of end-tidal air. For this purpose, a portable device with an inbuilt infrared sensor (IRMATM ICU, PHASEIN Medical Technologies, SE-182 33 Danderyd, Sweden) for real-time CO2 monitoring was developed. The most important improvements compared to currently available breath sampling techniques are: (i) heated (45 ◦ C) transfer lines and valves, preventing condensation of exhaled water and depletion of hydrophilic substances. (ii) Fully automated collection of end-tidal air only when exhaled CO2-level surpasses the threshold (in the range of 0–10%) preset by the operator. This was achieved by synchronization of a fast responding ( 0. Data gathered from 28 healthy non-smokers. Compounds released ex vivo from lung tissues are labeled with (T), while those released in vitro from lung cell lines are labeled with (CL).
Substance
CAS
Sampling device
BSD versus straw p-value
n>0
Mean
Median
VOCs detected at significantly higher levels in the end-tidal air collected with breath sampling device (E)-1,3-Dimethylcyclopentane 1759-58-6 5.9E-07 93% 1.3E + 06 1.1E + 06 n-Heptane 142-82-5 1.8E-06 100% 4.7E + 06 3.3E + 06 112-40-3 0.0001 96% 2.5E + 06 2.1E + 06 n-Dodecane (T) 2-Methyl-1-butene 563-46-2 0.0002 64% 9.5E + 05 9.6E + 05 2,2-Dimethylbutane 75-83-2 0.0002 96% 1.4E + 06 1.4E + 06 2,3-Dimethylbutane 79-29-8 0.0004 82% 4.7E + 05 4.7E + 05 111-65-9 0.0004 100% 1.7E + 06 1.5E + 06 n-Octane (T) (CL) 2213-23-2 0.0004 75% 1.9E + 06 6.5E + 05 2,4-Dimethylheptane (T) (CL) γ -Butyrolactone 96-48-0 0.0005 64% 6.2E + 05 6.2E + 05 Methyl tert-butyl ether 1634-04-4 0.0005 64% 5.7E + 05 3.7E + 05 Isobutane 75-28-5 0.0005 100% 9.2E + 06 8.0E + 06 Pyrrole (T) 109-97-7 0.0007 93% 5.1E + 05 4.8E + 05 o-Cymene 527-84-4 0.0010 93% 4.0E + 06 3.0E + 06 Ethylene oxide 75-21-8 0.0015 79% 7.5E + 06 8.8E + 06 n-Butane 106-97-8 0.0016 100% 3.5E + 07 2.2E + 07 4-Heptanone 123-19-3 0.0018 50% 4.4E + 05 2.2E + 05 n-Decane 124-18-5 0.0020 100% 3.6E + 06 3.2E + 06 111-84-2 0.0020 100% 2.0E + 06 1.8E + 06 n-Nonane (T) 2-Methylbutane 78-78-4 0.0021 100% 1.4E + 07 5.7E + 06 Furan 110-00-9 0.0026 100% 6.0E + 05 5.9E + 05 (E)-1-(methylthio)-1-propene 42848-06-6 0.0040 79% 5.4E + 06 1.7E + 06 107-87-9 0.0051 100% 3.2E + 06 3.1E + 06 2-Pentanone(T) 110-54-3 0.0065 100% 1.7E + 06 1.5E + 06 n-Hexane(T) 1-Butene 106-98-9 0.0090 54% 6.8E + 05 8.7E + 05 2-Methylhexane 591-76-4 0.0095 96% 3.5E + 06 2.6E + 06 2-Methylpropene 115-11-7 0.0135 96% 2.0E + 06 1.9E + 06 50% 9.1E + 05 1.7E + 05 2,4-Dimethyl-1-heptene (T) (CL) 19549-87-2 0.0165 Ethyl tert-butyl ether 637-92-3 0.0181 93% 1.2E + 06 9.9E + 05 (Z)-3-Dodecene 7239-23-8 0.0239 39% 5.7E + 05 0 Dimethylsulfide 75-18-3 0.0248 100% 1.9E + 07 1.2E + 07 2-Methylbutanal (T) 96-17-3 0.0269 68% 7.7E + 05 6.3E + 05 75-05-8 0.0285 100% 5.4E + 06 3.9E + 06 Acetonitrile (T) Butanal 123-72-8 0.0292 29% 5.2E + 04 0 3-Methyloctane 2216-33-3 0.0342 43% 3.8E + 05 0 100-52-7 0.0399 100% 4.9E + 06 4.8E + 06 Benzaldehyde(T) γ -Terpinene 99-85-4 0.0440 25% 1.3E + 06 0 Carbonyl sulfide 463-58-1 0.0472 96% 1.0E + 06 1.1E + 06 Methanethiol 74-93-1 0.0481 100% 7.2E + 05 6.7E + 05
Straw n>0
Mean
Median
0% 14% 14% 0% 14% 21% 21% 7% 7% 7% 79% 50% 64% 21% 100% 0% 100% 36% 50% 100% 29% 71% 100% 14% 36% 64% 7% 43% 7% 57% 29% 64% 0% 7% 93% 0% 43% 64%
0 4.0E + 05 4.8E + 05 0 3.7E + 05 1.7E + 05 6.0E + 05 1.6E + 05 9.1E + 03 2.0E + 04 3.7E + 06 1.8E + 05 1.3E + 06 2.2E + 06 1.1E + 07 0 2.2E + 06 8.4E + 05 4.6E + 06 3.8E + 05 7.2E + 05 1.9E + 06 1.1E + 06 1.4E + 05 1.4E + 06 1.3E + 06 8.6E + 05 6.5E + 05 5.2E + 04 1.0E + 07 2.6E + 05 3.3E + 06 0 1.1E + 05 3.6E + 06 0 8.4E + 05 5.2E + 05
0 0 0 0 0 0 0 0 0 0 1.8E + 06 7.8E + 04 9.2E + 05 0 9.9E + 06 0 2.2E + 06 0 1.1E + 06 3.3E + 05 0 1.4E + 06 1.0E + 06 0 0 9.1E + 05 0 0 0 5.1E + 06 0 2.4E + 06 0 0 3.8E + 06 0 0 3.0E + 05
93% 71% 43% 43% 93% 29% 29% 29% 29% 29% 64% 43% 43% 21% 21% 79% 36% 79% 29%
6.3E + 06 4.7E + 05 4.5E + 07 1.4E + 07 9.0E + 06 4.2E + 05 2.0E + 06 2.8E + 05 2.7E + 06 4.6E + 05 1.0E + 06 1.5E + 06 2.7E + 05 8.5E + 04 2.5E + 05 7.5E + 05 2.1E + 06 9.3E + 05 4.2E + 05
5.4E + 06 4.1E + 05 0 0 8.5E + 06 0 0 0 0 0 5.7E + 05 0 0 0 0 6.8E + 05 0 8.2E + 05 0
VOCs detected at significantly higher in mixed-expiratory air collected with plastic straw N,N-Dimethylformamide Pyridine (T) Methenamine Formaldehyde N,N-Diethylformamide 3-Methyl-2-cyclohexen-1-one 2 H-Perfluoropropene Dimethoxymethane 2-Ethyl-m-xylene Cyclohexanon Nonanal 1,3-Dioxolan Pentanal Carveol (2E)-2-Butenylbenzene Hexanal (T) 2-Acetyl-5-methylfuran Benzonitrile β,β-Dimethylstyrene
68-12-2 110-86-1 100-97-0 50-00-0 617-84-5 1193-18-6 690-27-7 109-87-5 2870-04-4 108-94-1 124-19-6 646-06-0 110-62-3 99-48-9 1560-06-1 66-25-1 1193-79-9 100-47-0 768-49-0
6.2E-05 0.0002 0.0002 0.0002 0.0016 0.0034 0.0034 0.0034 0.0034 0.0034 0.0059 0.0083 0.0114 0.0121 0.0121 0.0148 0.0191 0.0201 0.0201
36% 14% 0% 0% 79% 0% 0% 0% 0% 0% 18% 11% 7% 0% 0% 32% 7% 61% 4%
5
1.3E + 06 7.7E + 04 0 0 3.5E + 06 0 0 0 0 0 2.8E + 05 4.0E + 04 1.2E + 05 0 0 4.1E + 05 3.4E + 05 4.4E + 05 7.3E + 04
0 0 0 0 1.7E + 06 0 0 0 0 0 0 0 0 0 0 0 0 3.4E + 05 0
W Filipiak et al
J. Breath Res. 8 (2014) 027111
Table 3. (Continued.)
