1H NMR To Evaluate the Metabolome of Bronchoalveolar Lavage

Subscriber access provided by UB + Fachbibliothek Chemie | (FU-Bibliothekssystem)

Article

1H-NMR TO EVALUATE THE METABOLOME OF BRONCHOALVEOLAR LAVAGE FLUID (BALf) IN BRONCHIOLITIS OBLITERANS SYNDROME (BOS): TOWARDS THE DEVELOPMENT OF A NEW APPROACH FOR BIOMARKER IDENTIFICATION Carlotta Ciaramelli, Marco Fumagalli, Simona Viglio, Anna Maria Bardoni, Davide Piloni, Federica Meloni, Paolo Iadarola, and Cristina Airoldi J. Proteome Res., Just Accepted Manuscript • DOI: 10.1021/acs.jproteome.6b01038 • Publication Date (Web): 28 Feb 2017 Downloaded from http://pubs.acs.org on March 2, 2017

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Journal of Proteome Research is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 45

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Proteome Research

1

H-NMR TO EVALUATE THE METABOLOME OF BRONCHOALVEOLAR LAVAGE FLUID (BALf)

IN BRONCHIOLITIS OBLITERANS SYNDROME (BOS): TOWARDS THE DEVELOPMENT OF A NEW APPROACH FOR BIOMARKER IDENTIFICATION Carlotta Ciaramelli,a Marco Fumagalli,b Simona Viglio,c Anna Maria Bardoni,c Davide Piloni,d Federica Meloni,d Paolo Iadarola,b* Cristina Airoldi a*

a

Department of Biotechnology and Biosciences, University of Milano-Bicocca, 20126, Milan, Italy. b

c

d

Department of Biology and Biotechnologies "L. Spallanzani" and

Department of Molecular Medicine, Biochemistry Unit, University of Pavia, Italy.

Department of Internal Medicine, Section of Pneumology, University of Pavia, & IRCCS

Foundation Policlinico San Matteo, Department of Cardiothoracic and Vascular Department, Pneumology Unit, Pavia, Italy.

*Corresponding authors:

1 ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Paolo Iadarola Department of Biology and Biotechnologies "L. Spallanzani", Biochemistry Unit University of Pavia, Via Taramelli 3, 27100 Pavia, Italy. Tel. +39 0382 98 7264 Fax. +39 0382 423108 [email protected]

Cristina Airoldi Department of Biotechnology and Biosciences, University of Milano-Bicocca, P.zza della Scienza 2, 20126, Milan, Italy. Tel. +39 02 6448 3303 Fax. +39 02 6448 3565 [email protected]


Page 2 of 45

Page 3 of 45

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


ABSTRACT This report describes the application of NMR spectroscopy to the profiling of metabolites in BronchoAlveolar Lavage fluid (BALf) of lung transplant recipients without Bronchiolitis Obliterans Syndrome (BOS) (stable, S, n=10), and with BOS at different degree of severity (BOS 0p, n=10; BOS I, n=10). Through the fine tuning of a number of parameters concerning both sample preparation/processing and variations of spectra acquisition modes, an efficient and reproducible protocol was designed for the screening of metabolites in a pulmonary fluid that should reflect the status of airway inflammation/injury. Exploiting the combination of mono and bi-dimensional NMR experiments, 38 polar metabolites, including amino acids, Krebs cycle intermediates, monoand di-saccharides, nucleotides and phospholipid precursors were unequivocally identified. To determine which signature could be correlated with the onset of BOS, metabolites’ content of the above recipients was analyzed by multivariate (PCA and OPLS-DA) statistical methods. PCA analysis (almost) totally differentiated S from BOS I and this discrimination was significantly improved by the application of OPLS-DA, whose model was characterized by excellent fit and prediction values (R2=0.99 and Q2=0.88). The analysis of S vs BOS 0p and of BOS 0p vs BOS I samples showed a clear discrimination of considered cohorts, although with a poorer efficiency compared to those measured for S vs BOS I patients. The data shown in this work assess the suitability of NMR approach in monitoring different pathological lung conditions.

KEYWORDS: BALf (BronchoAlveolar Lavage fluid,) BOS (Bronchiolitis Obliterans Syndrome), CLAD (Chronic Lung Allograft Dysfunction), NMR-based metabolomics.

INTRODUCTION Early and long-term graft and patient survival after lung transplantation continue to be challenged respectively by primary graft dysfunction and by Chronic Lung Allograft Dysfunction (CLAD),



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

whose main phenotype is represented by Bronchiolitis Obliterans Syndrome (BOS), a condition characterized by an irreversible obstructive graft dysfunction due to fibro-obliterative lesion of small airways. 1 In addition, a restrictive phenotype of CLAD, named Restrictive Allograft Syndrome (RAS) was recently identified as being characterized by a restrictive functional pattern and typical chest imaging (CT scan) findings (upper lobe dominant fibrosis, irregular opacities, and traction bronchiectasis). From a clinical standpoint, CLAD diagnosis relies on functional parameters, i.e. an irreversible decline in lung function compared to the best post-transplant value and on chest imaging. A significant heterogeneity is also depicted in the time of onset and evolution of CLAD, varying from an insidious onset and gradual decline in graft function to a more abrupt onset and more severe decline. Given this heterogeneity, attempts to early identify and unravel specific molecular pathways associated to different CLAD phenotypes, mainly BOS, have been made, by far with scanty results.2,3 Therefore, tools that will help unraveling the complexity of the disease and identifying possible useful predictive markers are urgently needed. On the assumption that any pathological condition leads to a number of physiological changes, in particular distinctive shifts in the metabolic profile, a better understanding of the metabolic variations that accompany a given disorder would obviously provide new insights into disease pathophysiology. Being changes in levels of metabolic intermediates of a sequential series of reactions often more pronounced than variations in enzymatic kinetics or individual fluxes, a technique that involves the holistic profiling of metabolites within a biological matrix would be the most appropriate analytical tool for the analysis of their fluctuations. As such, metabolomics is indeed a competent platform capable of capturing disease-relevant metabolic profile changes and molecular signatures of disease processes.4,5,6 Despite numerous metabolomic analyses have already been performed on human biofluids such as urine and blood,7 to date only relatively few studies have been specifically designed to address the application of this approach to respiratory samples in general, and to BronchoAlveolar Lavage fluid


Page 4 of 45

Page 5 of 45

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


(BALf) in particular. It can be hypothesized that one of the reasons that may limit the utility of this biofluid for metabolomic studies is the combination of its high protein/salt content and the low concentration of metabolites. Nevertheless, being BALf routinely collected as part of routine follow up of lung recipients, it was considered the most appropriate source of data of interest. The rationale for this choice was that BALf might contain components, which likely reflect the distress signals of the injured compartment, in other words biomarkers of the disease. Thus, in an effort to understand whether significantly up- or down-regulated metabolites could be identified in BALf of patients after BOS development with respect to stable recipients or to patients suffering from other pulmonary pathologies, a non-targeted approach was applied to analyze this biofluid. By investigating all metabolites detectable in a fluid or tissue, this non-specific approach allows to capture as much information as possible, thus providing a functional fingerprint of the physiological/pathological state of the fluid submitted to analysis. One of the leading analytical tools for metabolomic research is currently high-resolution Nuclear Magnetic Resonance spectroscopy (hr-NMR) whose inherently quantitative signals and nontargeted nature are indeed important advantages. NMR also shows a good reproducibility and it is not destructive with regard to the sample for which little or no preparation is required.7 Previously published reports dealing with the application of 1H-NMR to the analysis of BALf have been essentially focused on: i) the investigation of the effect of naphthalene/1-nitronaphthalene on the respiratory system of rats and mice exposed to these chemicals,8,9,10 ii) the analysis of miniBALf (reflecting mainly bronchial micro-environment) collected from adults affected by different lung injures,11,12,13 iii) the study of preterm infants with Respiratory Distress Syndrome (RDS), aimed at understanding whether the metabolic profile of these very low birth weight infants was different from that of healthy controls14 and iv) the study of pediatric patients with Cystic Fibrosis (CF), in an effort to correlate the degree of airway inflammation to the amount of metabolites in BALf.15 In some cases,13,15 NMR proved to be a powerful tool in demonstrating a good correlation between



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

both the number and the concentration of metabolites in BAlf of those patients and the level of lung inflammation. Based on these encouraging results, the utility of NMR platform for profiling the BALf metabolome in lung transplant recipients with or without BOS, was explored in this pilot study. Detailed insights into biologically and technically influencing factors, which may hamper analysis of samples and limit interpretation of spectral data, will also be presented.

