Skip to main content
PMC home page
Chest logoLink to Chest
. 2016 Oct 21;151(2):262–277. doi: 10.1016/j.chest.2016.10.008

Asthma Metabolomics and the Potential for Integrative Omics in Research and the Clinic

Rachel S Kelly a, Amber Dahlin a, Michael J McGeachie a, Weiliang Qiu a, Joanne Sordillo c, Emily S Wan a,b, Ann Chen Wu c, Jessica Lasky-Su a,∗
PMCID: PMC5310123  PMID: 27776981

Abstract

Asthma is a complex disease well-suited to metabolomic profiling, both for the development of novel biomarkers and for the improved understanding of pathophysiology. In this review, we summarize the 21 existing metabolomic studies of asthma in humans, all of which reported significant findings and concluded that individual metabolites and metabolomic profiles measured in exhaled breath condensate, urine, plasma, and serum could identify people with asthma and asthma phenotypes with high discriminatory ability. There was considerable consistency across the studies in terms of the reported biomarkers, regardless of biospecimen, profiling technology, and population age. In particular, acetate, adenosine, alanine, hippurate, succinate, threonine, and trans-aconitate, and pathways relating to hypoxia response, oxidative stress, immunity, inflammation, lipid metabolism and the tricarboxylic acid cycle were all identified as significant in at least two studies. There were also a number of nonreplicated results; however, the literature is not yet sufficiently developed to determine whether these represent spurious findings or reflect the substantial heterogeneity and limited statistical power in the studies and their methods to date. This review highlights the need for additional asthma metabolomic studies to explore these issues, and, further, the need for standardized methods in the way these studies are conducted. We conclude by discussing the potential of translation of these metabolomic findings into clinically useful biomarkers and the crucial role that integrated omics is likely to play in this endeavor.

Key Words: asthma, biomarkers, integrative omics, metabolomics, proteomics

Abbreviations: AUC, area under the curve; EBC, exhaled breath condensate; MS, mass spectrometry; NMR, nuclear magnetic resonance spectroscopy; PLS-DA, partial least squares- discriminant analysis; SNP, single nucleotide polymorphism; VOC, volatile organic compounds

Asthma is a complex disease with both environmental and genetic influences; however, the role of molecular determinants as mediators of asthma is not yet fully understood.1 Metabolomics, the systematic analysis of small molecules, including carbohydrates, amino acids, organic acids, nucleotides, and lipids, has identified new biomarkers and novel pathogenic pathways for a number of complex chronic diseases.2 Metabolomics is well-suited to the study of diseases with an environmental etiological component because it has the potential to capture the history of the cellular response to past exposures. Metabolite fluctuations represent an integrated pathophysiologic profile encompassing genetic and environmental interactions; therefore, metabolic profiles can be instrumental in elucidating the understanding of the biologic mechanisms of asthma. Although the application of metabolomics to study asthma is recent, the body of literature is rapidly growing. Critical analysis of this literature will afford an improved understanding of the status of asthma metabolomics and help to inform future studies.

Methods

The National Center for Biotechnology Information PubMed database was searched to identify studies of asthma in humans using mass spectrometry (MS) or nuclear magnetic resonance spectroscopy (NMR) to identify and quantify metabolites associated with asthma or asthma-related outcomes. The references of each identified study were evaluated to identify additional qualifying manuscripts. Twenty-one studies using metabolomic profiling of exhaled breathe condensate (EBC) (n = 11),3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13 urine (n = 4),14, 15, 16, 17 serum (n = 3),18, 19, 20 and plasma (n = 3)21, 22, 23 were identified (Table 1). Twelve studies evaluated children,3, 4, 5, 6, 7, 8, 9, 10, 14, 15, 21, 22 eight evaluated adults,12, 13, 16, 17, 18, 19, 20, 23 and one included both.11 The majority used MS-based methods; six used NMR.10, 11, 12, 13, 15, 20 All but four16, 17, 21, 22 were case-control in design, and the total number of people with asthma ranged from 1016 to 343.9 The primary aim of most studies was to examine the differences between asthma cases and healthy control patients, with a smaller number of studies examining disease severity or phenotypes. One study of recurrent wheeze was also included.7

Table 1.

Characteristics of the 21 Asthma Metabolomic Studies Conducted in Humans

Biological Sample Age Group Method Authors No. of Cases No. of Control Patients Diagnostic Criteria Main Aim Population Metabolomic Profiling
EBC Children LC-MS Esther et al (2009)3 11 28 PD Metabolomic profile of asthma vs healthy United States Targeted: adenosine, AMP, and purine biomarkers
Montuschi (2009)4 20 atopic patients without asthma, 25 steroid-naïve atopic mild patients with asthma, 22 atopic mild-to-moderate patients with asthma 15 PD; skin-prick testing Leukotriene profile of asthma vs healthy Italy Targeted: leukotrienes
Carraro et al (2012)5 31 patients with nonsevere asthma, 11 patients with severe asthma 15 PD; GINA guidelines Discrimination of different asthma phenotypes Italy Untargeted
GC-MS Caldeira et al (2012)6 32 atopic patients with asthma 27 PD Metabolomic profile of asthma vs healthy Portugal Targeted: alkanes, alkenes, aldehydes, and ketones
van de Kant et al (2013)7 202 recurrent wheezers 50 ≥ 2 parental-reported episodes of wheeze during life Metabolomic profile of recurrent wheeze vs no recurrent wheeze ADEM study, Netherlands Targeted: VOCs
Gahleitner et al (2013)8 11 12 Health questionnaire; respiratory examination Metabolomic profile of asthma vs healthy United Kingdom Targeted: VOCs
EBC Children GC-MS Smolinska et al (2014)9 343 185 healthy and 546 transient wheezers PD Metabolomic profile of asthma vs transient wheeze ADEM study, Netherlands Targeted: VOCs
NMR Carraro et al (2007)10 17 patients with persistent asthma treated with inhaled corticosteroids, 8 corticosteroid-naïve intermittent patients with asthma 11 PD; GINA guidelines Metabolomic profile of asthma vs healthy Italy Untargeted
All ages NMR Sinha et al (2012)11 7 adults with asthma, 58 children with asthma 10 PD Metabolomic profile of asthma vs healthy India Untargeted
Adults NMR Ibrahim et al (2013)12 82 35 Reported symptoms; treatment Metabolomic profile of asthma vs healthy ASMAL study, United Kingdom Untargeted
Motta et al (2014)13 35 patients with mild asthma 35 PD; GINA guidelines; DSS Metabolomic profile of asthma vs healthy Italy Targeted and untargeted
Urine Children LC-MS Mattarucchi et al (2012)14 41 12 PD; GINA guidelines Metabolomic profile of asthma vs healthy Italy Untargeted
NMR Saude et al (2011)15 73 patients with stable asthma, 20 patients with unstable asthma 42 PD Metabolomic profile of asthma vs healthy, and of different asthma endotypes Canada Targeted
Adults GC-MS Loureiro et al (2014)16 7 patients with allergic asthma, 3 patients with nonallergic asthma NA PD Metabolomic changes with asthma exacerbation Portugal Targeted: aldehydes and alkanes and central metabolites
Loureiro et al (2016)17 57 NA PD Metabolomic profile of asthma severity Portugal Targeted: aliphatic aldehydes and alkanes
Serum Adults GC-MS Ried et al (2013)18 147 2,778 Self-report and medical examination Metabolomic profile of asthma vs healthy KORA Study, Germany Targeted
Chang et al (2015)19 17 patients with mild persistent asthma 17 PD; GINA guidelines Metabolomic profile of asthma vs healthy China Untargeted
Plasma Children LC-MS McGeachie et al (2015)21 20 NA PD Identification of predictors of asthma control CARE Network cohort, United States Targeted lipidomics
Fitzpatrick et al (2014)22 22 patients with mild/moderate asthma, 35 patients with severe asthma NA Spirometry Metabolomic profile of mild-moderate vs severe asthma United States Untargeted
Plasma Adults NMR Jung et al (2013)20 39 26 PD Metabolomic profile of asthma vs healthy South Korea Untargeted and targeted
MS Comhair et al (2015)23 20 10 ATS Workshop on Refractory Asthma Guidelines Metabolomic profile of asthma vs healthy, and of different asthma endotypes United States Untargeted and targeted
Open in a new tab