Substance
CAS
BSD versus straw p-value
3-Methyl-2(5 H)-furanone Dimethyl ether N,N-Dimethylacetamide Methyl formate Decanal (T) Styrene
22122-36-7 115-10-6 127-19-5 107-31-3 112-31-2 100-42-5
0.0211 0.0251 0.0268 0.0282 0.0373 0.0399
Sampling device
Straw
n>0
Mean
Median
n>0
Mean
Median
50% 50% 25% 21% 11% 100%
8.2E + 05 5.4E + 06 1.4E + 06 4.1E + 04 1.9E + 05 2.3E + 06
2.7E + 05 2.3E + 05 0 0 0 2.1E + 06
86% 93% 64% 57% 36% 100%
2.2E + 06 1.2E + 07 2.8E + 06 1.2E + 05 1.2E + 06 3.2E + 06
1.4E + 06 1.7E + 06 1.6E + 06 8.5E + 04 0 2.7E+06
sensor. The end-tidal air samples were collected with the here presented breath sampling device (BSD). The comparison of these two sampling techniques was performed on a cohort of 28 healthy non-smokers for which 176 VOCs could be detected in at least 10% of breath samples when collected with BSD compared to 152 VOCs when collected with a straw. Thirty-eight amongst these compounds, mostly hydrocarbons, were found at significantly (p < 0.05) higher levels in the samples collected with BSD as compared to samples collected with a straw. Contrary to mixed-expiratory air, collection of end-tidal air results also in higher occurrence of breath VOCs, as observed for all apart from four compounds (butane, decane, furan and hexane) (table 3). In turn, 25 compounds were found at significantly (p < 0.05) higher levels in the mixed-expiratory air collected with a plastic straw. However, many of them are known environmental pollutants (chlorinated and fluorinated hydrocarbons, aromatics, etc) or indoor-air contaminants (amides, diverse aldehydes) [34, 59–61], which accumulate in the anatomic dead space gas. Importantly, the concentration of ten metabolites that were found to be emitted from lung tissues studied here (listed in table 4) were significantly higher in end-tidal air samples collected with BSD as compared to only three VOCs released from lung tissues which were found at higher levels for mixed-expiratory air collected with a straw. These substances are additionally marked in table 3. Taken together, the results suggest that the BSD is superior to expired-air collection with a straw.
study cohorts, whereby the only ‘positive alveolar gradient’ for aldehydes was observed for 2-butenal in cancer patients and 2-methylbutanal in healthy individuals. Obviously, this complicates a comparison of such breath-profiles with those of in vitro and ex vivo studies. Altogether 23 VOCs were detected at significantly (p < 0.05) different levels in the breath gas of LC patients and healthy controls, mainly hydrocarbons. The most important are VOCs with positive alveolar gradients such as n-octane, nnonane and 2,3-butanedione, which were found at significantly higher levels in the cancer group either of breath samples or of lung tissue samples and even cell cultures (A549 for n-octane; see table 4 for more details). 3.3. VOCs detected in the headspace of lung tissues
Altogether 39 compounds were detected in the headspace of all lung specimens, out of which 30 were observed at elevated levels in tumor lung tissues (table 4) compared to healthy control tissues. For six of them a significant difference (tumor versus healthy tissues) was indicated by the Kruskal–Wallis test (p 0.05). Twenty-five substances were detected in at least 50% of LC tissues. Importantly, since human blood contains a myriad of organic molecules which are not necessarily of metabolic origin [37, 62–64] all lung specimens were rinsed in liquid buffer to eliminate the blood-borne substances, thus minimizing the background level for tissues’ headspace. This enables a more reliable comparison of VOCs excreted from tumor and healthy tissues. Amongst substances observed at significantly (p < 0.05) different concentrations in tumor compared to healthy tissues, ethanol is an important compound as it was detected more often and at higher levels in tumor tissues (average 3.5 ppbv) as compared to healthy tissues (average 0.08 ppbv) (table 4). Higher levels of ethanol have been observed for A549 cancer cell lines (229 ppbv compared to 63 ppbv for medium control) [41]. Other compounds present at significantly (p < 0.05) elevated levels in tumor tissues were pyridine, 4methylheptane, acetaldehyde, n-octane, and a similar tendency (but exceeding the significance level of 0.05) was observed for propanal, 2-propenal, 2,4-dimethylheptane, n-nonane, 2-methyl-1-pentene and 4-methyloctane 2,3-butanedione. Notably, apart from 4-methylheptane and 2-methyl-1-pentene (with ∼43% occurrence) all aforementioned VOCs were found in more than half of the LC specimens. Additionally, three compounds were detected at higher levels in healthy control tissues, namely 2-hexanone (p = 0.01, never found in tumor
3.2. Analysis of alveolar air from LC patients and healthy volunteers
Although patients rested for ∼10 min prior to breath sampling, the concentrations of exhaled VOCs (predominantly for aldehydes) were not at or near equilibrium with respective levels in inspired indoor-air (table 4). Therefore, we corrected expired VOCs for the respective inspired levels as proposed previously by Phillips et al [25], whereby the negative values of subtraction (expired–inspired) were not set to zero, revealing the so-called alveolar gradient of measured VOCs and visualizing the discrepancy between expired and inspired air. This was the case for 24 out of 39 VOCs reported here (table 4), for which a mean breath value indicating ‘negative alveolar gradient’ was found in at least one group of breath donors (20 VOCs in the lung cancer cohort, respectively 21 VOCs in the healthy controls). Particularly high levels in indoor-air (compared to breath) resulting in ‘negative alveolar gradient’ were observed for all aldehydes in both 6
lung tissue specimens. The results of breath analysis concern the mixed group of (non + ex)-smokers (36 LC patients and 28 healthy controls), whereby the VOCs levels in exhaled air were corrected by subtraction of inspired air level. Negative values for breath data were not normalized to zero to visualize difference between expired and inspired air levels, called ‘alveolar gradient’. VOCs significantly (p < 0.05) taken up by cancer and healthy lung cell under in vitro conditions are labeled with a downward arrow; VOCs significantly (p < 0.05) released from cell lines are labeled with an upward arrow. The difference between concentration [ppbv] in medium control and respective cell culture is given in parenthesis for each compound elucidated from the in vitro study. Concentrations [ppbv] of VOCs secreted from lung tissues (lobectomy specimens)
Concentrations [ppbv] of VOCs detected in breath gas of (non + ex)-smokers (expired - inspired)
Significant VOCs release from cancer and non-transformed cell lines under
HE versus CA
Healthy, n = 13
Cancer, n = 14
HE versus CA
Healthy, n = 28
Cancer, n = 36
Substance
CAS
p (KW)
n > 0 [%] Mean
n > 0 [%] Mean
p (KW)
n > 0 [%] Mean
n > 0 [%] Mean
Ethanol Pyridine 2-hexanone 4-Methylheptane
64-17-5 110-86-1 591-78-6 589-53-7
0.005 0.006 0.012 0.028
7.7 100 38.5 7.7
0.078 0.96 0.10 0
Compounds quantified to the part per billion units [ppbv] 57.1 3.51 2.60E-05 10.7 − 619.49 100 2.83 0.8037 7.1 − 0.03 0 0 0.4154 0 0 42.9 0.02 0.9410 10.7 − 0.02
51.4 25.7 8.6 14.3
− 4.65 − 0.02 0 − 0.02
Acetaldehyde
75-07-0
0.029
100
15.21
92.9
20.28
0.0031
28.6
− 15.21
65.7
− 2.92
Octane
111-65-9
0.046
84.6
0.07
85.7
0.14
0.0172
75
0.02
71.4
0.05
Propanal 2-Propenal (acrolein) 2,4-Dimethylheptane Nonane 2-Methyl-1-pentene 2-Butanon 4-Methyloctane 2-Propanol Butanedione
123-38-6 107-02-8 2213-23-2 111-84-2 763-29-1 78-93-3 2216-34-4 67-63-0 431-03-8
0.070 0.081 0.132 0.158 0.165 0.174 0.174 0.190 0.200
69.2 100 84.6 76.9 23.1 100 100 92.3 38.5
0.10 0.83 0.09 0.06 0.03 2.00 0.07 46.51 0.26
78.6 100 85.7 85.7 42.9 100 100 100 50.0
− 7.13 − 0.81 0.07 − 0.01 0.01 0.67 0.12 − 393.56 4.78
34.3 37.1 25.7 51.4 37.1 65.7 25.7 17.1 100
− 0.14 − 0.13 − 0.12 0.03 0.03 0.46 − 0.14 − 46.51 21.51
Benzaldehyde Acetic acid Propene 6-Methyl-5-heptene-2-one N-Dodecan Decanal Acetonitrile 2,3,4-Trimethylpentane
100-52-7 64-19-7 115-07-1 110-93-0 112-40-3 112-31-2 75-05-8 565-75-3
0.308 0.335 0.374 0.386 0.400 0.402 0.466 0.478
100 0 61.5 30.8 23.1 30.8 84.6 53.8
2.40E + 06 0 5.90E + 05 1.40E + 06 2.50E + 05 3.90E + 05 2.20E + 06 3.20E + 05
92.9 7.1 50.0 14.3 35.7 35.7 92.9 28.6
0.19 1.10E-08 7.1 1.11 0.0002 14.3 0.11 0.2554 28.6 0.11 0.0088 28.6 0.07 0.2658 25 1.51 0.8791 82.1 0.09 0.166 25 35.42 5.70E-08 3.6 0.79 7.20E-05 85.7 Compounds quantified to the peak area 2.90E + 06 0.5067 10.7 5.40E + 05 0.0041 28.6 2.50E + 05 0.0002 7.1 9.50E + 05 0.037 7.1 5.00E + 05 0.0011 10.7 1.20E + 06 0.0121 0 3.60E + 06 0.9559 82.1 4.10E + 05 0.0986 10.7
in vitro conditions
7
20 65.7 28.6 11.4 25.7 8.6 80 2.9
− 1.80E + 06 1.70E + 07 − 5.30E + 05 − 8.30E + 06 − 2.40E + 05 − 5.70E + 06 5.00E + 06 4.10E + 04
Lung Cells
Ref.