MATERIALS AND METHODS Materials Unless otherwise stated, all materials used in this work were from Sigma-Aldrich (St. Louis, MO, USA). Reagent-grade water used to prepare all solutions was obtained from a Milli-Q (Millipore, Bedford, MA, USA) purification system. Patient selection Thirty five samples from lung transplant recipients, some of them being stable and others suffering from BOS, and ten samples obtained from patients with other respiratory diseases (four cases of active stage 2 sarcoidosis, four cases of Extrinsic Allergic Alveolitis –EAA-, one from bacterial pneumonia and one from inflammatory pulmonary consolidations) were provided from the department of Respiratory diseases and transplant follow up unit of IRCCS San Matteo Foundation (Pavia, Italy) after obtaining informed consent. Samples from lung recipients were obtained within the surveillance FU program at different post-transplant time points and retrospectively selected for this study according to the results of the clinical, microbiological and functional examinations performed at the same post-transplant control visit. Demographic and clinical features of patients are detailed in Table 1 (including age, gender, lung transplant indication, immunosuppressive regimen, CLAD occurance, length of follow-up, other treatment strategies such as azithromycin and ECP).


Page 6 of 45

Page 7 of 45

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


It should be pointed out that the key functional feature of BOS is the development of airway obstruction with a reduction of Forced Expiratory Volume in 1 second (FEV1) that does not respond to bronchodilators and has no other alternative clinical cause. Thus, the current classification of severity of BOS is based on changes in FEV1 and is indicated as BOS 0p (potential BOS) if FEV1 is 81-90% of best FEV1 value obtained after transplantation; BOS I when FEV1 is 66-80% of best value; BOS II when FEV1 is 51-65% of best and BOS III if FEV1 is ≤ 50%. The study protocol was approved by the local ethics committee. BALf collection Lung lavage fluid (BAL) was obtained from patients during fiberoptic bronchoscopy performed for routine surveillance for transplant rejection. Briefly, 150 mL of 0.9% sterile saline solution were instilled in 5 subsequent 30 mL aliquots, which were sequentially retrieved by gentle aspiration through the suction port of the bronchoscope. Return from the first aliquot was used for microbiological and virological studies. The returned fluid of the second to fifth aliquots was pooled, filtered through a 40-mm filter, submitted to centrifugation (at 1,500 rpm for 10 min) to remove the cellular components and further processed as the BALf. To avoid repeated freezing and thawing for subsequent studies, BALfs from all subjects were aliquoted (2 mL each) and stored frozen at -80 °C until analysis. To check for potentially deleterious effects of storage temperature on sample integrity, a few BALf aliquots (n= 5) were stored also at -20 °C. Whole BAL samples (which include both cellular and a-cellular components) from five patients (affected by different pulmonary disorders) were also stored, to be analyzed “in parallel” with the corresponding BALfs. Experimental conditions for NMR analysis A variety of experimental parameters was explored to evaluate the effect of different conditions on BALf integrity. These included submission or not of samples to a lyophilization step prior to their



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

analysis. Freeze-dried samples were reconstituted in 10 mM deuterated phosphate buffer, either at the same or at higher concentration (1.75x, 3.5x, 5x or 10x) of the original specimen. Nonlyophilized ones were analyzed simply after adding 10% D2O to sample solution. 4,4-dimethyl-4silapentane-1-sulfonic acid (DSS, final concentration 0.1 mM) was added to all samples as internal reference of both concentration and chemical shift. pH values were measured with a Microelectrode (Mettler-Toledo, Greifensee, Switzerland) for 5 mm NMR tubes and adjusted to 7.4 with small amounts (few µL) of NaOD or DCl. pH* reading in D2O were corrected for the deuterium isotope effect at the glass electrode.16 The acquisition temperature was 25 °C. All spectra were acquired on a Bruker AVANCE III 600 MHz NMR spectrometer equipped with a QCI (1H, 13C, 15N/31P and 2H lock) cryogenic probe. 1D 1H-NMR spectra were recorded with water suppression (cpmgpr1d or noesygppr1d pulse sequences in Bruker library) and 1024 scans, spectral width of 20 ppm, relaxation delay of 5 s. They were processed with a line broadening of 0.3 Hz and automatically phased and baseline corrected. Chemical shifts values were internally calibrated to the DSS peak at 0.0 ppm. Metabolite identification and assignment were performed with the support of 2D NMR experiments, the Human Metabolome Database, 17,18,19 the Biological Magnetic Resonance Data Bank 20 and the SMA analysis tool integrated in MestreNova software. 21 In particular, 1H,1HTOCSY (TOtal Correlation SpectroscopY) spectra (dipsi2esgpphpp pulse sequence in Bruker library) were acquired with 120 scans and 512 increments, a mixing time of 80 ms and the relaxation delay was 2 s. 1H,13C-HSQC (Heteronuclear Single Quantum Coherence) spectra (hsqcetgppr pulse sequence in Bruker library) were acquired with 180 scans and 256 increments, a relaxation delay of 2.5 s. NMR spectra processing and peak peaking were performed by MNova software package of Mestrelab (MestReNova v 10.0, 2016 Mestrelab Research S.L.). For metabolite quantification, the SMA analysis tool integrated in MestreNova software package of Mestrelab (MestReNova v 10.0,


Page 8 of 45

Page 9 of 45

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


2016 Mestrelab Research S.L.) was exploited21 and, thus, the GSD (global spectrum deconvolution) algorithm. In this way, overlapping regions were deconvoluted and absolute quantification performed also for metabolites with resonances in crowded spectral areas. When possible, for the different compounds, the concentration was calculated from the mean value of the different assigned signals.22 Before starting with the analysis of all samples, the reproducibility of runs was checked by assessing within-day and between-day repeatability, according to the protocol previously described and validated.23 In particular, for 4 BALf samples (2 belonging to BOS patients and 2 to stable subjects) five 1H-NMR spectra were recorded (1) to comply with conditions for repeatability (measurements performed under the same operating conditions, over a short period of time), considering consecutive runs without removing the NMR tube from the magnet; and (2) to comply with conditions for intermediate precision, considering at least 24 h delay between runs, removal of the NMR tube from the magnet from run to run, with sample storage at 4 °C. Summarizing, spectra have been acquired in three different sessions: (i) three consecutive runs per NMR tube (spectra 1, 2 and 3); (ii) one run per NMR tube delayed 48 h from the first session (spectrum 4); (iii) one run per NMR tube delayed 48 h from the second session (spectrum 5). Spectra were divided in four regions (4.5-2.95 ppm; 2.88-2.22 ppm; 2.16-1.35 ppm; 1.31-0.7 ppm), excluding signals from residual HDO, DSS and the anesthetic lidocaine; each region was integrated and normalized to the total spectrum area, obtaining four parameters on which the coefficient of variation (CV%) was calculated. For within-day repeatability, CV% values (calculated for spectra 1, 2 and 3) were in the range 0.3 – 3.5%; for between-day repeatability (calculated for spectra 1, 4 and 5), CV% values were in the range 0.7 – 4.1%. Therefore, all CV% values were considerably below the accepted CV% limit of 10%.24 Statistical analysis



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Statistical analysis was performed exploiting the real-time interactive web-based application Metaboanalyst 3.0.25,26,27 To discriminate different BALf samples through NMR spectra, multivariate statistical analyses were carried out. Data were uploaded as a table reporting spectra peak-peaking for each sample, where a sample corresponds to the BALf spectrum of a different individual, classified as Stable (S), BOS 0p, BOS I or affected by Other Pulmonary Pathologies (OPP). Assuming that most missing values are caused by low abundance metabolites (i.e. below the detection limit), to avoid problems for downstream analysis, small values (the half of the minimum value present in the original data) were arbitrarily accounted for missing values.25,26,27 Because of the large difference in the total amount of metabolites among the diverse samples (mainly due to differences in samples dilution, depending on the intrinsic variability of the sample collecting procedure), for all analyses, data were firstly normalized to the constant sum, then transformed with log normalization and scaled with Pareto scaling, increasing the contribution of lower concentration metabolites in the generated models compared with models where no scaling has been employed. Data were subjected to multivariate analyses in both unsupervised (through Principal Component Analysis, PCA), and supervised mode (by Orthogonal Partial Least Squares Discriminant Analysis, OPLS-DA). Leave-one-out-cross validation (LOOCV) has been applied to validate the model. OPLS-DA models have been validated in order to understand whether the separation was statistically significant or due to random noise. This hypothesis was tested using the permutation tests in each permutation, as implemented in Metaboanalyst: an OPLS-DA model was built between the data (X) and the permuted class labels (Y) using the optimal number of components determined by previous cross validation calculations and based on the original class assignment. Finally, to compare S, BOS 0p and BOS I and S to BOS I, a hierarchical cluster analysis was performed. Clustering was represented in the form of dendogram.