ADEM = Asthma Detection and Monitoring Study; AMP = adenosine monophosphate; ASMAL = Assessment of Manchester Asthmatics Longitudinally Study; ATS = American Thoracic Society; CARE = Childhood Asthma Research and Education Study; DSS = disease severity score; EBC = exhaled breath condensate; GC-MS = gas chromatography–mass spectrometry; GINA = Global Initiative for Asthma; KORA = Cooperative Health Research in the Region Augsburg Study; LC-MS = liquid chromatography–mass spectrometry; NA = not applicable; NMR = nuclear magnetic resonance spectroscopy; PD = physician diagnosed; VOC = volatile organic compounds.

Results

All 21 studies reported significant findings and concluded that metabolomic profiles in EBC, urine, and blood could distinguish asthma and asthma phenotypes (Table 2). The utility of such profiles is twofold: (1) the identification of metabolite biomarkers for asthma and (2) the improved understanding of the pathophysiology of asthma. The majority of the studies focused on the former by building metabolomic signatures that were subsequently assessed for discriminative ability. These signatures were created by identifying associated metabolites from the total number measured, which ranged by study from two4 to almost 9,000.14 However, the interrogation of the metabolites and pathways composing these signatures also provided important insights into asthma pathophysiology. In this review, we compare the metabolomic signatures and the biological information they impart. In particular, we focus on how different methods and techniques may affect metabolomic signatures, and the implications thereof, as the metabolomics field begins to shift toward clinical translation.

Table 2.