↑A549 (63.2 → 229.4) – ↑ hFB (0.6 → 2.3) ↑ HBEpC (0 → 0.6) ↑ hFB (0.8 → 2.9) ↓ HBEpC (741.8 → 365.7) ↓ hFB (883.1 → 167.7) ↓ CALU-1 (302.8 → 66.0) ↓ NCI-H2087 (203 → 0) ↑ hFB (8.0 → 12.3) ↑ A549 (1.6 → 2.9) – ↓ CALU-1 (4.4 → 2.5) ↑ CALU-1 (0.3 → 0.9) – ↑ A549 (1.7 → 7.0) ↓ CALU-1 (3.9 → 1.9) ↑ CALU-1 (1.6 → 2.8) – –
[41]
↓ hFB (36.3 → 2.8) – – – – ↓ CALU-1 (8.2 → 4.0) ↑ hFB (0.1 → 1.1)
[41]
[41], [42], [44]
[41]
[42] [42] [41] [42] [42]
[41]
[42] [41]
W Filipiak et al
− 1.90E + 06 4.30E + 06 − 2.70E + 06 − 4.20E + 06 − 7.90E + 05 − 1.80E + 06 1.90E + 06 4.90E + 04
J. Breath Res. 8 (2014) 027111
Table 4. Quantitative analysis of volatile compounds released from healthy (n = 13) or tumorous lung tissues (n = 14). VOCs are arranged according to the increasing p-values of significance for
Concentrations [ppbv] of VOCs secreted from lung tissues (lobectomy specimens)
Concentrations [ppbv] of VOCs detected in breath gas of (non + ex)-smokers (expired - inspired)
Significant VOCs release from cancer and non-transformed cell lines under
HE versus CA
Healthy, n = 13
Cancer, n = 14
HE versus CA
Healthy, n = 28
Cancer, n = 36
n > 0 [%] Mean
p (KW)
n > 0 [%] Mean
n > 0 [%] Mean
8
Substance
CAS
p (KW)
n > 0 [%] Mean
Methyl acetate 1(R)-a-Pinene 3-Methylbutanal
79-20-9 7785-70-8 590-86-3
0.534 0.534 0.610
7.7 7.7 92.3
1.10E + 04 14.3 4.40E + 04 14.3 6.60E + 05 92.9
1.10E + 06 0.0123 2.00E + 05 0.0012 7.60E + 05 0.4284
89.3 50 21.4
4.30E + 06 91.4 1.40E + 05 85.7 − 3.40E + 05 34.3
Pyrrole Hexanal
109-97-7 66-25-1
0.680 0.697
92.3 76.9
4.60E + 05 92.9 5.00E + 05 85.7
5.00E + 05 0.3524 5.30E + 05 0.0005
53.6 0
− 6.00E + 04 54.3 − 1.90E + 06 11.4
2-Butenal
123-73-9
0.737
38.5
7.80E + 04 35.7
1.30E + 05 0.2377
10.7
− 2.70E + 04 42.9
2,4-Dimethyl-1-heptene 19549-87-2 0.764
30.8
1.60E + 05 28.6
2.10E + 05 0.0052
17.9
− 6.10E + 05 11.4
Hexane Octanal Acetone
110-54-3 124-13-0 67-64-1
0.771 0.805 0.884
84.6 69.2 100
3.60E + 05 85.7 5.20E + 05 64.3 1.10E + 09 100
4.30E + 05 6.50E-06 6.10E + 05 0.0074 1.00E + 09 0.6985
57.1 0 100
2.00E + 05 91.4 − 1.30E + 06 5.7 9.50E + 08 100
2-Pentanone
107-87-9
0.884
100
2.50E + 06 100
2.60E + 06 0.3469
100
2.70E + 06 82.9
2-Methylpentane 2-Methylpropanal
107-83-5 78-84-2
0.903 0.915
92.3 7.7
2.50E + 06 92.9 1.90E + 04 7.1
2.50E + 06 0.0024 1.40E + 04 0.0032
42.9 7.1
− 6.70E + 05 65.7 − 2.70E + 05 28.6
3-Methylhexane Methanol 2-Methylbutanal
589-34-4 67-56-1 96-17-3
0.923 1.000 1.000
100 100 7.7
8.30E + 05 100 8.10E + 06 100 1.70E + 04 7.1
8.20E + 05 0.023 7.60E + 06 0.0128 2.20E + 04 0.416
32.1 85.7 42.9
5.60E + 05 57.1 4.40E + 07 80 6.70E + 04 28.6
J. Breath Res. 8 (2014) 027111
Table 4. (Continued.)
in vitro conditions Lung Cells
Ref.
1.60E + 06 ↑ HBEpC (0.9 → 5.2) 5.90E + 06 – − 2.10E + 05 ↓ HBEpC (54.1 → 0.6) ↓ hFB (158.9 → 3.8) ↓ CALU-1 (147.5 → 2.0) ↓A549 (191.8 → 2.6) ↓ NCI-H2087 (8.8 → 1.6) ↓ NCI-H1666 (12.3 → 1.3) 1.50E + 05 ↓ A549 (1.0 → 0) − 9.40E + 05 ↓ HBEpC (1093 → 192) ↓ CALU-1 (2.8 → 0.8) ↓ NCI-H1666 (3.1 → 0) 7.00E + 04 ↓ HBEpC (3.4 → 0) ↓ hFB (4.3 → 0) − 9.90E + 05 ↑ HBEpC (1.1 → 4.6) ↑ hFB (6.1 → 16.2) ↑ A549 (3.1 → 9.9) 1.40E + 07 ↑ NCI-H2087 (0.2 → 0.7) − 3.70E + 06 ↓ HBEpC (6.4 → 1.2) 1.30E + 09 ↑ HBEpC (13.9 → 39.5) ↑ A549 (193.4 → 357.8) 2.40E + 06 ↑ HBEpC (0 → 1.3) ↑ hFB (1.0 → 9.4) ↑ A549 (0.6 → 2.3) 1.20E + 07 ↑ NCI-H2087 (1.8 → 6.6) − 9.30E + 04 ↓ HBEpC (18.7 → 0.2) ↓ hFB (98.4 → 0.6) ↓ CALU-1 (21.3 → 1.4) ↓ A549 (59.8 → 2.5) ↓ NCI-H2087 (8.0 → 5.2) 6.00E + 05 – 1.60E + 07 – − 1.10E + 05 ↓ hFB (214.8 → 1.6) ↓ NCI-H2087 (14.1 → 8.2)
[41] [41, 42], [44]
[41] [41], [42], [43]
[41] [41]
[44] [41] [41] [41, 44]
[41, 42], [44]
[41, 44] W Filipiak et al
W Filipiak et al
J. Breath Res. 8 (2014) 027111
specimens), 2-butanone (p = 0.17) and 2-propanol (p = 0.19). For all remaining VOCs no difference was found between their levels in healthy versus cancer tissue (table 4).