Page 10 of 45

Page 11 of 45

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


The Pathway Analysis was performed trough the respective module present in MetaboAnalyst. It combines results from powerful pathway enrichment analysis with the pathway topology analysis to help researchers to identify the most relevant pathways involved in the conditions under study using the high-quality KEGG metabolic pathways as the backend knowledgebase. Data were uploaded as compound concentration table with samples in columns and compounds (metabolites able to differentiate S and BOS I groups identified by previous multivariate analysis) in rows. Analysis were performed by selecting Homo sapiens (human) pathway libraries. The results from pathway analysis are presented graphically.

RESULTS Setting and optimization of NMR experiments To design an efficient and reproducible protocol for the screening of metabolites in BALf, preparation and processing of samples were optimized by exploring a range of experimental conditions. These included testing: i) different sample concentrations; ii) variations of sample temperature and time of storage and iii) diverse spectra acquisition modes. The low metabolite concentration observed in the “original” samples resulted, for some of them, in very poor NMR signals. This prevented detection and quantification of a good number of analytes. To overcome this drawback, samples were lyophilized and re-dissolved in a smaller volume of deuterated phosphate buffer although this meant facing a high concentration of salts (BALf is collected in 0.9% saline solution) which interfere with probe (particularly cryo-probe) performances, in terms of sensitivity.28 Among the different sample, concentration factors explored (1.75x, 3.5x, 5x, 10x), the 5x factor was found to be the most appropriate and, unless otherwise stated, all further analyses were performed on lyophilized samples which were re-dissolved in the appropriate volume of deuterated 10 mM phosphate buffer to obtain a final concentration factor of 5x. The profile of a BALf sample (representative of all other samples) obtained under these



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

conditions is shown in panel A of Figure 1. To account for possible sample degradation after storage at different temperatures and/or for different times, the metabolic profiles of the same sample stored in diverse conditions were compared. Spectra were divided in four regions (4.5-2.95 ppm; 2.88-2.22 ppm; 2.16-1.35 ppm; 1.31-0.7 ppm), excluding the signals from residual HDO, DSS and the aesthetic lidocaine; each region was integrated and the coefficient of variation (CV%) was calculated on these four parameters. The similarity/identity of spectra, that can be inferred from their visual inspection (Figure 1), and the calculated CV% suggest that storage for several months (panel B, CV% = 5) and at low/very low (-20/-80 °C) temperatures (panel C, CV% = 6) did not affect sample integrity. As far as 1H-NMR spectra acquisition is concerned, both 1D NOESY-presat sequence and CPMG relaxation-editing sequence with pre-saturation were tested. While both allowing the saturation of water (or HDO) signal, the peak intensities observed in the CPMG sequence, edited on the basis of NMR relaxation times, highlighted the signals from low molecular weight metabolites. This appears clear in Figure 1 (panel D) that shows how resonances attributable to macromolecules (basically proteins and lipids, evidenced by arrows in spectrum 1) are erased with CPMG sequence only, the spectrum also showing a flatter baseline. This feature is of great importance since it allows a more accurate multiplicity analysis and integration of metabolite signals that could be critical for identification/quantification of small molecules. Being CPMG relaxation-editing sequence more suitable for this kind of experiments, it was employed for the analysis of all BALf samples. Identification of metabolites in BALf samples Exploiting the combination of mono (Figures 2 and S1 in Supporting Information) and bidimensional NMR spectra (Figure S2), 38 polar metabolites, including amino acids, Krebs cycle intermediates, mono- and di-saccharides, nucleotides and phospholipid precursors were unequivocally identified in BALf samples, in substantial agreement with previously reported assignments.12,13 The resonances of 4 additional metabolites have not been assigned yet. A careful


Page 12 of 45

Page 13 of 45

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


observation of profiles shows the constant presence of five signals, two of which of high intensity (at 2.19 and 7.25 ppm), unambiguously identified as lidocaine (Figure 2 and Table 2). Being topical lidocaine usually administered through the flexible bronchoscope to reduce excessive coughing and patient discomfort during BAL collection, its finding was not such surprising. This signal was detectable also in BALf spectra produced by other groups15 although, to the best of our knowledge, it has never been identified before. The complete list of metabolites is reported in Table 2. Comparison of BOS to other respiratory pathologies To assess the suitability of this approach in monitoring different medical conditions, profiles of BALf from BOS, sarcoidosis, pneumonia, pulmonary consolidations and EAA patients were compared. A NMR spectrum representative of each group of patients is shown in Figure S3. While appearing similar, profiles of patients with different pathological conditions evidenced distinctive differences in terms of presence/absence of some specific metabolites and/or of their levels. Multivariate statistical analysis allowed to assess that, on the basis of NMR metabolomics profiling, subjects affected by BOS (in particular BOS I) could be distinguished from those affected by the other lung pathologies (OPP). As a matter of fact, while the unsupervised PCA analysis showed a partial overlapping between the 95% confidence regions of the two groups (Figure 3A), when the supervised OPLS-DA method was applied, a complete separation between these groups was observed (Figure 3B). The supervised model was tested by iteratively predicting the class membership of every sample. Results allowed to evaluate the model describing BOS I vs OPP. The calculated values of R2Y = 0.9 and Q2 = 0.63 suggested the ability of this model to explain data variation with an acceptable predictivity. In fact, as by literature directions, acceptable values for the goodness of ﬁt (R2) and prediction (Q2) must be ≥0.5. 29 Permutation test (n=100) was performed, obtaining the following empirical p-values: R2Y p = 0.01 (1/100) and Q2 p < 0.01 (0/100).



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Comparison of stable-transplanted subjects and BOS patients These encouraging results prompted further investigations evaluating the usability of the same statistical analyses on samples from transplanted subjects without BOS (stable, S, n=10) or with BOS at different levels of severity (BOS 0p, n=10; BOS I, n=10), in an effort to detect signatures that could be correlated with the onset of the pathological condition. When the experimental data sets were submitted to PCA analysis, the score plot obtained, while being able to discriminate totally S from BOS I, also showed a partial overlapping of BOS 0p to both S and BOS I samples (Figure 4A). A graphical representation of the extent to which each peak of the NMR spectrum accounts for the variance in the data is shown by the corresponding PCA loading plot of the first two eigenvalues (PC 1 and PC 2) (Figure S4). That BOS I patients could be easily discriminated from stable subjects became also evident looking at the dendogram generated by cluster analysis of the NMR profiles (Figure 5). Not surprisingly, the same dendogram provided a further confirmation (if needed) that BOS 0p samples, being spread out into the two clusters, could not be completely separated from the other two groups. Interestingly, the application of OPLS-DA analysis allowed to observe that S, BOS 0p and BOS I samples were completely separated (Figure 4B), the generated model having R2Y = 0.95 and Q2 = 0.61. When permutation test (n = 100) was done, we obtained empirical p-values R2Y: p < 0.01 (1/100) and Q2: p < 0.01 (0/100). The OPLS-DA loading plot is shown in Figure S5. The loading plots from PCA and OPLS-DA analysis allowed to identify the peaks responsible for S, BOS 0p and BOS I separation and to assign them to the corresponding metabolites. The concentrations of the most relevant metabolites for class discrimination were calculated integrating the corresponding signal/s, as described in Material and Methods section, and are provided in Supporting Information (Table S2). Interestingly, while these metabolites are present in the NMR spectra of all groups, they allow group separation because of their different concentration in each


Page 14 of 45

Page 15 of 45

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


class. Figure 6 depicts the variation of single metabolite concentration for S, BOS 0p and BOS I and the corresponding mean values are reported in Table 3. Comparison of cohorts two-by-two The question of whether groups under investigation could be differentiated two-by-two was answered by submitting to statistical analyses S vs BOS I; BOS 0p vs BOS I and S vs BOS 0p, respectively. The confirmation of data reliability came from the results of PCA analysis which showed that the two groups S and BOS I were almost totally differentiated (Figure 7A) and even better was the outcome of the OPLS-DA model that showed the complete differentiation of these two groups with excellent fit and prediction values (R2Y=0.99 and Q2=0.88) (Figure 7B). Permutation test (n = 100) was performed, yielding the following empirical p-values: R2Y p = 0.01 (0/100) and Q2 p = 0.01 (0/100). Taken together, these findings allowed to conclude that S could be unambiguously separated from BOS I subjects on the basis of their metabolites’ content, in accordance with the outcome of multivariate analysis performed on S, BOS 0p and BOS I samples already discussed above. This condition is graphically represented by the dendogram generated by cluster analysis shown in Figure S6. Analogous comparisons were carried out for BOS 0p vs S and for BOS 0p vs BOS I samples. The corresponding score plots obtained from PCA and OPLS-DA analyses are reported in Figure S7 and S8, respectively, of Supporting Information. In both cases, it can be observed that the OPLSDA model is characterized by R2Y and Q2 values (RY2 = 0.97 and Q2 = 0.78 for BOS 0p vs S and R2Y = 0.87 and Q2 = 0.578 for BOS 0p vs BOS I) which, although with a poorer efficiency compared to those measured for S vs BOS I patients, still allow a discrimination of the considered cohorts. Permutation test (n = 100) gave empirical p-values R2Y: p < 0.01 (0/100) and Q2: p < 0.01 (0/100) for BOS 0p vs S, and R2Y: p < 0.03 (3/100) and Q2: p < 0.01 (0/100) for BOS 0p vs BOS I. The very poor quality of NMR spectra and the very low number of BOS II (n=2) and BOS III (n=3) patients included in this study, prevented the application of statistical analysis also to these samples.