Summary of Results for the 21 Asthma Metabolomic Studies in Humans

Authors No. of Metabolites Results Significant Metabolites Implicated Pathways Conclusions Validation
Esther et al (2009)3 6 Adenosine-to-urea ratio elevated in asthma (median, 1.5) vs control (median, 0.4) (P < .05) Adenosine Neutrophilic airway inflammation EBC adenosine-to-urea ratio is a potential noninvasive biomarker of airways disease No
Montuschi et al (2009)4 2 Exhaled leukotrienes were increased in children with asthma children and were highest in steroid-naive children Leukotrienes Leukotriene-related pathways, inflammatory pathways EBC leukotriene B4 and eicosanoids represent potential noninvasive biomarkers of airway inflammation and therapy monitoring No
Carraro et al (2012)5 NR PLS-DA models could distinguish severe asthma cases from healthy control patients (R2 = 0.93; Q2 = 0.75); and severe from nonsevere asthma cases (R2 = 0.84; Q2 = 0.47). Retinoic acid, adenosine, and vitamin D NR Metabolomic profiling of EBC could clearly distinguish asthmatic children Internal cross-validation
Caldeira et al (2012)6 134 PLS-DA model had a classification rate of 98% and showed 96% sensitivity and 95% specificity for distinguishing patients with asthma from healthy control patients Nonane, 2,2,4,6,6-pentamethylheptane, decane, 3,6-dimethyldecane, dodecane, and tetradecane Oxidative stress and inflammatory processes EBC metabolome is able to accurately distinguish healthy children from children with asthma No
van de Kant et al (2013)7 913 Sparse logistic regression model on the basis of 28 VOCs correctly classified 73% of recurrent wheezers (79% sensitivity, 50% specificity) 28 VOCs NR VOC profiles in EBC are able to distinguish children with and without recurrent wheeze Internal cross-validation
Gahleitner et al (2013)8 NR PLS-DA model on the basis of 8 metabolites distinguished patients with asthma from healthy children with 100% accuracy 1-(methylsulfanyl)propane, ethylbenzene, 1,4-dichlorobenzene, 4-isopropenyl-1-methylcyclohexene, 2-octenal, octadecyne, 1-isopropyl-3-methylbenzene, and 1,7-dimethylnaphtalene NR VOC profiles in EBC are able to distinguish children with and without asthma Internal cross-validation
Smolinska et al (2014)9 NR PLS-DA model on the basis of 17 VOCs distinguished children with asthma from transient wheezers with a prediction rate of 80% Alkanes, acetone, 2,4-dimethylpentane, 2,4-dimethylheptane, 2,2,4-trimethylheptane, 1-methyl-4-(1-methylethenyl) Cyclohexen, 2,3,6-trimethyloctane, 2-undecenal, Biphenyl, 2-ethenylnaphtalene, 2,6,10-trimethyldodecane, Octane, 2-methylpentane, 2,4-dimethylheptane, and 2-methylhexane Oxidative stress and lipid peroxidation VOCs in EBC predict development of asthma Split into a training and test set (80:20)
Carraro et al (2007)10 101 spectral regions NMR-based PLS-DA model distinguished patients with asthma from healthy children with a classification rate of 95% Oxidized and acetylated compounds. Oxidative stress Metabolomic profiling of EBC affords potential for noninvasive biomarker development No
Sinha et al (2012)11 NR Trident peak at 7 ppm reliably distinguishes EBC samples from patients with and without asthma Ammonium ions Glutamate-glutamine cycle Distinct metabolomic profiles of asthmatics and healthy control patients can be identified in NMR-based metabolomic profiling of EBC No
Ibrahim et al (2013)12 367 spectral bins 13 spectral regions discriminated patients with asthma from healthy control patients; AUC, 0.91; overall accuracy, 82.3%; PPV, 83.1%; NPV 78.6% Reported spectral regions NR Distinct metabolomic profiles of patients with asthma and healthy control patients can be identified in NMR-based metabolomic profiling of EBC Split into a training and test set (70:30)
Motta et al (2014)13 NR PLS-DA model distinguished patients with mild asthma from healthy subjects (R2 = 0.90, Q2 = 0.84) Saturated fatty acids, valine, adenosine, hippurate, alanine, formate, urocanic acid, proline, acetate, ethanol, methanol, isoleucine, propionate, 4OH-phenylacetate, tyrosine, arginine, trans-aconitate, and phenylalanine Histidine conversion pathways EBC metabolome is determined by asthma status External validation models (n = 40 drawn from same population)
Mattarucchi et al (2012)14 6,744 features PLS-DA models distinguished patients with well-controlled symptoms (resulting from drugs), well-controlled symptoms (not from drugs), and poorly controlled symptoms (despite using drugs). Prediction rate > 90% for all models Urocanic acid and methyl-imidazoleacetic acid Modulation of immunity LC-MS urinary metabolic profiles can characterize asthma in children Internal cross-validation
Saude et al (2011)15 70 PLS-DA model on the basis of 23 metabolites could distinguish patients with asthma from healthy children; sensitivity, 94%; specificity, 95%; R2= 0.84; Q2 = 0.74 2-oxaloglutarate, succinate, fuma- rate, 3-hydroxy-3-methylglutarate, threonine, and cis-aconitate and trans-aconitate Hypoxia, TCA cycle NMR urinary metabolomic profiles can characterize asthma in children Internal cross-validation
Loureiro et al (2014)16 32 During exacerbations, urine revealed increased levels of aldehydes and alkanes and alterations in a number of nonvolatile metabolites Threonine, lactate, alanine, carnitine, acetylcarnitine, trimethylamine-N-oxide, acetate, citrate, malonate, hippurate, dimethylglycine, and phenylacetylglutamine Oxidative stress, tricarboxylic acid cycle Urinary metabolic composition in asthmatics is highly altered during exacerbations No
Loureiro et al (2016)17 34 Metabolites related to lipid peroxidation levels could predict clinical and laboratory parameters including disease severity, lung function, FeNO, and blood eosinophils in nonobese patients (R2 0.53-0.90) Aliphatic aldehydes and alkanes Lipid peroxidation Metabolomics can provide vital insights into asthma mechanisms Internal cross-validation
Ried et al (2013)18 151 Identified 4 metabolites associated with asthma risk loci and asthma status Phosphatidylcholines, lyso-phosphatidylcholines, PC.aa.C42:2 and PC.aa.C42:4 Lipid metabolism GC-MS serum-based metabolomics affords potential for asthma biomarker development No
Chang et al (2015)19 272 14 discriminatory metabolites were identified. Top metabolite AUCs: 2-ketovaleric acid (0.874), 3,4-dihydroxybenzoic acid (0.965), 5-aminovaleric acid (0.948), ascorbate (0.917), dehydroascorbic acid (0.896), inosine (0.962), phenylalanine (0.927), and succinic acid (0.976) 2-ketova-leric acid, 3,4-dihydroxybenzoic acid, 5-aminovaleric acid, ascorbate, dehydroascorbic acid, inosine, phenylalanine, and succinic acid (succinate), β-glycerophosphoric acid, maleamate, maleic acid, monoolein, ribose, and trans-4-hydroxy-L-proline TCA cycle, nitrogen metabolism, glutamine and glutamate metabolism, ribose metabolism, and phenylalanine metabolism, alterations in amino acid metabolism, and hypoxia Distinct metabolomic profiles of asthmatics and healthy control patients can be identified in GC-MS based metabolomic profiling of serum Internal cross-validation
McGeachie et al (2015)21 25 Integrated genomic-metabolomic model could predict asthma control (AUC, 95%) monoHETE0863, and sphingosine-1-phosphate, arachidonic acid, PGE2 and S1P Cellular immune response, interferon signaling, and cytokine-related signaling Metabolomic profiling of plasma provides insight into the pathophysiology of asthma control Bootstrapping and cross- validation
Fitzpatrick et al (2014)22 8,953 features Identified 164 Discriminatory metabolites Glycine, serine, and threonine Oxidative stress: the glycine, serine, and threonine metabolism pathway and the N-acylethanolamine, and N-acyltransferase pathway Severe, corticosteroid refractory asthma in children is associated with metabolic derangements No
Jung et al (2013)20 64 PLS-DA model distinguished patients with asthma from healthy adults; training set AUC, 1 (P < .001); validation set: 0.9771 (P < .001). Prediction in validation set: 90.9% for asthma and 100% for control subjects Formate, methanol, acetate, choline, O-phosphocholine, arginine, and glucose Asthma status: hypermethylation, response to hypoxia, and immune reaction; severity: lipid metabolism (1)H-NMR–based metabolite profiling of serum may be useful for the effective diagnosis of asthma and a further understanding of its pathogenesis External validation models (n = 10 drawn from same population)
Comhair et al (2015)23 293 25 compounds were significantly different between cases and control patients, 18 differed by asthma severity levels Taurine, lathosterol, bile acids (taurocholate and glycodeoxycholate), nicotinamide, and adenosine-5-phosphate Asthma status: steroid and amino acid/protein metabolism, inflammatory and immune pathways. Severity: bile acid metabolism and taurine transport The plasma metabolome differs between patients with asthma and healthy control patients and by asthma severity No
Open in a new tab

AUC = area under the curve; FeNO = fractional exhaled nitric oxide; NPV = negative predictive value; NR = not reported; PGE2 = prostaglandin E2; PLS-DA = partial least squares discriminant analysis; PPV = positive predictive value; TCA = tricarboxylic acid; VOC = volatile organic compound. See Table 1 legend for expansion of other abbreviations.