Significantly higher VOC levels in end-tidal air (compared to mixed-expiratory air) can be explained by the fact that the blood-borne substances (e.g. carbon dioxide commonly used for capnography) come from the respiratory zone. Consequently, their concentration rises as the gas from anatomic dead space is increasingly washed out by alveolar gas until a steady state in their concentration is observed, representing alveolar air. On the other hand, the significantly higher level of 25 compounds being well known environmental pollutants and indoor air contaminants [57] were observed in mixed-expiratory air collected with a plastic straw (table 3) demonstrating breath gas contamination with indoor-air accumulated in the anatomic dead space. To clarify whether VOCs observed in expired air may be indicators of malignant transformation, we compared the results of our previously published in vitro studies [41–44] and ex vivo study (lung tissues) with the results of breath analysis. The majority of lung tissue specimens were collected from the patients with NSCLC, either adenocarcinoma (5 out of 14) or squamous-cell carcinoma (5 out of 14) (table 1). The investigated lung cancer cell lines also represented NSCLC, namely adenocarcinoma (NCI-H1666, NCI-H2087) and squamous-cell carcinoma (CALU-1, A549) [41]. Finally, 72% of the studied LC patients (breath donors) suffered from adenocarcinoma (n = 14) or squamous-cell carcinoma (n = 12) (table 2). All cell lines and tissue specimens investigated released a variety of organic compounds, mostly branched hydrocarbons (but also alcohols, esters and ethers), whereby n-octane and ethanol are of particular interest. Strong production of ethanol was observed for A549 LC cells but not for normal lung cells [41]. Ethanol was detected at significantly (p < 0.005) higher level in tumor than healthy lung tissues but also at significantly (p
Search
Collections
Journals
About
Contact us
My IOPscience
Comparative analyses of volatile organic compounds (VOCs) from patients, tumors and transformed cell lines for the validation of lung cancer-derived breath markers
This content has been downloaded from IOPscience. Please scroll down to see the full text. 2014 J. Breath Res. 8 027111 (http://iopscience.iop.org/1752-7163/8/2/027111) View the table of contents for this issue, or go to the journal homepage for more
Download details: IP Address: 132.203.227.61 This content was downloaded on 11/06/2014 at 10:27
Please note that terms and conditions apply.
Journal of Breath Research J. Breath Res. 8 (2014) 027111 (13pp)
doi:10.1088/1752-7155/8/2/027111
Comparative analyses of volatile organic compounds (VOCs) from patients, tumors and transformed cell lines for the validation of lung cancer-derived breath markers Wojciech Filipiak 1,2 , Anna Filipiak 1,2 , Andreas Sponring 1,2 , Thomas Schmid 3 , Bettina Zelger 4 , Clemens Ager 1,2 , Ewa Klodzinska 5 , Hubert Denz 1,6 , Alex Pizzini 6 , Paolo Lucciarini 3 , Herbert Jamnig 6 , Jakob Troppmair 7,8 and Anton Amann 1,2,8 1
Breath Research Institute of the University of Innsbruck, A-6850 Dornbirn, Austria Univ.-Clinic for Anesthesia and Intensive Care, Innsbruck Medical University, A-6020 Innsbruck, Austria 3 Department of Visceral, Transplant and Thoracic Surgery, University Hospital Innsbruck, A-6020 Innsbruck, Austria 4 Institute of Pathology, Innsbruck Medical University, A-6020 Innsbruck, Austria 5 Institute for Engineering of Polymer Materials and Dyes, Torun, Poland 6 Department of Pneumology, University Hospital Innsbruck, A-6161 Natters, Austria 7 Daniel Swarovski Research Laboratory, Department of Visceral, Transplant and Thoracic Surgery, Innsbruck Medical University, A-6020 Innsbruck, Austria 2
E-mail: [email protected] and [email protected] Received 12 December 2013, revised 14 March 2014 Accepted for publication 25 March 2014 Published 27 May 2014 Abstract
Breath analysis for the purpose of non-invasive diagnosis of lung cancer has yielded numerous candidate compounds with still questionable clinical relevance. To arrive at suitable volatile organic compounds our approach combined the analysis of different sources: isolated tumor samples compared to healthy lung tissues, and exhaled breath from lung cancer patients and healthy controls. Candidate compounds were further compared to substances previously identified in the comparison of transformed and normal lung epithelial cell lines. For human studies, a breath sampling device was developed enabling automated and CO2-controlled collection of the end-tidal air. All samples were first preconcentrated on multibed sorption tubes and analyzed with gas chromatography mass spectrometry (GC-MS). Significantly (p < 0.05) higher concentrations in all three types of cancer samples studied were observed for ethanol and n-octane. Additional metabolites (inter alia 2-methylpentane, n-hexane) significantly released by lung cancer cells were observed at higher levels in cancer lung tissues and breath samples (compared to respective healthy controls) with statistical significance (p < 0.05) only in breath samples. The results obtained confirmed the cancer-related origin of volatile metabolites, e.g. ethanol and octane that were both detected at significantly (p < 0.05) elevated concentrations in all three kinds of cancer samples studied. This work is an important step towards identification of volatile breath markers of lung cancer through the demonstration of cancer-related origin of certain volatile metabolites.
8
Authors to whom any correspondence should be addressed.
1752-7155/14/027111+13$33.00
1
© 2014 IOP Publishing Ltd
Printed in the UK
W Filipiak et al
J. Breath Res. 8 (2014) 027111
Keywords: lung cancer, breath analysis, tumor tissue, cancer cell lines, breath markers, adsorptive preconcentration, GC-MS (Some figures may appear in colour only in the online journal) to obtain a survey of candidate compounds, which may detect a transformed phenotype, i.e. LC. For this purpose we analyzed headspace samples of lung tumors and healthy lung tissues as well as breath samples from LC patients and healthy controls for VOC production. The data generated were further compared to the results of our in vitro experiments with LC cell lines [41–44].
1. Introduction The six most common cancers in humans are those of lung, breast, colon, rectum, stomach and prostate, with the highest mortality seen in lung cancer (LC) [1]. Although benign tumors can be cured by surgical resection [2], tumor size, type, location and associated co-morbidities predispose only 20– 30% of patients suffering from lung cancer to such treatment [3, 4]. An extensive study confirmed that most patients with stage III or IV lung cancer at the time of diagnosis time did not undergo surgery but instead were more likely to receive radiation treatment or even no treatment [5]. Worldwide, the most common method for lung cancer screening is chest x-ray (CXR) alone or in combination with cytology (sputum, bronchoscopy or biopsy). More recently, computed tomography (CT) [6–9] and positron emission tomography (PET) [10, 11] are also used. Though both imaging techniques have impressive sensitivity, which translates into a reduction in lung cancer deaths in the screened population [12–14], they also lead to the detection of many benign lung nodules and false positive results [15]. Furthermore, CT and PET techniques are expensive and require experienced operating personnel, and additional concerns about radiation exposure associated with frequent CT imaging persist [16]. Therefore, the development of an early diagnostic method which, due to its simplicity, non-invasiveness and low cost, eventually could be used by a regular physician is of utmost importance. Exhaled breath analysis has great potential for noninvasive diagnosis [17–19] as exemplified by the urea test for detection of Helicobacter pylori infection [20, 21] or the nitric oxide test for asthma monitoring [22, 23]. Particular effort has been made to develop breath analysis for the diagnosis of LC [24–32]. Numerous substances have been proposed as potential breath markers of LC [33, 34], but the clinical relevance for many of them is questionable [35]. Apart from strictly technical reasons, such as differences in breath sampling procedures or analytical techniques, an important issue is the limited knowledge about the origin of these compounds with regard to cellular sources (cancer cell, immune cells, infectious agents, etc) and biochemical pathways. Due to metabolic activity of organs like the liver, the kidneys or the lungs themselves, substances once released from an organ or a tumor site into the bloodstream may be further converted, resulting in an altered profile, when measured in exhaled air or other body emanations. It has been demonstrated recently that comparing volatile organic compounds (VOCs) from various sources yields complementary information with regard to the human volatolome [36, 37]. In our research approach we combine the comparative breath analysis of LC patients and normal subjects [37], ex vivo experiments with human lung tissues, in vitro experiments with normal and transformed cells as well as microorganisms frequently causing lung infections [38–40]. Here we describe our efforts
2. Materials and methods 2.1. Patient cohort and sample collection in ex vivo study with lung tissues
The lung tissues were collected from 14 patients: 8 women (amongst which 4 were active smokers) and 6 men (all nonsmokers) with different stages of LC (table 1) [45]. In all cases, tissue specimens were taken during segmentectomy and the sampling procedure is depicted in figure 1. From each patient undergoing surgery, tumor and healthy tissues were collected. Unfixed gross specimens were carefully examined by palpation and serial laminations. Grayish white, indurate tissue, macroscopically representing malignant neoplasia was taken in duplicates for our studies, and at least half of the tumor tissue was formalin fixed and paraffin embedded for diagnostic hematoxylin/eosin (H&E) staining and immunohistochemical evaluation according to an established classification scheme [45]. The same procedure was done with unaffected lung tissue for control samples whereby the amount of both tissue specimens (tumor and healthy) were collected in preferably the same amounts, mostly 2–3 g (dependent on tumor size). Immediately after surgical excision, specimens were rinsed in Ca2+ and Mg2+ free Dulbecco PBS buffer (Sigma-Aldrich, Steinheim, Germany) to remove blood, and carefully dried with a sterile swab prior to enclosing them in a gas-tight 20 ml glass vial. To minimize tissue necrosis, all glass vials containing specimens were kept in an ice-water bath at 0 ◦ C during transport while the adsorption of tissue headspace was done at 37 ◦ C and the entire procedure was done within approximately 30 min after the surgical resection of the specimen. Informed consent was obtained from all participants and the study was approved by the Ethics Committee of the Innsbruck Medical University. 2.2. Adsorptive preconcentration of headspace of lung tissues
For the dynamic headspace sampling a continuous flow (5 ml min−1) of synthetic air of purity 5.5 (Linde, Stadl-Paura, Austria) additionally purified on SupelcarbTM hydrocarbon trap (Bellefonte, PA, USA) was purging through the vial containing the specimen. Usage of air instead of nitrogen reduces the tissue necrosis, hence additionally minimizes the release of VOCs from dying cells. Apparently 100 ml of tissues’ headspace was adsorbed on the multibed sorption tube. To prevent excessive water uptake on carbon molecular sieves, 2
W Filipiak et al
J. Breath Res. 8 (2014) 027111
(a)
(b)
(c)
(d)
(f)
(e)
Figure 1. Procedure of lung tissues preparation and headspace sampling. All lung specimens were collected during segmentectomy (a). From the large tumor mass resected (b), healthy and tumor tissue were carefully separated by a pathologist (c), tissues were rinsed in buffer to remove blood (d), subsequently transferred to a tightly closed glass vial (e) and transported to the laboratory on ice. A custom-made system for VOC preconcentration ( f ) used a continuous flow (5 ml min−1) of purified synthetic air through the glass vial containing lung tissue with simultaneous sample dilution (1:6) with purified nitrogen to decrease the sample’s relative humidity to prevent excessive adsorption of water on multibed tube. VOC from the tissues’ headspace adsorbed on multibed sorption tubes were ultimately analyzed by GC-MS. Table 1. Lung tumor staging [45] and smoking habits of lung tissue donors.