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Pathway analysis The MetaboAnalyst 3.0 platform, applied to our list of metabolites using “human” as the specific model organism22, allowed to generate a pathway analysis aimed at defining the relationships among them. The metabolome view shown in Figure 8 visualizes all matched pathways according to p-values from pathway enrichment analysis and pathway impact values from pathway typology analysis. This plot revealed that, among the pathways observed in this study, a critical role is played by the pyruvate and the taurine/hypotaurine pathways. Both have a good pathway impact value, the p value being p< 0.03 for the former and p< 0.005 for the latter. 1

H-NMR spectra from whole BAL samples

Given the availability of BAL samples for a limited number of patients suffering from the different lung disorders mentioned in this work, also this matrix was analyzed with the aim of understanding whether, in terms of analyte loss, BALf recovery from BAL could be a critical step. Thus, both BAL and the corresponding BALf from five patients, chosen at random among all available, have been submitted to the same characterization protocol. As shown in Figure S9, NMR profiles for both BAL and BALf of the same individual appear very similar, if not identical. This indicates that the cell removal (by centrifugation) and the filtration step involved in BALf preparation from BAL does not significantly affect the recovery of metabolites. In our opinion these data, if confirmed on larger cohorts, are of great interest since they provide a rationale for considering whole BAL as a good biological sample, substitute of BALf, for future NMR studies on pulmonary disorders.

DISCUSSION The metabolic profile of BALf from adults has rarely been studied by NMR and none of the articles published so far reported on the analysis of this fluid from BOS patients. In fact, while NMR has been previously applied to the investigation of BALf from mice,8,9,10 to the best of our knowledge this is the first report dealing with the application of this technique to BALf samples from adults.


Page 16 of 45

Page 17 of 45

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Indeed, other works on humans have been mainly focused either on the use of miniBALf samples,11,12,13 or on BALf from pediatric subjects,14,15 both collected with different procedures from that used for this study. Moreover, BALf or miniBALf samples used in works mentioned above have been analyzed after dilution in phosphate buffer or by simply adding 10% D2O. Thus, to design an efficient and reproducible protocol for the screening of metabolites in BALf from adults, the preparation and processing of samples have been optimized. This resulted in the “tailoring” of a method that allowed to assess the suitability of the NMR approach in monitoring different medical conditions and to explore the metabolic profile of BALfs from lung transplant recipients with or without BOS. This represents a post-transplant complication whose syndrome is usually diagnosed at a relatively advanced stage, where the higher likelihood of irreversible lung damage limits the efficacy of available therapies and, thus, survival. Hence, the possibility to predict BOS could allow for early intervention, potentially preventing irreversible lung damage and improving outcomes. Aim of this study was bridging the level of lung complication in patients under investigation and both the number and the concentration of metabolites in their BALf, in the awareness that succeeding in identifying molecules that nobody expected to be there (and the metabolic pathway they are involved in) would be an important contribution which may open the door to clinical studies. The metabolites detected by the NMR platform belonged to a variety of different chemical classes, which included amino acids, Krebs cycle intermediates, mono- and di-saccharides, nucleotides and phospholipid precursors. This fact, other than being a source of confusion, was an incentive to the search of a rationale for reasoning on the role of (some of) these molecules in the onset of BOS, being possible biomarkers of the disorder. Thus, efforts have been devoted to answer the question of whether metabolites identified were ideally suited for this purpose. The finding that all groups of individuals investigated shared largely the same metabolites’ panel allowed to focus on their quantitative variations. The results shown here seem to establish a previously unrecognized link



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

between the amount of metabolites released in BALf and the extent of disorder severity. Collectively, our experimental observations point to these metabolites as appropriate indicators of patients’ clinical conditions. In fact, if the indisputable differentiation between stable subjects and BOS I patients is indicative of a founder effect, the observed “distribution” of potential-BOS subjects between these two cohorts of individuals is well-consistent with these findings. As shown by several studies, patients negative for BOS 0p at a given time may change to BOS 0p-positive during subsequent follow-up.30 This suggests an intrinsic difficulty in the definite identification of a solid BOS 0p–negative (or positive) population and implies that these subjects may be found either in S and in BOS I groups. Thus, if the prediction of BOS goes through the implementation of increasingly sensitive criteria for detection of early decline in pulmonary function, an essential step in both treating patients and devising clinical trials with early intervention would be the identification of markers of disease. As such, the metabolites identified in this study, in their capacity of being “predictors” of the disease, seem to address these requirements. As shown by the pathway analysis built by including our metabolites in the MetaboAnalyst 3.0 platform, the pyruvate and the taurine/hypotaurine pathways may play a critical role in BOS development. The detection of a few metabolites which may be related to the former pathway was not a matter of surprise. For example, lactate was previously shown to be present in airway secretions15 and, based on experimental evidences, it was found that an increase in lactate production may be triggered by lung inflammation.31 This allowed us to speculate that the increased lactate levels in BOS patients at different stages of severity could be consistent with an increased inflammatory state. The finding that the levels of a good number of branched-chain amino acids (valine, leucine, isoleucine) was increasing with the severity of the disease somehow supported this hypothesis. In fact, it has been previously demonstrated that the amount of these amino acids secreted in BALf is correlated with different inflammatory states, such as sepsis.15,32 However, these hypotheses do not exclude that the increased lactate levels in our patients may reflect the hypoxic environment of the tissue, very low


Page 18 of 45

Page 19 of 45

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


oxygen concentrations being present in the deepest layers of airway inflamed biofilms.33 As airway epithelia reside at these deepest layers, they are probably forced to rely on anaerobic metabolism for energy, which could result in increased production of lactate. The presence of a ketone body (acetone) within BALF samples of our patients is still poorly understood. In fact, on one side its finding was not completely surprising since not only relatively high and stable concentrations of this compound within the human airways have been previously described, but it has also been shown that its airway levels can be affected by diabetic chetoacidosis.34 As a matter of fact, some of our patients had diabetes but all of them were well-controlled with no sign of chetoacidosis at the time of BALf sampling. It is also a matter of speculation whether acetone may serve as a precursor of pyruvate, which can then enter gluconeogenesis, as it has been shown to happen in rats. 35 Although the capacity of this pathway is limited, evidence was cumulated that acetone, through a variety of reactions, may be a sort of “back door” for fatty acyl carbon to be turned back into glucose. In addition, it has also been estimated that, in fasting humans, up to 11% of endogenously produced glucose may be derived from acetone.36 If this hypothesis proves true, then also other C3 metabolites detected in our samples may be assigned to the pyruvate pathway according to the scheme shown in Figure 9. Obviously, further studies will be needed to establish whether or not this suggestive hypothesis is corroborated by clinical conditions of patients investigated. A few comments about the potential role of taurine/hypotaurine: it should be underlined that, besides its well-known involvement in bile acid conjugation, taurine is also acknowledged for having a role in the regulation of cell volume, as a membrane stabilizing agent, and for its antioxidative, anti-inflammatory and anti-apoptotic effects.37,38 In addition, taurine was shown to be a weak agonist for chloride-permeable gamma-aminobutyric acid type A receptors (GABAAR) and glycine receptors (GlyR) that are located not only in the neural synapsis but also in the central nervous system and in the lungs. 39 The suggested role of taurine in the lung was to potentiate relaxation of airway smooth muscle cells through its binding to the α4-subunit of GABAA receptors.