Metabolomic Biomarkers of Asthma

Predictive Indices

The indices of prediction reported by the included studies suggest extremely good classification accuracy (≥ 85%), particularly when differentiating asthma cases from healthy control patients. Discriminatory power was lower for mild vs severe asthma, but would still be ranked as good to excellent (≥ 80%) in most cases. There was no evidence that discriminatory ability differed in adults as opposed to children, with comparable values reported in both groups. The most commonly used model was partial least squares-discriminant analysis (PLS-DA), which is appropriate for analysis of datasets with many correlated predictors, as is common in metabolomics. The R2 and Q2, which are used to assess PLS-DA models, had high values in all studies; however, PLS-DA is known to overestimate predictive ability and only a few studies addressed this.24, 25 Four studies reported area under the receiver operator characteristic (AUC) curves, demonstrating strong predictive ability, with the highest AUC of 0.977 for a 10-biomarker profile.20 The highest AUC for a single metabolite was 0.976 for succinate in serum.19 On the basis of these studies, there was no evidence that including larger numbers of metabolites in the profile increased discriminatory ability.

Regardless of the reported predictive indices, the true test of discriminatory ability is validation. In the search for biomarkers several possible validation strategies can be used, including: (1) testing on a separately recruited and ascertained validation cohort; (2) using a hold-out data set; and (3) performing permutation testing by label shuffling.24, 25 Eight5, 7, 8, 14, 15, 17, 19, 21 and four9, 12, 13, 20 studies used permutation and a hold-out dataset, respectively. Both approaches provided support for accuracy of the reported biomarkers. However, the most robust measure of validation is replication in an entirely independent cohort; this was not performed by any included study. In this review, we highlight the metabolites and metabolomic pathways that are replicated between the included studies to inform the development of comprehensive and effective metabolomic asthma biomarkers.

Outcomes

A large number of common metabolites were associated with asthma case status, severity, exacerbations, and phenotype discrimination, suggesting metabolites contributing to disease onset may also contribute to its severity (Table 3). Even where individual metabolites were not concordant across similar studies, there was consistency in the enriched metabolic pathways (Table 4). Similarly, many were common to both the studies of adults and children. This finding of common metabolomic signatures in children and adults with asthma may support a shared etiology and pathophysiology for these two entities, in contrast to the prevailing belief that childhood asthma is influenced more by genetic predisposition, whereas adult asthma is more affected by environmental factors and obesity.26 In fact, it may be age at asthma onset that is most important in this regard. This was not reported in the studies and, given that 95% of asthma is postulated to start in childhood,27 it can be assumed the included studies are not representative of adult-onset asthma. Further metabolomic studies with rigorously characterized adult-onset asthma are required to determine if and how the metabolomic profiles of such cases differ.

Table 3.

Metabolites Identified as Significant in More Than 1 Study

Group Metabolite Biospecimen Population Method Outcome
Acid salt Formate Plasma,20 EBC13 Adults13, 20 NMR13, 20 AvH13, 20
Hippurate Urine,16 EBC13 Adults13, 16 GC-MS16; NMR13 AvH,13 exacerbation16
Succinate Serum,19 urine15 Children,15 adults19 GC-MS19; NMR15 AvH,15, 19 phenotype15
Alcohol Methanol Plasma,20 EBC13 Adults13, 20 NMR13, 20 AvH13, 20
Amino acid Alanine Urine,16 EBC13 Adults13, 16 GC-MS,16 NMR13 AvH,13 exacerbation16
Arginine Plasma,20 EBC13 Adults13, 20 NMR13, 20 AvH13, 20
Phenylalanine Serum,19 EBC13 Adults13, 19 GC-MS,19 NMR13 AvH13, 19
Threonine Plasma,22 urine15, 16 Children,15, 22 adults16 LC-MS,22 GC-MS,16 NMR15 AvH,15 phenotype,15, 22 exacerbation16
Intermediate in the catabolism of histidine Urocanic acid Urine,14 EBC13 Children,14 adults13 LC-MS,14 NMR13 AvH13, 14
Organic acid Trans-aconitate Urine,15 EBC13 Children,15 adults13 NMR13, 15 AvH,13, 15, phenotype15
Purine nucleoside Adenosine EBC,3, 5, 13 plasma23 Children3, 5; Adults13, 23 MS3, 5, 25; NMR13 AvH3, 13; Phenotype5, 23
Salt Acetate Plasma,20 urine,16 EBC9, 13 Children,9, 20 adults13, 16 LC-MS,20 GC-MS,9, 16 NMR13 AvH,13, 20 AvW,9 exacerbations16
VOC 1,4-dichloro-benzene EBC7, 8 Children7, 8 GC-MS7, 8 AvH,8 WvH7
2,4-dimethyl-1-heptene EBC7, 9 Children7, 9 GC-MS7, 9 WvH,7 AvW9
Open in a new tab

AvH = asthma cases vs healthy control patients; AvW = asthma cases vs wheeze cases; phenotype = measures of asthma phenotypes and severity; WvH = heeze cases vs healthy control patients. See Table 1 legend for expansion of other abbreviations.

Table 4.

Metabolomic Pathways Identified as Significant in More Than 1 Study

Pathway Biospecimen Population Method Outcome
Amino acid metabolism Serum,19 plasma23 Adults19, 23 GC-MS,19 MS23 AvH,19, 23 phenotype23
Glutamate-glutamine cycle; glutamine and glutamate metabolism EBC,11 serum19 Children,11 adults11, 19 NMR,11 GC-MS19 AvH11, 19
Hypoxia response pathways Serum,19 plasma,20 urine15 Children,15 adults19, 20 GC-MS,19 NMR15, 20 AvH,15, 19, 20 phenotype15
Immune pathways Plasma,20, 21, 23 urine14 Children,14, 21 adults20, 23 MS,23 LC-MS,14, 2 NMR20 AvH,14, 20, 23 phenotype,23 asthma control21
Inflammatory pathways EBC,4, 6 plasma23 Children,4, 6 adult23 GC-MS,6 LC-MS,4 MS23 AvH,4, 6, 23 phenotype23
Lipid metabolism Plasma,20 serum,18 EBC,9 urine17 Children,9 adults17, 18, 20 NMR,20 GC-MS9, 17, 18 AvH,18, 20 AvW,9 phenotype17
Oxidative stress EBC,6, 9, 10 plasma,22 urine16 Children,6, 9, 10, 22 adults16 GC-MS,6, 9, 16 LC-MS,22 NMR10 AvH,6, 9, 10 exacerbations,16 phenotype22
Tricarboxylic acid cycle Urine,16 serum,19 urine15 Children,15 adults16, 19 GC-MS,16, 19 NMR15 AvH,15, 19 phenotype15, 16
Open in a new tab

See Table 1 and 3 legends for expansion of abbreviations.