Nr
Gender
Age
Smoking status
Cancer group
Cancer type
Stage of lung cancer
N
M
Tumor staging
1 2 3 4 5 6 7 8 9 10 11 12 13 14
Female Male Female Female Female Male Female Female Male Male Male Female Female Male
68 65 47 59 69 52 63 61 78 72 80 78 62 59
Non Non Active Active Non Unknown Active Active Non Non Non Non Ex Non
NSCLC NSCLC NSCLC NSCLC NSCLC NSCLC NSCLC NSCLC NSCLC NSCLC NSCLC NSCLC SCLC Metastasis to the lungs
Adenocarcinoma Adenocarcinoma Adenocarcinoma Adenocarcinoma Adenocarcinoma Large-cell carcinoma Large-cell carcinoma Squamous-cell carcinoma Squamous-cell carcinoma Squamous-cell carcinoma Squamous-cell carcinoma Squamous-cell carcinoma Small-cell lung cancer Sarcoma
1 2 2 2 3 2 3 2 2 2 3 1 1 4
0 0 0 0 1 0 2 0 0 0 2 0 0 1
0 0 0 0 0 0 0 0 0 0 0 0 0 1
pT1 pT1 pT1 pT1 pT3 pT2 pT4 pT1 pT1 pT2 pT2 pT1 pT1 n.a.
at an elevated temperature of 40 ◦ C to avoid condensation of water and depletion of polar analytes.
the sample relative humidity was reduced by dilution with dry nitrogen, as proposed by others [46, 47]. For this purpose, the headspace sample was mixed in a ratio of 1:6 with a nitrogen of purity 6.0 (i.e. 999 999%, Air Liquide Germany GmbH, Krefeld, Germany) additionally purified on a Carboxen 1000 trap. All transfer lines within the sampling system were kept
2.3. Patient cohort and sample collection in breath analysis
A total of 36 LC patients and 28 healthy individuals were recruited for this study (table 2). All LC patients and healthy 3
W Filipiak et al
J. Breath Res. 8 (2014) 027111
Pump model M2K3, SCHEGO Schemel & Goetz GmbH & Co KG, Offenbach, Germany) creating an underpressure drawing breath gas into the sampling bag. Importantly, the pump is placed outside the path of sample flow, preventing contamination or loss. (iii) An alternative sampling mode offers breath collection when the expired CO2 exceeds 80% of its level in a previous exhalation. This is required for subjects with breathing difficulties (older persons, patients suffering from advanced COPD) with gradually decreasing CO2 levels in subsequent exhalations. (iv) Usage of inert materials (Teflon) for tubing and custom-made valves preventing sample contamination with non-patient-derived VOCs. (v) The possibility of indoor-air collection (blank samples for breath analysis) by an inbuilt small pump (M2K3, SCHEGO). This device has been designed and built in the laboratory of Breath Research Institute of the Austrian Academy of Sciences.
Table 2. Histological analyses of tumors from study participants. Only non-smokers and ex-smokers (with at least one year of non-smoking history) were recruited.
Lung cancer
Lung cancer + COPD
Cancer type
Male Female
Male Female
Total number
Small-cell LC Squamous-cell LC Adenocarcinoma Large-cell carcinoma Mesothelioma Carcinoid Sum
1 1 3 0 1 0 6
3 10 6 0 0 0 19
7 12 14 1 1 1 36
1 0 4 1 0 1 7
2 1 1 0 0 0 4
volunteers were non- or ex-smokers with at least one year of non-smoking history. The cancer patient cohort consisted of 25 males aged from 44 to 79 years and 11 females aged 42 to 74 years, all of them with LC confirmed by imaging techniques (CXR, CT or PET) and histology (sputum, bronchoscopy or biopsy). The healthy group included 12 males (22–87 years) and 16 females (22–78 years) without any lung diseases or injury. Although healthy controls were significantly younger (mean age healthy 52 ± 17 years versus mean age cancer 63 ± 7 years; p-value of Kruskal–Wallis test p = 0.016), no significant correlation between VOC concentrations and age could be found with Spearman’s correlation coefficient (ρ < 0.5, p > 0.05) for any substance significant for this study. The exclusion criteria for both groups were the previous surgical excision of a tumor (segmentectomy, lobectomy) and active smoking. To prevent fluctuations in breath gas composition caused by physical activity [48–51], all individuals rested for ∼10 min before breath collection and food consumption was not less than 2 h prior to sampling. All breath and indoor-air samples were collected in the morning and were stored in R bag for a maximum of 5 h, after which they the Tedlar were adsorbed on multibed sorption tubes and analyzed with GC-MS. Two breath sampling techniques were compared: (i) use of a plastic straw for collection of mixed-expiratory R R , Kynar or air (into sampling bags, such as Tedlar R Flexfilm bags, the usability of which has been discussed in other studies [52–54]) and (ii) use of a custom-made apparatus for collection of end-tidal air. For this purpose, a portable device with an inbuilt infrared sensor (IRMATM ICU, PHASEIN Medical Technologies, SE-182 33 Danderyd, Sweden) for real-time CO2 monitoring was developed. The most important improvements compared to currently available breath sampling techniques are: (i) heated (45 ◦ C) transfer lines and valves, preventing condensation of exhaled water and depletion of hydrophilic substances. (ii) Fully automated collection of end-tidal air only when exhaled CO2-level surpasses the threshold (in the range of 0–10%) preset by the operator. This was achieved by synchronization of a fast responding ( 0. Data gathered from 28 healthy non-smokers. Compounds released ex vivo from lung tissues are labeled with (T), while those released in vitro from lung cell lines are labeled with (CL).
