Metabolomics in asthma: A platform for discovery

Abstract

Asthma, characterized by airway hyperresponsiveness, inflammation and remodeling, is a chronic airway disease with complex etiology. Severe asthma is characterized by frequent exacerbations and poor therapeutic response to conventional asthma therapy. A clear understanding of cellular and molecular mechanisms of asthma is critical for the discovery of novel targets for optimal therapeutic control of asthma. Metabolomics is emerging as a powerful tool to elucidate novel disease mechanisms in a variety of diseases. In this review, we summarize the current status of knowledge in asthma metabolomics at systemic and cellular levels. The findings demonstrate that various metabolic pathways, related to energy metabolism, macromolecular biosynthesis and redox signaling, are differentially modulated in asthma. Airway smooth muscle cell plays pivotal roles in asthma by contributing to airway hyperreactivity, inflammatory mediator release and remodeling. We posit that metabolomic profiling of airway structural cells, including airway smooth muscle cells, will shed light on molecular mechanisms of asthma and airway hyperresponsiveness and help identify novel therapeutic targets.

1. Introduction

Metabolome identifies the global collection of small molecules generated from cellular metabolic processes (Sun and Hu, 2016). Technically, metabolomics and metabonomics are used interchangeably. Metabolomics analyzes and defines the entire metabolome in organisms and multicellular systems, while metabonomics describes the metabolic changes in the system often in response to an evoked phenotype or interventions (Nicholson and Lindon, 2008). Compared to other “omics” approaches measuring global cellular changes, metabolomics characterizes the dynamic changes of small molecules generated from a multitude of complex intracellular processes (Fiehn et al.,2000). In fact, the metabolomic profile of a cell represents the downstream output of the transcriptomic and proteomic changes and the upstream input to the phenotype. Therefore, metabolomic profile encompasses information captured in transcriptomic and proteomic profiles of the cell.

2. Targeted vs untargeted metabolomics

Based on analytical methods used, metabolomics can comprise untargeted and targeted metabolomics (Fig. 1). When it is driven by a specific hypothesis, metabolomic screening is conducted in a targeted manner for selected metabolites (Dudley et al., 2010). The targeted approach optimizes sample preparation and accurately identifies associations among pre-selected metabolites and a defined physiological context (Roberts et al., 2012). Toxicological studies in animal models and pharmacokinetic investigations in drug discovery benefit mostly from targeted metabolomic approaches (Kantae et al., 2017; Ramirez et al., 2013).

Fig. 1. Workflow of metabolomics-based asthma study.

A typical metabolomic workflow in asthma involves either targeted metabolomics to test a hypothesis or untargeted metabolomics approach to generate testable hypotheses. In targeted approach, quantitative comparison of pre-selected metabolites leads to identifying profiles related to disease states or treatment conditions. In the untargeted approach, disease or treatment-associated profiles can still be elucidated, however, bench validation is critical for testing hypotheses. The end result of both approaches is to identify disease-selective diagnostic biomarkers and understanding novel mechanisms of disease, which potentially lead to novel therapeutic targets. NMR – Nuclear Magnetic Resonance, LC-MS – Liquid Chromatography – Mass Spectrometry, HPLC-MS – High-Performance Liquid Chromatography – Mass Spectrometry, KEGG – Kyoto Encyclopedia of Genes and Genomes, HMDB – Human Metabolome Database, MassBank – MassBank Europe Spectral Database, OpenMS – Open-source software for mass spectrometry.

In untargeted metabolomics, the analysis is unbiased and screens all metabolites in the system in an exploratory manner to generate hypotheses. Untargeted metabolomic screening comprehensively reflects the metabolic profiles of the organism, tissue, or the selected cell types and facilitates identification of novel biomarkers and therapeutic targets (Patti et al., 2012).