Biological Samples

Unlike the genetic sequence, metabolite profiles can vary depending upon the biomaterial being assessed. Throughout the variety of biospecimens used in these studies, there was considerable consistency in the metabolites and metabolomic pathways identified as significant (Tables 3 and 4); however, larger numbers were not replicated between biospecimens. The lack of replication between studies using the same biospecimen should also be noted. This may be in part attributable to specimen collection conditions and processing procedures, which can affect the metabolome; however, there was no evidence of systematic bias relating to such variables for plasma or urine in these studies.

A number of variables should be considered for EBC, including whole breath vs end-tidal gases, collection device used as well as whether inhaled medications, spirometry, exercise, or other procedures have occurred before sample collection. Motta et al13 investigated the impact of different condensation temperatures on the EBC metabolome (−27.3 and −4.8°C). They reported that although the samples collected at both temperatures resulted in metabolomics profiles that could distinguish asthma cases from control patients, the constituent metabolites of the profiles varied. Their work and that of others highlights that susceptibility to such external influences tends to be metabolite dependent. Further, it underlines the importance of standardizing collection and metabolite assays for biospecimens—a goal that has yet to be achieved in this field. Last, none of the studies profiled metabolites in more than one biospecimen type; therefore, it is not possible to determine the relationship between metabolites across biospecimens from the same individual, nor whether discriminatory metabolites could be detected in different biospecimens from the same subject.

Metabolomic Profiling

Metabolomic profiling technique may also account for differences in study findings. NMR uses the magnetic properties of atomic nuclei to generate information on structure and thereby identify metabolites in the biofluid under investigation by their unique pattern of chemical shifts and peak intensities.28 Liquid or gas chromatography tandem mass spectrometry combines chromatography, a technique that separates metabolites, with MS, which measures their abundance. The complement of metabolites measured by NMR and MS may not always be comparable. Nevertheless, multiple metabolites and metabolomic pathways were defined as significant under both methods (Tables 3 and 4). Others were only identified as significant in the studies using NMR profiling such as arginine, formate, methanol,13, 20 and trans-aconitate.13, 15 However whether this is a function of the profiling method or other sources of heterogeneity between the studies cannot be discerned.

One fundamental source of heterogeneity is the use of a global untargeted metabolomic profiling approach, which aims to capture all metabolites in a biological system as opposed to a hypothesis-driven approach targeting specific metabolites, or metabolite classes. Fifteen of the included studies were targeted or semitargeted in nature: three focused on volatile organic compounds (VOCs)7, 9, 13; three on a combination of alkanes, alkenes, aldehydes, and ketones6, 16, 17; one on leukotrienes4; and the rest on a variety of metabolites from specific panels. A targeted approach allows for optimal sensitivity in the measurement of these metabolites because it uses tailored and calibrated methods; however, it lacks the broad range of an untargeted approach and may miss novel but important metabolites without an a priori biological hypothesis. Crucially, it also hampers the replication and validation of findings between studies.

Other Variables

Replication may also have been affected by heterogeneity in the diagnostic criteria for asthma, with variable use of physician diagnosis, spirometric criteria, and/or subject self-report. Even where a definitive diagnosis can be made, asthma reflects a broad spectrum of disorders of varying severity. The concordance between the studies across a range of asthma outcomes suggests a similar underlying pathogenesis, however.

The metabolome is known to fluctuate with factors such as BMI and to be highly sensitive to external influences including diet, smoking status, and treatment regime. Three7, 12, 20 studies withheld treatment for a set period before sample collection. Others reported on medication usage, some used treatment as a measure of asthma severity, and five3, 8, 9, 11, 19 did not report on treatment at all. Currently, the data are too limited to draw conclusions regarding the effect of treatment on the asthma metabolome. Similarly, the data on potential confounders are not yet comprehensive enough for analysis, but will benefit from the efforts of the wider metabolomics community to identify the metabolomic shifts induced by various environmental factors and physiological characteristics, and the “healthy” metabolome.

Biological Insights Into Asthma Pathophysiology and Treatment

An expanding list of human metabolites has been annotated and comprehensively mapped to specific biological pathways. In the reviewed studies, a large number of pathways were reported to be associated with asthma outcomes in a variety of biospecimens (Table 2). Although diverse, these pathways can be broadly categorized on the basis of general physiological or molecular roles: (1) immune response, signaling, and inflammation; (2) metabolism of amino acids, sugars, bile acids, steroids, and lipids; (3) oxidative stress and hypoxia; (4) cellular energy homeostasis; and (5) DNA hypermethylation. Among studies investigating asthma status, all were represented. In analyses of asthma phenotypes, immune responses, oxidative stress, energy metabolism, and metabolism of amino acids and lipids were enriched, whereas asthma control21 was associated primarily with immune response pathways.

Aberrant immune responses and acute inflammation are hallmark features of all asthmatic phenotypes, and the predominance of inflammatory and immunological response pathways is not surprising. An enrichment of pathways reflecting the increased metabolism of amino acids, lipids, steroids, and bile acids that are fundamental to asthma pathogenesis is also anticipated. Amino acids are mediators of immunological activities in asthma and have antioxidant functions; in particular, taurine, glycine, glutamine, and glutamate may have potentially protective effects, whereas phenylalanine may have adverse effects. Lipid mediators are key drivers of inflammatory responses in asthma and have well-characterized roles in T-cell recruitment and energy metabolism; therefore, the enrichment of lipid metabolism pathways in asthma metabolomic studies is consistent with the biological importance of these molecules in asthma pathogenesis. The role of oxidative stress in asthma has also been well-studied, and evidence suggests that an imbalance between oxidation and reducing systems, in the favor of oxidative states, contributes to asthma severity. Both endogenous and exogenous reactive oxygen species including superoxide and reactive nitrogen and hydrogen species, increase airway inflammation, and are key determinants of asthma severity. Activated inflammatory cells in the airway produce reactive oxygen species that contribute to poor asthma control by reducing the ability of the airway epithelium to repair damage resulting from oxidative stress.