Substance
CAS
Sampling device
BSD versus straw p-value
n>0
Mean
Median
VOCs detected at significantly higher levels in the end-tidal air collected with breath sampling device (E)-1,3-Dimethylcyclopentane 1759-58-6 5.9E-07 93% 1.3E + 06 1.1E + 06 n-Heptane 142-82-5 1.8E-06 100% 4.7E + 06 3.3E + 06 112-40-3 0.0001 96% 2.5E + 06 2.1E + 06 n-Dodecane (T) 2-Methyl-1-butene 563-46-2 0.0002 64% 9.5E + 05 9.6E + 05 2,2-Dimethylbutane 75-83-2 0.0002 96% 1.4E + 06 1.4E + 06 2,3-Dimethylbutane 79-29-8 0.0004 82% 4.7E + 05 4.7E + 05 111-65-9 0.0004 100% 1.7E + 06 1.5E + 06 n-Octane (T) (CL) 2213-23-2 0.0004 75% 1.9E + 06 6.5E + 05 2,4-Dimethylheptane (T) (CL) γ -Butyrolactone 96-48-0 0.0005 64% 6.2E + 05 6.2E + 05 Methyl tert-butyl ether 1634-04-4 0.0005 64% 5.7E + 05 3.7E + 05 Isobutane 75-28-5 0.0005 100% 9.2E + 06 8.0E + 06 Pyrrole (T) 109-97-7 0.0007 93% 5.1E + 05 4.8E + 05 o-Cymene 527-84-4 0.0010 93% 4.0E + 06 3.0E + 06 Ethylene oxide 75-21-8 0.0015 79% 7.5E + 06 8.8E + 06 n-Butane 106-97-8 0.0016 100% 3.5E + 07 2.2E + 07 4-Heptanone 123-19-3 0.0018 50% 4.4E + 05 2.2E + 05 n-Decane 124-18-5 0.0020 100% 3.6E + 06 3.2E + 06 111-84-2 0.0020 100% 2.0E + 06 1.8E + 06 n-Nonane (T) 2-Methylbutane 78-78-4 0.0021 100% 1.4E + 07 5.7E + 06 Furan 110-00-9 0.0026 100% 6.0E + 05 5.9E + 05 (E)-1-(methylthio)-1-propene 42848-06-6 0.0040 79% 5.4E + 06 1.7E + 06 107-87-9 0.0051 100% 3.2E + 06 3.1E + 06 2-Pentanone(T) 110-54-3 0.0065 100% 1.7E + 06 1.5E + 06 n-Hexane(T) 1-Butene 106-98-9 0.0090 54% 6.8E + 05 8.7E + 05 2-Methylhexane 591-76-4 0.0095 96% 3.5E + 06 2.6E + 06 2-Methylpropene 115-11-7 0.0135 96% 2.0E + 06 1.9E + 06 50% 9.1E + 05 1.7E + 05 2,4-Dimethyl-1-heptene (T) (CL) 19549-87-2 0.0165 Ethyl tert-butyl ether 637-92-3 0.0181 93% 1.2E + 06 9.9E + 05 (Z)-3-Dodecene 7239-23-8 0.0239 39% 5.7E + 05 0 Dimethylsulfide 75-18-3 0.0248 100% 1.9E + 07 1.2E + 07 2-Methylbutanal (T) 96-17-3 0.0269 68% 7.7E + 05 6.3E + 05 75-05-8 0.0285 100% 5.4E + 06 3.9E + 06 Acetonitrile (T) Butanal 123-72-8 0.0292 29% 5.2E + 04 0 3-Methyloctane 2216-33-3 0.0342 43% 3.8E + 05 0 100-52-7 0.0399 100% 4.9E + 06 4.8E + 06 Benzaldehyde(T) γ -Terpinene 99-85-4 0.0440 25% 1.3E + 06 0 Carbonyl sulfide 463-58-1 0.0472 96% 1.0E + 06 1.1E + 06 Methanethiol 74-93-1 0.0481 100% 7.2E + 05 6.7E + 05
Straw n>0
Mean
Median
0% 14% 14% 0% 14% 21% 21% 7% 7% 7% 79% 50% 64% 21% 100% 0% 100% 36% 50% 100% 29% 71% 100% 14% 36% 64% 7% 43% 7% 57% 29% 64% 0% 7% 93% 0% 43% 64%
0 4.0E + 05 4.8E + 05 0 3.7E + 05 1.7E + 05 6.0E + 05 1.6E + 05 9.1E + 03 2.0E + 04 3.7E + 06 1.8E + 05 1.3E + 06 2.2E + 06 1.1E + 07 0 2.2E + 06 8.4E + 05 4.6E + 06 3.8E + 05 7.2E + 05 1.9E + 06 1.1E + 06 1.4E + 05 1.4E + 06 1.3E + 06 8.6E + 05 6.5E + 05 5.2E + 04 1.0E + 07 2.6E + 05 3.3E + 06 0 1.1E + 05 3.6E + 06 0 8.4E + 05 5.2E + 05
0 0 0 0 0 0 0 0 0 0 1.8E + 06 7.8E + 04 9.2E + 05 0 9.9E + 06 0 2.2E + 06 0 1.1E + 06 3.3E + 05 0 1.4E + 06 1.0E + 06 0 0 9.1E + 05 0 0 0 5.1E + 06 0 2.4E + 06 0 0 3.8E + 06 0 0 3.0E + 05
93% 71% 43% 43% 93% 29% 29% 29% 29% 29% 64% 43% 43% 21% 21% 79% 36% 79% 29%
6.3E + 06 4.7E + 05 4.5E + 07 1.4E + 07 9.0E + 06 4.2E + 05 2.0E + 06 2.8E + 05 2.7E + 06 4.6E + 05 1.0E + 06 1.5E + 06 2.7E + 05 8.5E + 04 2.5E + 05 7.5E + 05 2.1E + 06 9.3E + 05 4.2E + 05
5.4E + 06 4.1E + 05 0 0 8.5E + 06 0 0 0 0 0 5.7E + 05 0 0 0 0 6.8E + 05 0 8.2E + 05 0
VOCs detected at significantly higher in mixed-expiratory air collected with plastic straw N,N-Dimethylformamide Pyridine (T) Methenamine Formaldehyde N,N-Diethylformamide 3-Methyl-2-cyclohexen-1-one 2 H-Perfluoropropene Dimethoxymethane 2-Ethyl-m-xylene Cyclohexanon Nonanal 1,3-Dioxolan Pentanal Carveol (2E)-2-Butenylbenzene Hexanal (T) 2-Acetyl-5-methylfuran Benzonitrile β,β-Dimethylstyrene
68-12-2 110-86-1 100-97-0 50-00-0 617-84-5 1193-18-6 690-27-7 109-87-5 2870-04-4 108-94-1 124-19-6 646-06-0 110-62-3 99-48-9 1560-06-1 66-25-1 1193-79-9 100-47-0 768-49-0
6.2E-05 0.0002 0.0002 0.0002 0.0016 0.0034 0.0034 0.0034 0.0034 0.0034 0.0059 0.0083 0.0114 0.0121 0.0121 0.0148 0.0191 0.0201 0.0201
36% 14% 0% 0% 79% 0% 0% 0% 0% 0% 18% 11% 7% 0% 0% 32% 7% 61% 4%
5
1.3E + 06 7.7E + 04 0 0 3.5E + 06 0 0 0 0 0 2.8E + 05 4.0E + 04 1.2E + 05 0 0 4.1E + 05 3.4E + 05 4.4E + 05 7.3E + 04
0 0 0 0 1.7E + 06 0 0 0 0 0 0 0 0 0 0 0 0 3.4E + 05 0
W Filipiak et al
J. Breath Res. 8 (2014) 027111
Table 3. (Continued.)
Substance
CAS
BSD versus straw p-value
3-Methyl-2(5 H)-furanone Dimethyl ether N,N-Dimethylacetamide Methyl formate Decanal (T) Styrene
22122-36-7 115-10-6 127-19-5 107-31-3 112-31-2 100-42-5
0.0211 0.0251 0.0268 0.0282 0.0373 0.0399
Sampling device
Straw
n>0
Mean
Median
n>0
Mean
Median
50% 50% 25% 21% 11% 100%
8.2E + 05 5.4E + 06 1.4E + 06 4.1E + 04 1.9E + 05 2.3E + 06
2.7E + 05 2.3E + 05 0 0 0 2.1E + 06
86% 93% 64% 57% 36% 100%
2.2E + 06 1.2E + 07 2.8E + 06 1.2E + 05 1.2E + 06 3.2E + 06
1.4E + 06 1.7E + 06 1.6E + 06 8.5E + 04 0 2.7E+06
sensor. The end-tidal air samples were collected with the here presented breath sampling device (BSD). The comparison of these two sampling techniques was performed on a cohort of 28 healthy non-smokers for which 176 VOCs could be detected in at least 10% of breath samples when collected with BSD compared to 152 VOCs when collected with a straw. Thirty-eight amongst these compounds, mostly hydrocarbons, were found at significantly (p < 0.05) higher levels in the samples collected with BSD as compared to samples collected with a straw. Contrary to mixed-expiratory air, collection of end-tidal air results also in higher occurrence of breath VOCs, as observed for all apart from four compounds (butane, decane, furan and hexane) (table 3). In turn, 25 compounds were found at significantly (p < 0.05) higher levels in the mixed-expiratory air collected with a plastic straw. However, many of them are known environmental pollutants (chlorinated and fluorinated hydrocarbons, aromatics, etc) or indoor-air contaminants (amides, diverse aldehydes) [34, 59–61], which accumulate in the anatomic dead space gas. Importantly, the concentration of ten metabolites that were found to be emitted from lung tissues studied here (listed in table 4) were significantly higher in end-tidal air samples collected with BSD as compared to only three VOCs released from lung tissues which were found at higher levels for mixed-expiratory air collected with a straw. These substances are additionally marked in table 3. Taken together, the results suggest that the BSD is superior to expired-air collection with a straw.