3. Technical approaches in metabolomics

Currently, nuclear magnetic resonance (NMR) and mass spectrometry (MS) represent widely used analytical platforms in metabolomics. Generally, NMR is considered a non-destructive method with high reproducibility and capacity to identify chemical structures. In recent years, however, MS-based metabolomics has gained momentum over NMR due to improved sensitivity and wide availability (Dunn et al., 2005; Markley et al., 2017; Zhang et al., 2012). The MS platforms have advanced with recently-introduced cutting-edge technologies, such as BAYESIL system, matrix assisted laser desorption/ionization (MALDI), and desorption/ionization on silicon (DIOS) system (Greving et al., 2011; Ravanbakhsh et al., 2015; Warren G. Lewisa, 2003; Wittmann and Heinzle, 2001). The critical component of unbiased metabolomics is the data analysis with bioinformatic interfaces to derive meaningful biological inferences. The Kyoto Encyclopedia of Genes and Genomes (KEGG), Human Metabolome Database (HMDB), and Open MS are the prominent databases currently in use (Du et al., 2013; Fessenden, 2016).

In summary, metabolomics provides an integrative platform to understand the pathophysiologic basis of a variety of diseases (Cirulli et al., 2019; Guasch-Ferre et al., 2016). Studies suggest that asthma, especially the obesity-associated asthma, is driven by altered metabolic status (Cottrell et al., 2011; Singh et al., 2013). In this review, we will focus on the application of metabolomics in asthma studies, with an emphasis on airway structural cells.

4. Metabolomics studies in asthma

Asthma is a heterogeneous chronic airways disease, with a complex etiology (Diette et al., 2008; Ober and Hoffjan, 2006). Over the last three decades, “omics” approaches encompassing genomics, transcriptomics, epigenomics, and proteomics, have immensely helped understand the underlying disease mechanisms of asthma (Kan et al., 2017). However, the complex etiology and heterogeneity of asthma can also benefit from the use of novel tools such as metabolomics.

Recent research links metabolomic profiles to asthma by focusing on metabolites as systemic biomarkers (Kelly et al., 2017a; Licari et al., 2018; Pite et al., 2018; Snowden et al., 2012). We will review studies that have focused on asthma-derived biospecimens and metabolites associated with disease onset and severity.

4.1. Urine-based metabolomics

Because of its non-invasive and convenient sampling methods, urine is a widely used body fluid in metabolomic studies aimed to identify biomarkers of physiological status (Bouatra et al., 2013). A variety of metabolites in urine reportedly differentiate between asthma and healthy participants (Loureiro et al., 2016; Quan-Jun et al., 2017; Saude et al., 2011). According to urine-based metabolomic profiles, Citric Acid Cycle (Tricarboxylic Acid cycle, TCA cycle) appears to be a predominant pathway altered in asthma patients. TCA cycle is the terminal common pathway for oxidative catabolism of the three macronutrients, carbohydrates, lipids, and proteins. As the central metabolic hub in eukaryotic cells, the TCA cycle provides substrates for ATP generation and facilitates biosynthesis of non-essential fatty acids and amino acids (Jeremy, Berg, 2002; Victor W. Rodwell et al., 2018). In children with asthma, urinary metabolomic profiles, dominated by TCA metabolites, such as 2-oxoglutarate, succinate, threonine, and cis-aconitate, differentiated participants with stable and unstable asthma (Saude et al., 2011). These data suggest a perturbed TCA cycle is associated with childhood asthma and could serve as potential urinary biomarkers to predict poorly controlled pediatric asthma. However, potential confounding effects of asthma medications, such as glucocorticoids, should be carefully considered. Oxidative stress also represents a hallmark of airway inflammatory disorders, such as asthma and COPD (Kirkham and Rahman, 2006; MacNee, 2001). Mitochondrial electron transport chain (ETC) consists of a series of complexes transferring electrons at the inner membrane. Electrons leaked during ETC react with oxygen molecules to generate reactive oxygen species (ROS), causing oxidative stress in the cells (Pan et al., 2019). The lipids in membranes and biological fluids provide prime targets of oxidative damage induced by the ROS. Therefore, lipid peroxidation products are often measured to assess oxidative stress in vivo (Yoshida et al., 2013). In the asthma patients, lipid peroxidation products in urine, such as 2-methylpentane and octane, correlated with clinical parameters of asthma severity, including Forced Expiratory Volume in 1 s (FEV₁) and eosinophilia (Loureiro et al., 2016). Eicosanoids-comprised of prostaglandins, cysteinyl leukotrienes and isoprostanes-are lipid mediators synthesized and rapidly metabolized in response to inflammatory stimuli and therefore studied in asthma as diagnostic and treatment response biomarkers (Sokolowska et al., 2021; von Moltke et al., 2012). For instance, elevated urinary leukotriene E4 (LTE4) and prominent PGD2 metabolites predicted severe and Th2-biased asthma in pediatric and adult patients. (Kolmert et al., 2021). On the other hand, urinary metabolomics can also be useful in measuring metabolic outcomes of asthma therapy. The conventional asthma treatment includes β₂ adrenergic receptor (β₂AR) agonists as bronchodilators and glucocorticoids as anti-inflammatory agents. Children with asthma treated with β₂AR and glucocorticoids manifested urinary markers of altered TCA cycle and pyruvate metabolism, characterized by decreased citrate and increased lactate levels (Quan-Jun et al., 2017).