Pathways related to hypoxia were also significantly enriched. Increased hypoxic responses by the inflamed airway have been observed in asthma and were reported to lead to exacerbations in acute and chronic experimental allergic models of asthma, but not in healthy, noninflamed lung tissue.29 The increase in oxidative and hypoxic stress responses in asthma coincides with considerable alterations in cellular energy metabolism. Levels of metabolites participating in the tricarboxylic acid cycle were altered in asthma, and fluctuations of metabolites in pathways involved in cellular energy metabolism in the lungs have been observed in mouse models of experimental asthma.30 Potentially, alterations in these pathways may reflect the reduced ability of the damaged lung to meet the substantial energy demands of activated inflammatory cells in the allergic airway. Finally, epigenetic effects have a strong impact on asthma severity, and metabolites related to the methyl transfer pathway were also reported.13, 20 DNA methylation may increase airway inflammation by predisposing immune responses towards a Th2 phenotype; increased hypermethylation may therefore represent a novel epigenetic mechanism underlying asthma pathogenesis.20

Discussion: The Future of Asthma Omics

Asthma metabolomics studies to date are limited but encouraging and report a number of replicated biologically plausible metabolites and metabolomic pathways associated with the development and manifestation of asthma. Whether the nonreplicated results represent spurious findings or heterogeneity between the studies cannot be assessed with the literature available to date. Much of this heterogeneity stems from lack of standardization in the field, and highlights the need for the development of a rigorous set of criteria for conducting and reporting metabolomic studies.

If clinical translation is the end goal, several factors must be considered. First, the determination of specificity: the biomarkers must be specific to the asthma phenotype rather than representing a general profile of a biological system in a dysregulated physiological state. In these studies, the VOC profile of wheeze7 was similar to many of the asthma profiles. This is perhaps not unexpected; however, Esther et al3 also reported similarities with cystic fibrosis profiles. In the wider literature, more distinct respiratory disorders such as ARDS as well as exposure to environmental pollutants that may affect lung function31 were also associated with a number of the metabolites identified in this review. Perhaps most importantly, many “asthma metabolites,” particularly the amino acids and those involved in choline metabolism, have been associated with other chronic diseases including multiple malignancies.32 Although this does not negate their possible involvement in the pathogenesis of asthma, it does call into question their utility as stand-alone biomarkers.

A further question involves the role of the biomarkers. Most studies focused on distinguishing asthma cases from healthy control patients; however, established clinical markers and criteria for the diagnosis of asthma already exist. A more useful role for metabolomic biomarkers may be in the discrimination of different subtypes, which are currently not well-defined. Prediction is arguably of the greatest clinical use. No studies used a prospective design to identify predictive biomarkers, although one focused on wheezing in preschoolers, which could be considered an early asthma phenotype. In terms of the most optimal biospecimen, EBC is an attractive, noninvasive method approach for collecting samples with more direct relevance to the end organ of interest. However, in the included studies, there was no evidence that EBC-based biomarkers outperformed blood or urine.

Clearly, further refinement of biomarkers is required before clinical translation is a viable option. Metabolomic data alone may be insufficient to fully characterize complex pathologies.33 The integration of metabolomics with other omic data to identify the interactions and synergisms between the different hierarchical components of the “central biological dogma” represents a potential strategy that will allow the visualization of a biological system on a global level.

Ried et al18 integrated metabolomic profiles with asthma-associated single nucleotide polymorphisms (SNPs) and observed that several SNPs at the asthma susceptibility locus 17q21 influenced asthma-associated metabolites, particularly phosphatidylcholines, and concluded that the simultaneous analysis of metabolite and genetic data provide an improved understanding of diseases mechanisms on a molecular and functional level. McGeachie et al21 expanded on this approach by additionally incorporating gene expression and methylation data into their analysis. This led to both an increased understanding of physiology and an increased predictive accuracy, relative to the use of a single omic technology, again supporting the integration of multiple data types.

To date, no studies have integrated metabolomics and proteomic data, although this may in fact be the most informative integrative strategy. A single gene can generate multiple different proteins through alternative splicing, and posttranslational modifications and proteolytic processes. These proteins form the main structural components of all cells and control the majority of their biological functions.34 One crucial role is as enzymes to catalyze metabolic and signaling pathways; however, it can be difficult to ascertain the endogenous physiological function of these different enzymes because they often exist as part of large networks and are regulated by posttranslational events. Metabolomic profiling of the substrates involved in these reactions can help assign biochemical functions to these enzymes providing access to “a portion of biomolecular space that is inaccessible to genomics and proteomics”35 and have the potential to identify functionally relevant biological targets.

The field of proteomics has developed almost in parallel with metabolomics, although the terminology was coined slightly earlier in 1994.36 Analogous to metabolomics, some of the most commonly used technologies for protein separation and identification include liquid chromatography and mass spectrometry.37 The proteome is also similarly dynamic and sensitive to exogenous exposures and intracellular stimuli.38 Together, these omes provide a downstream read out of the genome and its direct interaction with the environment.

As with metabolomics, proteomics studies remain limited in respiratory medicine, specifically asthma.34 Studies are hindered by many of the same issues, including sample size, lack of standards for sample collection, handling and storage, multiple incompatible profiling technologies, and underdeveloped analytical methods.38 Similarly, quantification and identification of proteins is challenging; as with the metabolome, the entire proteome is yet to be characterized. Unidentified proteins may account for up to 60% of the total proteomic database and it is unclear exactly how large it is.37, 39 Mapping of the proteome is more advanced than the metabolome, however, with drafts of the human and lung proteomes available.34 Crucially, as with metabolomics, the majority of reported proteomic findings have yet to be validated.40

One notable difference between metabolomics and proteomics is in the biospecimens used. EBC, which has a low protein content, forms only a minority of the literature, whereas sputum, lung epithelial lining, or BAL fluid, which more directly reflect the lungs activity, are much more commonly used.39 The abundance of proteins in plasma and serum can be both an advantage and a disadvantage, particularly for the measurement of the less abundant, lower molecular weight proteins.41 The choice of biospecimen may therefore affect the findings34 and is an important consideration for integrated omics moving forward.

The proteomics asthma literature to date has been summarized in several comprehensive reviews.38, 40 In plasma proteins involved in iron metabolism, coagulation cascade, acute-phase response, responses to stress and pathogens, and in complement cascades have all been reported. Complement cascades were also identified among the sputum and BAL fluid literature, together with signaling; calcium-binding and lung remodeling proteins; proteins involved in cellular movement, immune cell trafficking, collagen fibrillogenesis and chemotaxis; cytokines; chemokines; matrix metalloproteinases; signaling; and, crucially for integration, metabolic enzymes. Yet no proteomic biomarkers with clinical applications in asthma have emerged thus far.38 These findings broadly support those of the existing metabolomics studies, which is, in particular, an important role for the immune system and for identifying novel potentially important processes. Going forward, however, it may be anticipated that the exact points(s) of dysregulation, within a pathway, and therefore more clinically relevant information, can be identified by actually combining these two data sets.