study cohorts, whereby the only ‘positive alveolar gradient’ for aldehydes was observed for 2-butenal in cancer patients and 2-methylbutanal in healthy individuals. Obviously, this complicates a comparison of such breath-profiles with those of in vitro and ex vivo studies. Altogether 23 VOCs were detected at significantly (p < 0.05) different levels in the breath gas of LC patients and healthy controls, mainly hydrocarbons. The most important are VOCs with positive alveolar gradients such as n-octane, nnonane and 2,3-butanedione, which were found at significantly higher levels in the cancer group either of breath samples or of lung tissue samples and even cell cultures (A549 for n-octane; see table 4 for more details). 3.3. VOCs detected in the headspace of lung tissues
Altogether 39 compounds were detected in the headspace of all lung specimens, out of which 30 were observed at elevated levels in tumor lung tissues (table 4) compared to healthy control tissues. For six of them a significant difference (tumor versus healthy tissues) was indicated by the Kruskal–Wallis test (p 0.05). Twenty-five substances were detected in at least 50% of LC tissues. Importantly, since human blood contains a myriad of organic molecules which are not necessarily of metabolic origin [37, 62–64] all lung specimens were rinsed in liquid buffer to eliminate the blood-borne substances, thus minimizing the background level for tissues’ headspace. This enables a more reliable comparison of VOCs excreted from tumor and healthy tissues. Amongst substances observed at significantly (p < 0.05) different concentrations in tumor compared to healthy tissues, ethanol is an important compound as it was detected more often and at higher levels in tumor tissues (average 3.5 ppbv) as compared to healthy tissues (average 0.08 ppbv) (table 4). Higher levels of ethanol have been observed for A549 cancer cell lines (229 ppbv compared to 63 ppbv for medium control) [41]. Other compounds present at significantly (p < 0.05) elevated levels in tumor tissues were pyridine, 4methylheptane, acetaldehyde, n-octane, and a similar tendency (but exceeding the significance level of 0.05) was observed for propanal, 2-propenal, 2,4-dimethylheptane, n-nonane, 2-methyl-1-pentene and 4-methyloctane 2,3-butanedione. Notably, apart from 4-methylheptane and 2-methyl-1-pentene (with ∼43% occurrence) all aforementioned VOCs were found in more than half of the LC specimens. Additionally, three compounds were detected at higher levels in healthy control tissues, namely 2-hexanone (p = 0.01, never found in tumor
3.2. Analysis of alveolar air from LC patients and healthy volunteers
Although patients rested for ∼10 min prior to breath sampling, the concentrations of exhaled VOCs (predominantly for aldehydes) were not at or near equilibrium with respective levels in inspired indoor-air (table 4). Therefore, we corrected expired VOCs for the respective inspired levels as proposed previously by Phillips et al [25], whereby the negative values of subtraction (expired–inspired) were not set to zero, revealing the so-called alveolar gradient of measured VOCs and visualizing the discrepancy between expired and inspired air. This was the case for 24 out of 39 VOCs reported here (table 4), for which a mean breath value indicating ‘negative alveolar gradient’ was found in at least one group of breath donors (20 VOCs in the lung cancer cohort, respectively 21 VOCs in the healthy controls). Particularly high levels in indoor-air (compared to breath) resulting in ‘negative alveolar gradient’ were observed for all aldehydes in both 6
lung tissue specimens. The results of breath analysis concern the mixed group of (non + ex)-smokers (36 LC patients and 28 healthy controls), whereby the VOCs levels in exhaled air were corrected by subtraction of inspired air level. Negative values for breath data were not normalized to zero to visualize difference between expired and inspired air levels, called ‘alveolar gradient’. VOCs significantly (p < 0.05) taken up by cancer and healthy lung cell under in vitro conditions are labeled with a downward arrow; VOCs significantly (p < 0.05) released from cell lines are labeled with an upward arrow. The difference between concentration [ppbv] in medium control and respective cell culture is given in parenthesis for each compound elucidated from the in vitro study. Concentrations [ppbv] of VOCs secreted from lung tissues (lobectomy specimens)
Concentrations [ppbv] of VOCs detected in breath gas of (non + ex)-smokers (expired - inspired)
Significant VOCs release from cancer and non-transformed cell lines under
HE versus CA
Healthy, n = 13
Cancer, n = 14
HE versus CA
Healthy, n = 28
Cancer, n = 36
Substance
CAS
p (KW)
n > 0 [%] Mean
n > 0 [%] Mean
p (KW)
n > 0 [%] Mean
n > 0 [%] Mean
Ethanol Pyridine 2-hexanone 4-Methylheptane
64-17-5 110-86-1 591-78-6 589-53-7
0.005 0.006 0.012 0.028
7.7 100 38.5 7.7
0.078 0.96 0.10 0
Compounds quantified to the part per billion units [ppbv] 57.1 3.51 2.60E-05 10.7 − 619.49 100 2.83 0.8037 7.1 − 0.03 0 0 0.4154 0 0 42.9 0.02 0.9410 10.7 − 0.02
51.4 25.7 8.6 14.3
− 4.65 − 0.02 0 − 0.02
Acetaldehyde
75-07-0
0.029
100
15.21
92.9
20.28
0.0031
28.6
− 15.21
65.7
− 2.92
Octane
111-65-9
0.046
84.6
0.07
85.7
0.14
0.0172
75
0.02
71.4
0.05
Propanal 2-Propenal (acrolein) 2,4-Dimethylheptane Nonane 2-Methyl-1-pentene 2-Butanon 4-Methyloctane 2-Propanol Butanedione
123-38-6 107-02-8 2213-23-2 111-84-2 763-29-1 78-93-3 2216-34-4 67-63-0 431-03-8
0.070 0.081 0.132 0.158 0.165 0.174 0.174 0.190 0.200
69.2 100 84.6 76.9 23.1 100 100 92.3 38.5
0.10 0.83 0.09 0.06 0.03 2.00 0.07 46.51 0.26
78.6 100 85.7 85.7 42.9 100 100 100 50.0
− 7.13 − 0.81 0.07 − 0.01 0.01 0.67 0.12 − 393.56 4.78
34.3 37.1 25.7 51.4 37.1 65.7 25.7 17.1 100
− 0.14 − 0.13 − 0.12 0.03 0.03 0.46 − 0.14 − 46.51 21.51
Benzaldehyde Acetic acid Propene 6-Methyl-5-heptene-2-one N-Dodecan Decanal Acetonitrile 2,3,4-Trimethylpentane
100-52-7 64-19-7 115-07-1 110-93-0 112-40-3 112-31-2 75-05-8 565-75-3
0.308 0.335 0.374 0.386 0.400 0.402 0.466 0.478
100 0 61.5 30.8 23.1 30.8 84.6 53.8
2.40E + 06 0 5.90E + 05 1.40E + 06 2.50E + 05 3.90E + 05 2.20E + 06 3.20E + 05
92.9 7.1 50.0 14.3 35.7 35.7 92.9 28.6
0.19 1.10E-08 7.1 1.11 0.0002 14.3 0.11 0.2554 28.6 0.11 0.0088 28.6 0.07 0.2658 25 1.51 0.8791 82.1 0.09 0.166 25 35.42 5.70E-08 3.6 0.79 7.20E-05 85.7 Compounds quantified to the peak area 2.90E + 06 0.5067 10.7 5.40E + 05 0.0041 28.6 2.50E + 05 0.0002 7.1 9.50E + 05 0.037 7.1 5.00E + 05 0.0011 10.7 1.20E + 06 0.0121 0 3.60E + 06 0.9559 82.1 4.10E + 05 0.0986 10.7
in vitro conditions
7
20 65.7 28.6 11.4 25.7 8.6 80 2.9
− 1.80E + 06 1.70E + 07 − 5.30E + 05 − 8.30E + 06 − 2.40E + 05 − 5.70E + 06 5.00E + 06 4.10E + 04
Lung Cells
Ref.