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Thus, under normal conditions, a release of taurine via GABAA and glycine receptors may stimulate the relaxation of lung smooth muscle cells and the secretion of mucus from goblet cells. The regulation of cell volume occurs by means of taurine accumulation via the Na+-dependent TauT and the proton-coupled PAT-1 transporters and subsequent taurine release using both volumeinsensitive and volume-sensitive leak pathways (VSOAC).37,40 The Tau-T activity is strongly downregulated by acidification, osmotic cell swelling and exposure to reactive oxygen species (ROS). At the same time, the ROS species act by amplifying the effect of volume sensitive tyrosine kinases (PTK) through the inhibition of specific phosphatases and kinases, thus prolonging the open probability of VSOAC, which is also up-regulated by eicosanoids released during inflammation.40 In summary, the release of ROS species by neutrophils and macrophages during inflammatory states, together with the production of eicosanoids by leukocytes, could stimulate a decrease of the intracellular taurine intake and an increase of the taurine release in extracellular fluids with the consequent increase of mucus secretion. The eicosanoides released during the inflammatory state, in particular leukotriene D4, also induce autocrine stimulation of cysteinyl leukotriene receptor 1 (CysLT1), that is a strong constricting agent and might counteract the relaxing effects of taurine on smooth muscle cells thereby worsening lung conditions.40 On the other hand, since taurine accumulates intracellularly in airway epithelia to regulate osmotic balance, its increased concentrations in BALf samples of BOS patients with different severity may reflect inflammationinduced osmotic stresses or epithelial cell damage. Moreover, taurine accumulates in the cytoplasm of neutrophils and other leukocytes, and its presence may reflect the presence of these cells in the inflamed airways.15

Limitations of the work We are aware that, to define which, among the metabolites identified, may be the best biomarker(s) of this disease, our “model” would have included also BALf samples from BOS II and BOS III


Page 20 of 45

Page 21 of 45

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


patients. Unfortunately, the intrinsic difficulty to obtain specimens from individuals severely affected and the very poor number of patients belonging to these categories available in our structure did not allow us to perform this comprehensive study. Moreover, while roughly the same volumes of fluid are injected and then picked up from each patient, the NMR spectra may differ significantly in the "total" content of metabolites due to the level of airway obstruction that can be significantly different among patients. However, while being an important step to confirm the speculations discussed above, the analysis of these samples does not seem essential. In fact, if the excellent fit and prediction values between S and BOS I prove reliable, all together or a limited number of the metabolites identified may be the hallmark of the disorder. Thus, although there is still a long way to go, it seems obvious to emphasize that the identification of the disease at a very early stage would be of the utmost importance to set up immediately a patient's treatment program.

CONCLUSIONS We have generated, for the first time, a profile of metabolites in human BALf that could be relevant for a better understanding of the biological activities involved in the transition from a state of stability to that of established Bronchiolitis Obliterans Syndrome in patients after lung transplant. This pilot study indeed demonstrates that the NMR profiles not only discriminate BOS from other pulmonary disorders, but can also differentiate stable from BOS I patients. This aspect is crucial for an early identification of the disorder onset. The fact that BOS 0p patients can be separated much less efficiently by stable subjects and by BOS I patients, rather than being a limitation, in our opinion confirms the reliability of the work. In fact, it shows how the method is “sensitive” to the change of clinical conditions that occur during the evolution from stability to the established disease. The closeness of these group pairs in terms of FEV1 and of medical conditions in general, did not allow their unambiguous discrimination. Nevertheless, data are promising and we are



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

confident that sufficient “sensitivity” to discriminate also these cohorts (for example by increasing the numerosity of samples) will soon be attained. The possibility to include in our study also significant numbers of BOS II and BOS III patients would be the essential step to confirm this speculation and to allow identifying which, among the metabolites identified, may be the best biomarker(s) of this disease. Obviously, identification of these markers and mechanistic analysis of their role will open the way to gain insights on the molecular mechanisms that govern the evolution of the disease. Our data, although partial, also suggest that future studies may rely on the use of BAL. The experimental advantage would be to avoid processing steps that are time-consuming and, if not correctly performed, may introduce biases affecting reproducibility.

SUPPORTING INFORMATION Table S1. Cytological features of all included BAL samples. Table S2. Most relevant metabolite concentrations in BALf samples from stable subjects, BOS 0p and BOS I patients. Figure S1. 1H-NMR spectra of BALf samples containing the metabolites sucrose and glycerol. Figure S2. Aliphatic region of the 1H,1H-TOCSY spectrum of a representative BALf sample with the assignment of most abundant metabolites. Figure S3. Comparison of 1H-NMR profiles of representative BALf samples collected from patients suffering from different pulmonary pathologies. Figure S4. PCA loading plot for the comparison between S, BOS 0p and BOS I. Figure S5. OPLS-DA loading plot for the comparison between S, BOS 0p and BOS I. Figure S6. Clustering results shown as for S and BOS I sample comparison. 22 ACS Paragon Plus Environment

Page 22 of 45

Page 23 of 45

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure S7. PCA and OPLS-DA score plots between BOS 0p and S. Figure S8. PCA and OPLS-DA score plots between BOS 0p and BOS I. Figure S9. Comparison between spectra acquired on BAL and BALf samples.

ACKNOWLEDGEMENTS This study was funded by Fondazione CARIPLO (Milan, Italy) under project agreement 2013-0820 (BALf metabolomics in chronic lung rejection: an innovative approach to identify predictive markers and sub-phenotypes). Dr. Cristina Airoldi thanks Fondazione CARIPLO for funding project 2015-0763. The authors are grateful to Prof. Dario Pescini and Prof. Nadia Solaro (Department of Statistics and Quantitative Methods, University of Milano-Bicocca) for the helpful discussion about statistical analysis.

REFERENCES (1) Belperio, J. A.; Lake, K.; Tazelaar, H.; Keane, M. P.; Strieter, R. M.; Lynch, J. P. Bronchiolitis Obliterans Syndrome Complicating Lung or HeartLung Transplantation. Semin. Respir. Crit. Care Med. 2003, 24 (5), 1-47. (2) Verleden, S. E.; Ruttens, D.; Vos, R.; Vandermeulen, E.; Moelants, E.; Mortier, A.; Van Raemdonck, D. E.; Proost, P.; Schols, D.; Verleden, G. M.; Vanaudenaerde, B. M. Differential cytokine, chemokine and growth factor expression in phenotypes of chronic lung allograft dysfunction. Transplantation. 2015, 99, 86-93. (3) Meloni, F.; Salvini, R.; Bardoni, A. M.; Passadore, I.; Solari, N.; Vitulo, P.; Oggionni, T.; Viganò, M.; Pozzi, E.; Fietta, A. M. Bronchoalveolar lavage fluid proteome in bronchiolitis



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

obliterans syndrome: possible role for surfactant protein A in disease onset. J. Heart Lung Transplant. 2007, 26, 1135-1143. ( 4 ) Bonneau, E.; Tétreault, N.; Robitaille, R.; Boucher, A.; De Guire, V. Metabolomics: Perspectives on potential biomarkers in organ transplantation and immunosuppressant toxicity. Clinical Biochem. 2016, 49, 377-384. (5) Benahmed, M. A.; Santelmo, N.; Elbayed, K.; Frossard, N.; Noll, E.; Canuet, M.; Pottecher, J.; Diemunsch, P.; Piotto, M.; Massard, G.; Namer, I. J. The Assessment of the Quality of the Graft in an Animal Model for Lung Transplantation Using the Metabolomics 1H High-Resolution Magic Angle Spinning NMR Spectroscopy. Magnetic Resonance in Medicine. 2012, 68, 1026–1038. (6) Nobakht M. Gh., B. F.; Aliannejad, R.; Rezaei-Tavirani, M.; Taheri, S.; Oskouie, A. A. The metabolomics of airway diseases, including COPD, asthma and cystic fibrosis. Biomarkers. 2015, 20, 5-16. (7) Beckonert, O.; Keun, H. C.; Ebbels, T. M. D.; Bundy, J.; Holmes, E.; Lindon, J. C.; Nicholson, J. K. Metabolic profiling, metabolomic and metabonomic procedures for NMR spectroscopy of urine, plasma, serum and tissue extracts. Nature Protocols. 2007, 2, 2692-703. (8) Azmi, J.; Connelly, J.; Holmes, E.; Nicholson, J. K.; Shore, R. F.; Griffin, J. L. Characterization of the biochemical effects of 1-nitronaphthalene in rats using global metabolic profiling by NMR spectroscopy and pattern recognition. Biomarkers. 2005, 10, 401-16. (9) Hong, J. H.; Lee, W. C.; Hsu, Y. M.; Liang, H. J.; Wan, C. H.; Chien, C. L.; Lin, C. Y. Characterization of the biochemical effects of naphthalene on the mouse respiratory system using NMR-based metabolomics. J. Appl. Toxicol. 2014, 34, 1379–1388. (10) Lin, C. Y.; Huang, F. P.; Ling, Y. S.; Liang, H. J.; Lee, S. H.; Hu, M. Y.; Tsao, P. N. Use of nuclear magnetic resonance-based metabolomics to characterize the biochemical effects of naphthalene on various organs of tolerant mice. PLoS One. 2015, 10, e0120429.