Conclusion

Omics technologies remain in the early stages. Although increasingly promising results are being reported, metabolomics and proteomics in particular are limited by a lack of standards in the field and uncertainty in the optimal analytical methods. Additionally it is becoming increasingly clear that integrated omic analyses are necessary to maximally leverage these data. As more large population-based studies begin to generate multiomic data, it is likely to represent the newest frontier in asthma research. However, clinical utility has yet to be demonstrated, and whether the future of asthma metabolomics and integrative omics lie in the development of biomarkers, or whether it is better suited to increasing the understanding of its underlying biology remains to be determined.

Acknowledgments

Author contributions: R. S. K. takes responsibility for the content of the manuscript. J. L.-S. conceived of the original article; R. S. K. performed the literature search, which was validated by all authors; J. L.-S., R. S. K., A. D., M. J. M., W. Q., J. S., E. S. W., and A. C. W. contributed significantly to the writing of the manuscript.

Financial/nonfinancial disclosures: J. L.-S. is a consultant for Metabolon, Inc. None declared (R. S. K., A. D., M. J. M., W. Q., J. S., E. S. W., A. C. W.).

Role of sponsors: The sponsors had no role in the design of the study, the collection and analysis of the data, or the preparation of the manuscript.

Other contributions: We thank all members of the Integrative Metabolite-Phenotype Analyses of Complex Traits Consortium.

Footnotes

FUNDING/SUPPORT: Drs Lasky-Su and Kelly are supported by National Institutes of Health (NIH) grant R01HL123915-01; Dr Dahlin is supported by NIH grant 1K01HL130629-01A1; Dr McGeachie is supported by a grant from the Parker B Francis Foundation; and Dr Wan is supported by a Department of Veterans Affairs Rehabilitation Research and Development Award 1K2RX002165.