↑A549 (63.2 → 229.4) – ↑ hFB (0.6 → 2.3) ↑ HBEpC (0 → 0.6) ↑ hFB (0.8 → 2.9) ↓ HBEpC (741.8 → 365.7) ↓ hFB (883.1 → 167.7) ↓ CALU-1 (302.8 → 66.0) ↓ NCI-H2087 (203 → 0) ↑ hFB (8.0 → 12.3) ↑ A549 (1.6 → 2.9) – ↓ CALU-1 (4.4 → 2.5) ↑ CALU-1 (0.3 → 0.9) – ↑ A549 (1.7 → 7.0) ↓ CALU-1 (3.9 → 1.9) ↑ CALU-1 (1.6 → 2.8) – –
[41]
↓ hFB (36.3 → 2.8) – – – – ↓ CALU-1 (8.2 → 4.0) ↑ hFB (0.1 → 1.1)
[41]
[41], [42], [44]
[41]
[42] [42] [41] [42] [42]
[41]
[42] [41]
W Filipiak et al
− 1.90E + 06 4.30E + 06 − 2.70E + 06 − 4.20E + 06 − 7.90E + 05 − 1.80E + 06 1.90E + 06 4.90E + 04
J. Breath Res. 8 (2014) 027111
Table 4. Quantitative analysis of volatile compounds released from healthy (n = 13) or tumorous lung tissues (n = 14). VOCs are arranged according to the increasing p-values of significance for
Concentrations [ppbv] of VOCs secreted from lung tissues (lobectomy specimens)
Concentrations [ppbv] of VOCs detected in breath gas of (non + ex)-smokers (expired - inspired)
Significant VOCs release from cancer and non-transformed cell lines under
HE versus CA
Healthy, n = 13
Cancer, n = 14
HE versus CA
Healthy, n = 28
Cancer, n = 36
n > 0 [%] Mean
p (KW)
n > 0 [%] Mean
n > 0 [%] Mean
8
Substance
CAS
p (KW)
n > 0 [%] Mean
Methyl acetate 1(R)-a-Pinene 3-Methylbutanal
79-20-9 7785-70-8 590-86-3
0.534 0.534 0.610
7.7 7.7 92.3
1.10E + 04 14.3 4.40E + 04 14.3 6.60E + 05 92.9
1.10E + 06 0.0123 2.00E + 05 0.0012 7.60E + 05 0.4284
89.3 50 21.4
4.30E + 06 91.4 1.40E + 05 85.7 − 3.40E + 05 34.3
Pyrrole Hexanal
109-97-7 66-25-1
0.680 0.697
92.3 76.9
4.60E + 05 92.9 5.00E + 05 85.7
5.00E + 05 0.3524 5.30E + 05 0.0005
53.6 0
− 6.00E + 04 54.3 − 1.90E + 06 11.4
2-Butenal
123-73-9
0.737
38.5
7.80E + 04 35.7
1.30E + 05 0.2377
10.7
− 2.70E + 04 42.9
2,4-Dimethyl-1-heptene 19549-87-2 0.764
30.8
1.60E + 05 28.6
2.10E + 05 0.0052
17.9
− 6.10E + 05 11.4
Hexane Octanal Acetone
110-54-3 124-13-0 67-64-1
0.771 0.805 0.884
84.6 69.2 100
3.60E + 05 85.7 5.20E + 05 64.3 1.10E + 09 100
4.30E + 05 6.50E-06 6.10E + 05 0.0074 1.00E + 09 0.6985
57.1 0 100
2.00E + 05 91.4 − 1.30E + 06 5.7 9.50E + 08 100
2-Pentanone
107-87-9
0.884
100
2.50E + 06 100
2.60E + 06 0.3469
100
2.70E + 06 82.9
2-Methylpentane 2-Methylpropanal
107-83-5 78-84-2
0.903 0.915
92.3 7.7
2.50E + 06 92.9 1.90E + 04 7.1
2.50E + 06 0.0024 1.40E + 04 0.0032
42.9 7.1
− 6.70E + 05 65.7 − 2.70E + 05 28.6
3-Methylhexane Methanol 2-Methylbutanal
589-34-4 67-56-1 96-17-3
0.923 1.000 1.000
100 100 7.7
8.30E + 05 100 8.10E + 06 100 1.70E + 04 7.1
8.20E + 05 0.023 7.60E + 06 0.0128 2.20E + 04 0.416
32.1 85.7 42.9
5.60E + 05 57.1 4.40E + 07 80 6.70E + 04 28.6
J. Breath Res. 8 (2014) 027111
Table 4. (Continued.)
in vitro conditions Lung Cells
Ref.
1.60E + 06 ↑ HBEpC (0.9 → 5.2) 5.90E + 06 – − 2.10E + 05 ↓ HBEpC (54.1 → 0.6) ↓ hFB (158.9 → 3.8) ↓ CALU-1 (147.5 → 2.0) ↓A549 (191.8 → 2.6) ↓ NCI-H2087 (8.8 → 1.6) ↓ NCI-H1666 (12.3 → 1.3) 1.50E + 05 ↓ A549 (1.0 → 0) − 9.40E + 05 ↓ HBEpC (1093 → 192) ↓ CALU-1 (2.8 → 0.8) ↓ NCI-H1666 (3.1 → 0) 7.00E + 04 ↓ HBEpC (3.4 → 0) ↓ hFB (4.3 → 0) − 9.90E + 05 ↑ HBEpC (1.1 → 4.6) ↑ hFB (6.1 → 16.2) ↑ A549 (3.1 → 9.9) 1.40E + 07 ↑ NCI-H2087 (0.2 → 0.7) − 3.70E + 06 ↓ HBEpC (6.4 → 1.2) 1.30E + 09 ↑ HBEpC (13.9 → 39.5) ↑ A549 (193.4 → 357.8) 2.40E + 06 ↑ HBEpC (0 → 1.3) ↑ hFB (1.0 → 9.4) ↑ A549 (0.6 → 2.3) 1.20E + 07 ↑ NCI-H2087 (1.8 → 6.6) − 9.30E + 04 ↓ HBEpC (18.7 → 0.2) ↓ hFB (98.4 → 0.6) ↓ CALU-1 (21.3 → 1.4) ↓ A549 (59.8 → 2.5) ↓ NCI-H2087 (8.0 → 5.2) 6.00E + 05 – 1.60E + 07 – − 1.10E + 05 ↓ hFB (214.8 → 1.6) ↓ NCI-H2087 (14.1 → 8.2)
[41] [41, 42], [44]
[41] [41], [42], [43]
[41] [41]
[44] [41] [41] [41, 44]
[41, 42], [44]
[41, 44] W Filipiak et al
W Filipiak et al
J. Breath Res. 8 (2014) 027111
specimens), 2-butanone (p = 0.17) and 2-propanol (p = 0.19). For all remaining VOCs no difference was found between their levels in healthy versus cancer tissue (table 4).
Significantly higher VOC levels in end-tidal air (compared to mixed-expiratory air) can be explained by the fact that the blood-borne substances (e.g. carbon dioxide commonly used for capnography) come from the respiratory zone. Consequently, their concentration rises as the gas from anatomic dead space is increasingly washed out by alveolar gas until a steady state in their concentration is observed, representing alveolar air. On the other hand, the significantly higher level of 25 compounds being well known environmental pollutants and indoor air contaminants [57] were observed in mixed-expiratory air collected with a plastic straw (table 3) demonstrating breath gas contamination with indoor-air accumulated in the anatomic dead space. To clarify whether VOCs observed in expired air may be indicators of malignant transformation, we compared the results of our previously published in vitro studies [41–44] and ex vivo study (lung tissues) with the results of breath analysis. The majority of lung tissue specimens were collected from the patients with NSCLC, either adenocarcinoma (5 out of 14) or squamous-cell carcinoma (5 out of 14) (table 1). The investigated lung cancer cell lines also represented NSCLC, namely adenocarcinoma (NCI-H1666, NCI-H2087) and squamous-cell carcinoma (CALU-1, A549) [41]. Finally, 72% of the studied LC patients (breath donors) suffered from adenocarcinoma (n = 14) or squamous-cell carcinoma (n = 12) (table 2). All cell lines and tissue specimens investigated released a variety of organic compounds, mostly branched hydrocarbons (but also alcohols, esters and ethers), whereby n-octane and ethanol are of particular interest. Strong production of ethanol was observed for A549 LC cells but not for normal lung cells [41]. Ethanol was detected at significantly (p < 0.005) higher level in tumor than healthy lung tissues but also at significantly (p