Page 24 of 45

Page 25 of 45

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


(11) Singh, C.; Rai, R. K.; Azim, A.; Sinha, N.; Baronia, A. K. Mini-bronchoalveolar lavage fluid can be used for biomarker identification in patients with lung injury by employing 1H NMR spectroscopy. Critical Care. 2013, 17, 430. ( 12 ) Viswan, A.; Sharma, R. K.; Azim, A.; Sinha, N. NMR-based metabolic snapshot from minibronchoalveolar lavage fluid: an approach to unfold human respiratory metabolomics. J. Proteome Res. 2016, 15, 302−310. (13) Rai, R. K.; Azim, A.; Sinha, N.; Sahoo, J. N.; Singh, C.; Ahmed, A.; Saigal, S.; Baronia, A. K.; Gupta, D.; Gurjar, M.; Poddar B.; Singh, R. K. Metabolic profiling in human lung injuries by highresolution nuclear magnetic resonance spectroscopy of bronchoalveolar lavage fluid (BALF). Metabolomics. 2013, 9, 667–676. (14) Fabiano, A.; Gazzolo, D.; Zimmermann, L. J.; Gavilanes, A. W.; Paolillo, P.; Fanos, V.; Caboni, P.; Barberini, L.; Noto, A.; Atzori, L. Metabolomic analysis of bronchoalveolar lavage fluid in preterm infants complicated by respiratory distress syndrome: preliminary results. J. Matern. Fetal Neonatal Med. 2011, 2, 55-8. (15) Wolak, J. E.; Esther, C. R. Jr.; O’Connell, T. M. Metabolomic analysis of bronchoalveolar lavage fluid from cystic fibrosis patients. Biomarkers. 2009, 14 (1), 55-60. (16) Glasoe, P. K.; Long, F. A. Use of glass electrodes to measure acidities in deuterium oxide. J. Phys. Chem. 1960, 64, 188-9. (17) Wishart, D. S.; Tzur, D.; Knox, C.; Eisner, R.; Guo, A. C.; Young, N.; Cheng, D.; Jewell, K.; Arndt, D.; Sawhney, S.; Fung, C.; Nikolai, L.; Lewis, M.; Coutouly, M. A.; Forsythe, I.; Tang, P.; Shrivastava, S.; Jeroncic, K.; Stothard, P.; Amegbey, G.; Block, D.; Hau, D. D.; Wagner, J.; Miniaci, J.; Clements, M.; Gebremedhin, M.; Guo, N.; Zhang, Y.; Duggan, G. E.; Macinnis, G. D.; Weljie, A. M.; Dowlatabadi, R.; Bamforth, F.; Clive, D.; Greiner, R.; Li, L.; Marrie, T.; Sykes, B.



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

D.; Vogel, H. J.; Querengesser, L. HMDB: the Human Metabolome Database. Nucleic Acids Res. 2007, 35, D521-6. (18) Wishart, D. S.; Knox, C.; Guo, A. C.; Eisner, R.; Young, N.; Gautam, B.; Hau, D. D.; Psychogios, N.; Dong, E.; Bouatra, S.; Mandal, R.; Sinelnikov, I.; Xia, J.; Jia, L.; Cruz, J. A.; Lim, E.; Sobsey, C. A.; Shrivastava, S.; Huang, P.; Liu, P.; Fang, L.; Peng, J.; Fradette, R.; Cheng, D.; Tzur, D.; Clements, M.; Lewis, A.; De Souza, A.; Zuniga, A.; Dawe, M.; Xiong, Y.; Clive, D.; Greiner, R.; Nazyrova, A.; Shaykhutdinov, R.; Li, L.; Vogel, H. J.; Forsythe, I. HMDB: a knowledgebase for the human metabolome. Nucleic Acids Res. 2009, 37, D603-10. (19) Wishart, D. S.; Jewison, T.; Guo, A. C.; Wilson, M.; Knox, C.; Liu, Y.; Djoumbou, Y.; Mandal, R.; Aziat, F.; Dong, E.; Bouatra, S.; Sinelnikov, I.; Arndt, D.; Xia, J.; Liu, P.; Yallou, F.; Bjorndahl, T.; Perez-Pineiro, R.; Eisner, R.; Allen, F.; Neveu, V.; Greiner, R.; Scalbert, A. HMDB 3.0-The Human Metabolome Database in 2013. Nucleic Acids Res. 2013, 41, D801-7. (20) Ulrich, E. L.; Akutsu, H.; Doreleijers, J. F.; Harano, Y.; Ioannidis, Y. E.; Lin, J.; Livny, M.; Mading, S.; Maziuk, D.; Miller, Z.; Nakatani, E.; Schulte, C. F.; Tolmie, D. E.; Wenger, R. K.; Yao, H.; Markley, J. L. BioMagResBank. Nucleic Acids Res. 2008, 36, D402-8. (21) Cobas, S.; Seoane, F.; Dominguez, S.; Davies, A. N. A new approach to improving automated analysis of proton NMR. Spectrosc. Eur. 2010, 23, 26-30. (22) Airoldi, C.; Ciaramelli, C.; Fumagalli, M.; Bussei, R.; Mazzoni, V.; Viglio, S.; Iadarola, P.; Stolk, J. 1H NMR to Explore the Metabolome of Exhaled Breath Condensate in α1-Antitrypsin Deficient Patients: A Pilot Study. J. Proteome Res. 2016, 15 (12), 4569–4578. (23) Gallo, V.; Intini, N.; Mastrorilli, P.; Latronico, M.; Scapicchio, P.; Triggiani, M.; Bevilacqua, V.; Fanizzi, P.; Acquotti, D.; Airoldi, C.; Arnesano, F.; Assfalg, M.; Benevelli, F.; Bertelli, D.; Cagliani, L. R.; Casadei, L.; Cesare Marincola, F.; Colafemmina, G.; Consonni, R.; Cosentino, C.; Davalli, S.; Angelica, S.; De Pascali, S. A.; D'Aiuto, V.; Faccini, A.; Gobetto, R.; Lamanna, R.;


Page 26 of 45

Page 27 of 45

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Liguori, F.; Longobardi, F.; Mallamace, D.; Mazzei, P.; Menegazzo, I.; Milone, S.; Mucci, A.; Napoli, C.; Pertinhez, T. A.; Rizzuti, A.; Rocchigiani, L.; Schievano, E.; Sciubba, F.; Sobolev, A. S.; Tenori, L.; Valerio, M. Performance assessment in fingerprinting and multi component quantitative NMR analyses. Anal. Chem. 2015, 87 (13), 6709–6717. ( 24 ) Gallo, V.; Scapicchio, P.; Vezzoso, M.; Fanizzi, P. Risonanza Magnetica e Controllo Qualità−Presentazione dei Risultati del Primo Confront Interlaboratorio Nazionale; Camera di Commercio

di

Bari,

2013

(available

free

of

charge

at

http://www.samer.it/articoli/allegato/Samer_rev1_def.pdf and at http://www.innovativesolutions. it/wp-content/uploads/2013/09/Pubblicazione-con-ISBN.pdf). (25) Xia, J.; Sinelnikov, I.; Han, B.; Wishart, D. S. MetaboAnalyst 3.0 - making metabolomics more meaningful. Nucl. Acids Res. 2015, 43, 251-257. (26) Xia, J.; Mandal, R.; Sinelnikov, I.; Broadhurst, D.; Wishart, D. S. MetaboAnalyst 2.0 - a comprehensive server for metabolomic data analysis. Nucl. Acids Res. 2012, 40, 127-133. ( 27 ) Xia, J.; Psychogios, N.; Young, N.; Wishart, D. S. MetaboAnalyst: a web server for metabolomic data analysis and interpretation. Nucl. Acids Res. 2009, 37, 652-660. (28) Lane, A. N.; Arumugam, S. Improving NMR sensitivity in room temperature and cooled probes with dipolar ions. Journal of Magnetic Resonance. 2005, 173, 339–343. (29) Eriksson, L.; Johansson, E.; Kettaneh-Wold, N.; Trygg, J.; Wikstrom, C.; Wold, S. Multi- and Megavariate Data analysis; Umetrics Academy: Umeå, Sweden, 2006; Vol. 1. (30) Lama, V. N.; Murray, S.; Mumford, J. A.; Flaherty, K. R.; Chang, A.; Toews, G. B.; PetersGolden, M.; Martinez, F. J. Prognostic value of bronchiolitis obliterans syndrome stage 0-p in single-lung transplant recipients. Am. J. Respir. Crit. Care Med. 2005, 172, 379-383. (31) De Backer, D.; Creteur, J.; Zhang, H.; Norrenberg, M.; Vincent, J. L. Lactate production by the lungs in acute lung injury. Am. J. Respir. Crit. Care Med. 1997, 156, 1099–1104.



