Metabolomic fingerprinting and systemic inflammatory profiling of asthma COPD overlap (ACO)

Background Asthma-COPD overlap (ACO) refers to a group of poorly studied and characterised patients reporting with disease presentations of both asthma and COPD, thereby making both diagnosis and treatment challenging for the clinicians. They exhibit a higher burden in terms of both mortality and morbidity in comparison to patients with only asthma or COPD. The pathophysiology of the disease and its existence as a unique disease entity remains unclear. The present study aims to determine whether ACO has a distinct metabolic and immunological mediator profile in comparison to asthma and COPD. Methods Global metabolomic profiling using two different groups of patients [discovery (D) and validation (V)] were conducted. Serum samples obtained from moderate and severe asthma [n = 34(D); n = 32(V)], moderate and severe COPD [n = 30(D); 32(V)], ACO patients [n = 35(D); 40(V)] and healthy controls [n = 33(D)] were characterized using gas chromatography mass spectrometry (GC-MS). Multiplexed analysis of 25 immunological markers (IFN-γ (interferon gamma), TNF-α (tumor necrosis factor alpha), IL-12p70 (interleukin 12p70), IL-2, IL-4, IL-5, IL-13, IL-10, IL-1α, IL-1β, TGF-β (transforming growth factor), IL-6, IL-17E, IL-21, IL-23, eotaxin, GM-CSF (granulocyte macrophage-colony stimulating factor), IFN-α (interferon alpha), IL-18, NGAL (neutrophil gelatinase-associated lipocalin), periostin, TSLP (thymic stromal lymphopoietin), MCP-1 (monocyte chemoattractant protein- 1), YKL-40 (chitinase 3 like 1) and IL-8) was also performed in the discovery cohort. Results Eleven metabolites [serine, threonine, ethanolamine, glucose, cholesterol, 2-palmitoylglycerol, stearic acid, lactic acid, linoleic acid, D-mannose and succinic acid] were found to be significantly altered in ACO as compared with asthma and COPD. The levels and expression trends were successfully validated in a fresh cohort of subjects. Thirteen immunological mediators including TNFα, IL-1β, IL-17E, GM-CSF, IL-18, NGAL, IL-5, IL-10, MCP-1, YKL-40, IFN-γ, IL-6 and TGF-β showed distinct expression patterns in ACO. These markers and metabolites exhibited significant correlation with each other and also with lung function parameters. Conclusions The energy metabolites, cholesterol and fatty acids correlated significantly with the immunological mediators, suggesting existence of a possible link between the inflammatory status of these patients and impaired metabolism. The present findings could be possibly extended to better define the ACO diagnostic criteria, management and tailoring therapies exclusively for the disease.


GC-MS Data Acquisition
Two μl of derivatized serum sample was loaded using splitless mode to RTx-5 column (5% diphenyl, 95% dimethylpolysiloxane; 30 m × 0.25 mm ×0.25 μm; Restek USA) in a GC−MS (7890A GC, 5975 MSD from Agilent Technologies, USA) for separation using an automatic liquid sampler (7683B ALS, Agilent, USA). Helium was used as a carrier gas at a constant flow rate of 1 mL/min. The front inlet temperature was fixed at 250 °C during injection; temperature gradients of 50 to 150 °C (ramp of 10 °C/min) and 150 to 310 °C (ramp of 7 °C/min) with a hold time of 3 min between two ramps and after reaching final temperature were used. Electron ionization (EI) mode was fixed at −70 eV with a scan range of 35 to 600 m/z. Maximum scan speed was 5 spectra/sec with a 6 min solvent delay. The ion-source temperature and quadrupole temperatures were fixed at 230 and 150 °C, respectively. Sample introduction to data acquisition parameters (both GC separation and mass spectrometry) were controlled using ChemStation software (Agilent Technologies, USA), and the run time was 38.43 minutes per sample. Instrument performance over time and metabolite extraction efficiency were evaluated using peak area and retention time of internal standard in samples and quality controls (QCs). Further, as quality check, a mixture of metabolite standards at a known concentration (25 ng/10 μl) was injected after every 8 samples.

Data Pre-processing
Before analysis, the sample codes were opened by a team member not participating in sample processing and GC−MS data acquisition. GC-MS metabolite profiles were processed using Agilent Chemstation data analysis software. Chromatographic processing such as integration and convolution was performed in Agilent ChemStation software using automatic spectral deconvolution (AMDIS) algorithm. The detectable spectral features after background subtraction were annotated using NIST 14 standard mass spectral database (NIST, Gaithersburg, MD). Consistent metabolites with minimum 30% base peak intensity were considered for quantitation. Features with database matching percentage above 80% were only considered for further analysis. After chromatographic integration, peak areas of corresponding metabolites were noted. The peak areas of annotated features were extracted as a data matrix in .CSV file format.

Metabolomic data analysis using multivariate and univariate statistical analysis
Initially, the peak area metabolomic dataset generated post pre-processing was filtered.
Features with >50% missing values were removed from the data. The resulting data were further normalised to constant sum, log transformed and mean scaled using Metaboanalyst 4.0. These data pre-processing strategies such as transformation and scaling help in making features comparable. This normalised data was further subjected to univariate (UVA) and multivariate statistical analysis (MVA).
In order to reveal the pattern and draw conclusion from the metabolomics study, the normalised data files were independently subjected to MVA using SIMCA 13.0.1 software (Umetrics, Sweden) [1]. At first, principal component analysis (PCA) (an unsupervised multivariate statistical approach) was used to generate an overview of data distribution across samples and detect possible outliers. PCA was also performed on the QC samples to ensure the data quality was not compromised. Subsequently, supervised multivariate statistical tools i.e. partial least squares discriminant analysis (PLS-DA) and orthogonal projections to latent structures discriminant analysis (OPLS-DA) were applied to enhance group separation.
Parameters including R2 (goodness of the fit), Q2 (predictive ability), and analysis of variance testing of cross validated predictive residuals (CV-ANOVA) score were used to detect robustness of the OPLS-DA model [2,3]. The model robustness was evaluated by performing permutation of the model and comparing it with 200 randomly permutated models. Significant metabolites for group separation in OPLS-DA model were identified using variable importance in the projection (VIP) score. Metabolites with VIP score above 1.3 were considered relevant for group discrimination.
In addition to MVA, UVA was performed to assess statistically significant differences in the expression levels of these metabolites. Statistical significance of the metabolites between the groups was obtained using one way analysis of variance (ANOVA) (Dunnett's post hoc test) or Kruskal-Wallis test (Dunn's post hoc test) (GraphPad Prism version 7.00 for Windows, GraphPad Software, San Diego, CA, USA). Statistical significance was considered to be p ≤0.05. The metabolites were also adjusted for multiple hypothesis testing using FDR correction using Metaboanalyst 4.0. Fold change (FC) was calculated for ACO vs asthma and ACO vs COPD. Metabolites common to both ACO vs COPD and ACO vs asthma which collectively qualify the criteria of VIP, p-value, and FDR were considered significant. These significantly altered metabolites identified in the discovery phase were further validated in a fresh cohort of subjects by performing quantitative UVA post data matrix generation.