References

  • 1.Ober C., Yao T.-C. The genetics of asthma and allergic disease: a 21(st) century perspective. Immunol Rev. 2011;242(1):10–30. doi: 10.1111/j.1600-065X.2011.01029.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2.Johnson C.H., Ivanisevic J., Siuzdak G. Metabolomics: beyond biomarkers and towards mechanisms. Nat Rev Mol Cell Biol. 2016;17(7):451–459. doi: 10.1038/nrm.2016.25. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3.Esther C.R., Boysen G., Olsen B.M. Mass spectrometric analysis of biomarkers and dilution markers in exhaled breath condensate reveals elevated purines in asthma and cystic fibrosis. Am J Physiol Lung Cell Mol Physiol. 2009;296(6):L987–L993. doi: 10.1152/ajplung.90512.2008. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.Montuschi P. LC/MS/MS analysis of leukotriene B4 and other eicosanoids in exhaled breath condensate for assessing lung inflammation. J Chromatogr B Analyt Technol Biomed Life Sci. 2009;877(13):1272–1280. doi: 10.1016/j.jchromb.2009.01.036. [DOI] [PubMed] [Google Scholar]
  • 5.Carraro S., Giordano G., Reniero F. Asthma severity in childhood and metabolomic profiling of breath condensate. Allergy. 2012;68(1):110–117. doi: 10.1111/all.12063. [DOI] [PubMed] [Google Scholar]
  • 6.Caldeira M., Perestrelo R., Barros A.S. Allergic asthma exhaled breath metabolome: a challenge for comprehensive two-dimensional gas chromatography. Jf Chromatogr A. 2012;1254:87–97. doi: 10.1016/j.chroma.2012.07.023. [DOI] [PubMed] [Google Scholar]
  • 7.van de Kant K.D.G., van Berkel J.J.B.N., Jöbsis Q. Exhaled breath profiling in diagnosing wheezy preschool children. Eur Resp J. 2013;41(1):183–188. doi: 10.1183/09031936.00122411. [DOI] [PubMed] [Google Scholar]
  • 8.Gahleitner F., Guallar-Hoyas C., Beardsmore C.S., Pandya H.C., Thomas C.P. Metabolomics pilot study to identify volatile organic compound markers of childhood asthma in exhaled breath. Bioanalysis. 2013;5(18):2239–2247. doi: 10.4155/bio.13.184. [DOI] [PubMed] [Google Scholar]
  • 9.Smolinska A., Klaassen E.M.M., Dallinga J.W. Profiling of volatile organic compounds in exhaled breath as a strategy to find early predictive signatures of asthma in children. PLoS ONE. 2014;9(4):e95668. doi: 10.1371/journal.pone.0095668. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Carraro S., Rezzi S., Reniero F. Metabolomics applied to exhaled breath condensate in childhood asthma. Am J Respir Crit Care Med. 2007;175(10):986–990. doi: 10.1164/rccm.200606-769OC. [DOI] [PubMed] [Google Scholar]
  • 11.Sinha A., Krishnan V., Sethi T. Metabolomic signatures in nuclear magnetic resonance spectra of exhaled breath condensate identify asthma. Eur Respir J. 2012;39(2):500–502. doi: 10.1183/09031936.00047711. [DOI] [PubMed] [Google Scholar]
  • 12.Ibrahim B., Marsden P., Smith J.A., Custovic A., Nilsson M., Fowler S.J. Breath metabolomic profiling by nuclear magnetic resonance spectroscopy in asthma. Allergy. 2013;68(8):1050–1056. doi: 10.1111/all.12211. [DOI] [PubMed] [Google Scholar]
  • 13.Motta A., Paris D., D’Amato M. NMR metabolomic analysis of exhaled breath condensate of asthmatic patients at two different temperatures. J Proteome Res. 2014;13(12):6107–6120. doi: 10.1021/pr5010407. [DOI] [PubMed] [Google Scholar]
  • 14.Mattarucchi E., Baraldi E., Guillou C. Metabolomics applied to urine samples in childhood asthma; differentiation between asthma phenotypes and identification of relevant metabolites. Biomed Chromatogr. 2012;26(1):89–94. doi: 10.1002/bmc.1631. [DOI] [PubMed] [Google Scholar]
  • 15.Saude E.J., Skappak C.D., Regush S. Metabolomic profiling of asthma: diagnostic utility of urine nuclear magnetic resonance spectroscopy. J Allergy Clin Immunol. 2011;127(3):757–764.e1-6. doi: 10.1016/j.jaci.2010.12.1077. [DOI] [PubMed] [Google Scholar]
  • 16.Loureiro C.C., Duarte I.F., Gomes J. Urinary metabolomic changes as a predictive biomarker of asthma exacerbation. J Allergy Clin Immunol. 2014;133(1):261–263.e265. doi: 10.1016/j.jaci.2013.11.004. [DOI] [PubMed] [Google Scholar]
  • 17.Loureiro C.C., Oliveira A.S., Santos M. Urinary metabolomic profiling of asthmatics can be related to clinical characteristics. Allergy. 2016;71(9):1362–1365. doi: 10.1111/all.12935. [DOI] [PubMed] [Google Scholar]
  • 18.Ried J.S., Baurecht H., Stückler F. Integrative genetic and metabolite profiling analysis suggests altered phosphatidylcholine metabolism in asthma. Allergy. 2013;68(5):629–636. doi: 10.1111/all.12110. [DOI] [PubMed] [Google Scholar]
  • 19.Chang C., Guo Z-g, He B., Yao W-z. Metabolic alterations in the sera of Chinese patients with mild persistent asthma: a GC-MS-based metabolomics analysis. Acta Pharmacol Sinica. 2015;36(11):1356–1366. doi: 10.1038/aps.2015.102. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Jung J., Kim S.H., Lee H.S. Serum metabolomics reveals pathways and biomarkers associated with asthma pathogenesis. Clin Exp Allergy. 2013;43(4):425–433. doi: 10.1111/cea.12089. [DOI] [PubMed] [Google Scholar]
  • 21.McGeachie M.J., Dahlin A., Qiu W. The metabolomics of asthma control: a promising link between genetics and disease. Immun Inflamm Dis. 2015;3(3):224–238. doi: 10.1002/iid3.61. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22.Fitzpatrick A.M., Park Y., Brown L.A., Jones D.P. Children with severe asthma have unique oxidative stress-associated metabolomic profiles. J Allergy Clin Immunol. 2014;133(1):258–261.e258. doi: 10.1016/j.jaci.2013.10.012. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Comhair S.A.A., McDunn J., Bennett C., Fettig J., Erzurum S.C., Kalhan S.C. Metabolomic endotype of asthma. J Immunol. 2015;195(2):643–650. doi: 10.4049/jimmunol.1500736. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Westerhuis J.A., Hoefsloot H.C.J., Smit S. Assessment of PLSDA cross validation. Metabolomics. 2008;4(1):81–89. [Google Scholar]
  • 25.Esbensen K.H., Geladi P. Principles of proper validation: use and abuse of re-sampling for validation. J Chemometrics. 2010;24(3-4):168–187. [Google Scholar]
  • 26.de Nijs S.B., Venekamp L.N., Bel E.H. Adult-onset asthma: is it really different? Eur Resp Rev. 2013;22(127):44–52. doi: 10.1183/09059180.00007112. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27.Masoli M., Fabian D., Holt S., Beasley R. The global burden of asthma: executive summary of the GINA Dissemination Committee report. Allergy. 2004;59(5):469–478. doi: 10.1111/j.1398-9995.2004.00526.x. [DOI] [PubMed] [Google Scholar]
  • 28.Smith I.C.P., Blandford D.E. Nuclear magnetic resonance spectroscopy. Anal Chem. 1995;67(12):509–518. doi: 10.1021/ac00108a037. [DOI] [PubMed] [Google Scholar]
  • 29.Ahmad T., Kumar M., Mabalirajan U. Hypoxia response in asthma. Am J Resp Cell Mol Biol. 2012;47(1):1–10. doi: 10.1165/rcmb.2011-0203OC. [DOI] [PubMed] [Google Scholar]
  • 30.Ho W.E., Xu Y.-J., Xu F. Metabolomics reveals altered metabolic pathways in experimental asthma. Am J Resp Cell Mol Biol. 2013;48(2):204–211. doi: 10.1165/rcmb.2012-0246OC. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31.Stringer K.A., McKay R.T., Karnovsky A., Quémerais B., Lacy P. Metabolomics and its application to acute lung diseases. Front Immunol. 2016;7:44. doi: 10.3389/fimmu.2016.00044. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32.Kelly R.S., Vander Heiden M.G., Giovannucci E., Mucci L.A. Metabolomic biomarkers of prostate cancer: prediction, diagnosis, progression, prognosis and recurrence. Cancer Epidemiol Biomarkers Prev. 2016;25(6):887–906. doi: 10.1158/1055-9965.EPI-15-1223. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33.Wanichthanarak K., Fahrmann J.F., Grapov D. Genomic, proteomic, and metabolomic data integration strategies. Biomarker Insights. 2015;10(Suppl 4):1–6. doi: 10.4137/BMI.S29511. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 34.Priyadharshini V.S., Teran L.M. Personalized medicine in respiratory disease: role of proteomics. In: Rossen D., editor. Vol. 102. Academic Press; Cambridge, MA: 2016. pp. 115–146. (Advances in Protein Chemistry and Structural Biology). [DOI] [PubMed] [Google Scholar]
  • 35.Saghatelian A., Cravatt B.F. Global strategies to integrate the proteome and metabolome. Curr Opin Chem Biol. 2005;9(1):62–68. doi: 10.1016/j.cbpa.2004.12.004. [DOI] [PubMed] [Google Scholar]
  • 36.Crameri R. The potential of proteomics and peptidomics for allergy and asthma research. Allergy. 2005;60(10):1227–1237. doi: 10.1111/j.1398-9995.2005.00873.x. [DOI] [PubMed] [Google Scholar]
  • 37.Dompeling E., Jöbsis Q. Proteomics of exhaled breath condensate: a realistic approach in children with asthma? Clin Exp Allergy. 2011;41(3):299–301. doi: 10.1111/j.1365-2222.2010.03686.x. [DOI] [PubMed] [Google Scholar]
  • 38.Teran L.M., Montes-Vizuet R., Li X., Franz T. Respiratory proteomics: from descriptive studies to personalized medicine. J Proteome Res. 2015;14(1):38–50. doi: 10.1021/pr500935s. [DOI] [PubMed] [Google Scholar]
  • 39.O'Neil S.E., Lundbäck B., Lötvall J. Proteomics in asthma and COPD phenotypes and endotypes for biomarker discovery and improved understanding of disease entities. J Proteomics. 2011;75(1):192–201. doi: 10.1016/j.jprot.2011.10.008. [DOI] [PubMed] [Google Scholar]
  • 40.Terracciano R., Pelaia G., Preianò M., Savino R. Asthma and COPD proteomics: Current approaches and future directions. Proteomics. 2015;9(1-2):203–220. doi: 10.1002/prca.201400099. [DOI] [PubMed] [Google Scholar]
  • 41.Pereira-Fantini P.M., Tingay D.G. The proteomics of lung injury in childhood: challenges and opportunities. Clin Proteomics. 2016;13(1):1–8. doi: 10.1186/s12014-016-9106-0. [DOI] [PMC free article] [PubMed] [Google Scholar]

Articles from Chest are provided here courtesy of American College of Chest Physicians

RESOURCES