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

(32) Freund, H. R.; Ryan, J. A. Jr.; Fischer, J. E. Amino acid derangements in patients with sepsis: treatment with branched chain amino acid rich infusions. Ann. Surg. 1978, 188, 423–430. (33) Worlitzsch, D.; Tarran, R.; Ulrich, M.; Schwab, U.; Cekici, A.; Meyer, K. C.; Birrer, P.; Bellon, G.; Berger, J.; Weiss, T.; Botzenhart; K.; Yankaskas, J. R.; Randell, S.; Boucher, R. C.; Döring, G. Effects of reduced mucus oxygen concentration in airway Pseudomonas infections of cystic fibrosis patients. J. Clin. Invest. 2002, 109, 317–325. (34) Anderson, J. C.; Lamm, W. J. E.; Hlastala, M. P. Measuring airway exchange of endogenous acetone using a single-exhalation breathing maneuver. J. Appl. Physiol. 2006, 100, 880–889. (35) Casazza, J. P.; Felver, M. E.; Veech, R. L. The metabolism of acetone in rat. J. Biol. Chem. 1984, 259, 231-236. (36) Reichard, G. A. Jr.; Haff, A. C.; Skutches, C. L.; Paul, P.; Holroyde, C. P.; Owen, O. E. Plasma acetone metabolism in the fasting human. J. Clin. Invest. 1979, 63, 619-626. (37) Lambert, I. H. Regulation of the cellular content of the organic osmolyte taurine in mammals. Neurochem. Res. 2004, 29, 27-63. (38) Lambert, I. H.; Hoffmann, E. K.; Pedersen S. F. Cell volume regulation physiology and pathophysiology. Acta Physiol. Scand. 2008, 194, 255-282. (39) Gallos, G.; Yim, P.; Chang, S.; Zhang, Y.; Xu, D.; Cook, J. M.; Gerthoffer, W. T.; Emala, C. W. Targeting the restricted alpha-subunit repertoire of airway smooth muscle GABAA receptor augments airway smooth muscle relaxation. Am. J. Physiol. Lung Cell Mol. Physiol. 2012, 302, 248-256. (40) Lambert, I. H.; Kristensen, D. M.; Holm, J. B.; Mortensen, O. H. Physiological role of taurinefrom organism to organelle. Acta Physiol. 2015, 213, 191-212.


Page 28 of 45

Page 29 of 45

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


TABLES Table 1. Demographic and clinical features of patients (including age, gender, lung transplant indication, immunosuppressive regimen, CLAD occurrence, length of follow-up, other treatment strategies such as azithromycin and ECP). Abbreviations: DL, double lung transplant; SL, single lung transplant; HL, heart-lung transplant; PVOD, pulmonary veno-occlusive disease; NSIP, nonspecific

interstitial

pneumonia;

lymphangioleiomyomatosis;

CPFE,

EAA,

extrinsic

combined

pulmonary

allergic fibrosis

alveolitis; and

LAM,

emphysema;

IS,

immunosuppressive; ECP, extra-corporeal photopheresis; N/A, not applicable.

sample

age

sex

Underlying disease

type of Tx

time of tx (months)

time of CLAD (months)

IS therapy

azitromycin therapy

ECP

S 01

56

M

CPFE

DL

13,37

N/A

Tacrolimus, Mycophenolate mofetil, Prednisone

no

no

S 02

64

M

Idiopathic pulmonary fibrosis

SL dx

3,33

N/A

Tacrolimus, Prednisone

no

no

S 03

55

F

PVOD

DL

12,15

N/A

S 04

40

M

Histiocytosis X

DL

66,70

N/A

S 05

41

F

CTEPH

HL

5,78

N/A

S 06

64

M


SL dx

62,67

N/A

S 07

50

F

Cystic fibrosis

DL

83,20

N/A

S 08

66

M

Emphysema

DL

110,13

N/A

S 09

64

M

Emphysema

DL

3,50

N/A

S 10

57

F

LAM

DL

1,91

N/A

BOS 0p 01

37

M

Cystic fibrosis

DL

5,51

N/A

BOS 0p 02

37

M

Cystic fibrosis

DL

6,27

N/A

BOS 0p 03

40

F

Bronchiectasis

DL

185,18

N/A

BOS 0p 04

53

M

Cystic fibrosis

DL

199,41

N/A

Tacrolimus, Mycophenolate mofetil, Prednisone Tacrolimus, Everolimus, Prednisone Tacrolimus, Mycophenolate mofetil, Prednisone

no

no

yes

no

yes

no


yes

no

no

no

yes

no

no

no

no

no

yes

no

yes

no

yes

no

yes

no

Tacrolimus, Mycophenolic acid, Prednisone Cyclosporine, Mycophenolic acid, Prednisone Tacrolimus, Mycophenolate mofetil, Prednisone Tacrolimus, Mycophenolate mofetil, Prednisone Tacrolimus, Prednisone Cyclosporine, Prednisone Tacrolimus, Prednisone Tacrolimus, Everolimus, Prednisone



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

BOS 0p 05

62

M

Pulmonary hypertension

DL

51,98

N/A

BOS 0p 06

66

F

Bronchiectasis

DL

40,63

N/A

BOS 0p 07

54

F

PVOD

DL

3,40

N/A

BOS 0p 08

40

F

NSIP

SL dx

20,60

N/A

BOS0p 09

59

M


SL dx

63,70

N/A

BOS 0p 10

59

F

Emphysema

DL

83,83

N/A

BOS I 01

22

M

Cystic fibrosis

DL

83,60

2,54

BOS I 02

32

F

Cystic fibrosis

DL

91,22

59,14

BOS I 03

33

F

Cystic fibrosis

DL

94,98

62,90

BOS I 04

69

F

Emphysema

DL

89,47

64,49

BOS I 05

41

F

NSIP

SL dx

37,82

3,9

SL sx

65,64

46,80

SL dx

148,35

6,01

Idiopathic pulmonary fibrosis Idiopathic pulmonary fibrosis

Page 30 of 45

Tacrolimus, Mycophenolate mofetil, Prednisone Tacrolimus, Mycophenolate mofetil, Prednisone Tacrolimus, Mycophenolate mofetil, Prednisone Tacrolimus, Mycophenolate mofetil, Prednisone Tacrolimus, Everolimus, Prednisone Tacrolimus, Everolimus, Prednisone Tacrolimus, Mycophenolate mofetil, Prednisone Tacrolimus, Mycophenolate mofetil, Prednisone Tacrolimus, Mycophenolate mofetil, Prednisone Tacrolimus, Mycophenolate mofetil, Prednisone Tacrolimus, Prednisone Tacrolimus, Mycophenolate mofetil, Prednisone

no

no

yes

no

no

no

no

no

yes

no

yes

no

yes

no

yes

no

yes

no

yes

no

yes

yes

yes

yes

yes

no

yes

no

yes

no

BOS I 06

52

M

BOS I 07

73

M

BOS I 08

42

M

Eisenmenger sd

HL

245,58

63,70

BOS I 09

61

M


SL dx

84,98

26,20

BOS I 10

49

F

Bronchiectasis

DL

147,99

69,83

OP 01

45

M

N/A

N/A

OP 02

45

F

N/A

N/A

OP 03

59

F

Sarcoidosis Pulmonary consolidations Pneumonia

N/A

N/A

N/A

N/A

N/A

N/A

OP 04

51

M

EAA

N/A

N/A

N/A

N/A

N/A

N/A

OP 05

52

F

EAA

N/A

N/A

N/A

N/A

N/A

N/A

OP 06

37

M

EAA

N/A

N/A

N/A

N/A

N/A

N/A

OP 07

44

F

EAA

N/A

N/A

N/A

N/A

N/A

N/A

OP 08

44

M

Sarcoidosis

N/A

N/A

N/A

N/A

N/A

N/A

OP 09

38

M

Sarcoidosis

N/A

N/A

N/A

N/A

N/A

N/A

OP 10

34

F

Sarcoidosis

N/A

N/A

N/A

N/A

N/A

N/A


yes

no

N/A

Tacrolimus, Mycophenolate mofetil, Prednisone Tacrolimus, Mycophenolate mofetil, Prednisone Tacrolimus, Mycophenolate mofetil, Prednisone N/A

N/A

N/A

N/A

N/A

N/A

N/A


Page 31 of 45

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Table 2. Assignment of the metabolites identified in BALf from stable and BOS patients. 1

Metabolite

Assignment

1

Acetate

CH3

1,92 (s)

2

Acetone

2 x CH3

2,22 (s)

3

Adenine

C8H

8,19 (s)

C2H

8,21 (s)

βCH3

1,47 (d)

αCH

3,78 (q)

γCH2

1,65 (m)

βCH2

1,90 (m)

δCH2

3,24 (t)

αCH

3,77 (t)

βCH2

2,86-2,94 (m)

αCH

4,00 (dd)

βCH2

2,67-2.80 (m)

αCH

3,89 (m)

N-(CH3)3

3,25 (s)

CH2

3,89 (s)

N-(CH3)3

3,20 (s)

βCH2

3,52 (m)

αCH2

4,06 (s)

CH3

3,04 (s)

CH2

3,95 (s)

CH3

3,03 (s)

CH2

4,06 (s)

O-CH2

3,14 (t)

N-CH2

3,81 (t)

4 5

6 7 8 9

10 11 12

Alanine Arginine

Asparagine Aspartate Betaine Choline

Creatine Creatinine Ethanolamine

H chemical shift

13

Fatty acids

CH3

0,88 (t)

14

Formate

CH

8,45 (s)

15

α-Glucose

C4H

3,40 (m)

C2H

3,54 (m)

C3H

3,71 (m)

C6H2

3,76-3,83 (m)

C5H

3,85 (m)

C1H

5,25 (d)

C2H

3,24 (m)

C4H

3,41 (m)

C5H

3,48 (m)

16

β-Glucose



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

17

18

19

Glutamate

Glutamine

Glycerol

Page 32 of 45

C3H

3,49 (m)

C6H2

3,73-3,89 (m)

C1H

4,65 (d)

βCH2

2,08 (m)

γCH2

2,34 (m)

αCH

3,75 (m)

βCH2

2,12 (m)

γCH2

2,44 (m)

αCH

3,77 (t)

CH2

3,54 (dd)

CH2

3,64 (dd)

CH

3,77 (dd)

20

Glycine

αCH2

3,56 (s)

21

Histidine

βCH2

3,10-3,22 (m)

αCH

3,99 (m)

C5H ring

7,07 (s)

C2H ring

7,80 (s)

δCH3

0,93 (t)

22

Isoleucine

2

1,00 (d)

1

γ CH2

1,25- 1,45 (m)

βCH

1,96 (m)

αCH

3,67 (m)

CH3

1,32 (d)

CH

4,11 (q)

2 x δCH3

0,95 (m)

γCH

1,70 (m)

βCH2

1,72 (m)

αCH

3,74 (t)

γCH2

1,47 (m)

δCH2

1,72 (m)

βCH2

1,88 (m)

εCH2

3,03 (t)

αCH

3,74 (t)

S-CH3

2,13 (s)

βCH2

2,16 (m)

γCH2

2,63 (t)

αCH

3,86 (t)

βCH2

3,12-3,28 (m)

αCH

3,98 (m)

γ CH3

23 24

25

26

27

Lactate Leucine

Lysine

Methionine

Phenylalanine


Page 33 of 45

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


28

Proline

C2H, C6H ring

7,32 (d)

C4H ring

7,37 (m)

C3H, C5H ring

7,42 (t)

γCH2

2,00 (m)

βCH2

2,06-2,36 (m)

δCH2

2,34-3,42 (m)

αCH

4,13 (m)

29

Pyruvate

CH3

2,36 (s)

30

Serine

αCH

3,84 (dd)

βCH2

3,95-3,99 (m)

31

Succinate

2 x CH2

2,39 (m)

32

Sucrose

C4H

3,47 (t)

C2H

3,56 (dd)

C1'H2

3,67 (s)

C3H

3,76 (t)

C6H2, C6'H2

3,81-3,84 (m)

C5H, C5'H

3,87-3,89 (m)

C4'H

4,04 (t)

C3'H

4,21 (d)

C1H

5,41 (d)

S-CH2

3,26 (t)

N-CH2

3,42 (t)

γCH3

1,32 (d)

αCH

3,59 (d)

βCH

4,25 (dq)

βCH2

3,47-3,29 (m)

αCH

4,04 (m)

C5H, C6H ring

7,19 (t)

C5H, C6H ring

7,28 (t)

C2H ring

7,31 (s)

C7H ring

7,54 (d)

C4H ring

7,73 (d)

βCH2

3,05-3,19 (m)

αCH

3,94 (m)

C3H, C5H ring

6,9 (d)

C2H, C6H ring

7,18 (d)

33 34

35

36

Taurine Threonine

Tryptophan

Tyrosine

37

Urea

NH2

5,79 (m)

38

Valine

2 x γCH3

0,98 (d) - 1,04 (d)

βCH

2,27 (m)



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

39

Lidocaine *

Page 34 of 45

αCH

3,61 (d)

H aromatic

7,25 (dd) - 7,21 (d)

CH2

4,21 (s)

CH2

3,32 (q)

CH3

2,19 (s)

CH3 aromatic * This is a local anesthetic used during the medical procedure.

1,35 (t)

Table 3. Absolute concentrations (µM) of the metabolites that are mainly responsible for the three classes (S, BOS 0p and BOS I) discrimination, represented as mean values ± SEM for each group.a

Acetate Acetone Alanine Creatine Ethanolamine Formate Glucose Glutamate Glycerol Isoleucine Lactate Leucine Taurine Threonine Valine a

STABLE Mean SEM 0.583 0.204 0.207 0.073 0.353 0.130 0.116 0.043 0.250 0.122 0.250 0.068 0.250 0.568 2.099 0.698 2.004 0.710 0.439 0.249 1.652 0.536 0.907 0.380 1.273 0.458 1.637 0.590 0.475 0.246

BOS 0p Mean SEM 1.176 0.326 0.337 0.120 0.591 0.181 0.408 0.148 0.873 0.225 0.873 0.170 0.873 2.135 5.180 1.787 5.167 1.227 0.194 0.088 4.351 0.800 0.644 0.215 5.640 1.424 2.524 0.804 0.323 0.097

BOS I Mean SEM 0.928 0.228 0.790 0.376 1.173 0.381 0.392 0.070 1.265 0.323 1.265 0.150 1.265 1.826 8.857 2.563 6.738 1.083 0.444 0.113 7.387 1.697 1.834 0.512 6.619 1.083 3.510 0.796 0.436 0.101

NMR signals in each class were integrated and referred to the final DSS signal of known

concentration (0.1 mM) as determined by NMR signal integration.

FIGURES


Page 35 of 45

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 1. 1H-NMR spectra of the same BALf sample: A) at the “original” concentration (1) and concentrated 5x (2); B) stored for 18 months (1) and 30 months (2); C) stored at -80 °C (1) and at 20 °C (2); D) acquired with different acquisition sequences: 1D NOESY-presat sequence (1) and CPMG relaxation-editing sequence (2). Arrows in spectrum 1 indicate broad resonances belonging to macromolecules.



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Figure 2. Expansion of different regions of the

1

H-NMR spectrum of a BALf sample

(representative of all others) with the assignment of the most abundant metabolites. His, histidine; Trp, tryptophane; Phe, phenylalanine; Tyr, tyrosine; Glc, glucose; Thr, threonine; Pro, proline; Ser, serine; Gly, glycine; Lys, lysine; Asn, asparagine; Asp, aspartate; Gln, glutamine; Glu, glutamate; Arg, arginine; Ala, alanine; Val, valine; Ile, isoleucine; Leu, leucine; FA, fatty acids.


Page 36 of 45

Page 37 of 45

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 3. A) PCA and B) OPLS-DA score plots between BOS I and other pathologies (OPP) samples. The explained variances are shown in brackets. Ellipses display 95% confidence regions. In red are BOS I, in green OPP samples. 37 ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Figure 4. A) PCA and B) OPLS-DA score plots between S, BOS 0p and BOS I samples. The explained variances are shown in brackets. Ellipses display 95% confidence regions. In red are BOS 0p, in green BOS I and in blue S samples. 38 ACS Paragon Plus Environment

Page 38 of 45

Page 39 of 45

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 5. Clustering results shown as dendrogram (distance measure using euclidean, and clustering algorithm using ward.D) for S, BOS 0p and BOS I sample comparison.



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Figure 6. Graphical representation of the absolute concentrations (µM) of the metabolites that are mainly responsible for the discrimination of the three classes (S, BOS 0p and BOS I). For each class, metabolite concentrations are reported as mean values ± SEM.


Page 40 of 45

Page 41 of 45

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 7. A) PCA and B) OPLS-DA score plots obtained by comparing BOS I and stable (S) samples. The explained variances are shown in brackets. Ellipses display 95% confidence regions. In red are BOS I, in green S samples. 41 ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Figure 8. Summary of Pathway Analysis.


Page 42 of 45

Page 43 of 45

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 9. Possible metabolic pathway of C3 metabolites identified in BALf of BOS patients.



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

For Table of Contents Only


Page 44 of 45

Page 45 of 45

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


84x38mm (300 x 300 DPI)

ACS Paragon Plus Environment

1H NMR To Evaluate the Metabolome of Bronchoalveolar Lavage

Recommend Documents