Although urine-based metabolomics plays an important role in characterizing both pulmonary and systemic metabolism, the quality and accuracy of urinary metabolomic screening can be compromised by multiple factors. Firstly, it is difficult to precisely control the dilution of urine samples, influenced by an array of factors including hydration status, diuretics use, collection time, and renal pathology. For instance, a study found almost a quarter of screened metabolites in urine showed a rhythmic variation during 24 h cycle (Giskeodegard et al., 2015). However, semi-quantitative strategies, such as normalization to total ion current, osmolality, and common components like creatinine, help optimize urinary metabolomic analysis (Miller et al., 2019; Warrack et al., 2009). Additionally, the pH values and ionic species variations in urine may chemically modify some urinary metabolites, and confound the original metabolomic profiles (Vincent M. Asiago et al., 2008). These variations can be minimized by introduction of buffering systems in sample collection and processing (Vincent M. Asiago et al., 2008; Xiao et al., 2009).

4.2. Serum/plasma metabolomics

Currently, serum and plasma are the predominant biological fluids examined in metabolomic studies. Blood-based metabolic profiling has identified metabolomic endotypes of asthma, including increased synthesis of bile acids and altered steroid and amino acid metabolism (Comhair et al., 2015). Amino acids and fatty acids are macromolecules critical for a variety of cellular functions, such as biosynthesis, inflammation, and homeostasis. Plasma or serum-based metabolomic studies in human and animal models have illustrated the association of amino acids, nucleic acid metabolites, and lipid derivatives with asthma and identified metabotypes of asthma. The serum concentrations of lipid mediators and hormones, such as oleoylethanolamide, dehydroepiandrosterone sulfate, and cortisone, corelate with asthma severity (Reinke et al., 2017). Plasma metabolomic profiles in ovalbumin (OVA)-induced allergic AHR mouse model showed dysregulated metabolisms of purine, glycerophospholipids, and tryptophan (Yu et al., 2016). In mammals, purine metabolism generates uric acid, which at its homeostatic levels scavenges oxidative free radicals and provides antioxidant defense (Peden et al., 1990). As an indicator of oxidative stress, serum uric acid can be used as a biomarker for asthma severity (Abdulnaby et al., 2016). Elevated serum uric acid levels are also detected during acute asthma exacerbation (Li et al., 2014). Arginine is another amino acid linked to antioxidant pathways through its metabolism to nitric oxide (NO). NO functions as an antioxidant to limit oxidative injuries in the cells that is associated with both broncho-dilatory and anti-inflammatory effects (Prado et al., 2011; Ricciardolo, 2003; Wink et al., 2001). Studies in human subjects have identified metabolomic signatures of asthma linked to NO. Disrupted arginine synthesis in asthma results in reduced arginine bioavailability and NO deficiency (Jung et al., 2013; Lara et al., 2008; Xu et al., 2017). The Fraction of exhaled NO (FeNO) is a biomarker for eosinophilic airway inflammation. Studies showed that serum levels of metabolites taurine and nicotinamide are elevated in severe asthma patients and correlate with high FeNO levels (Comhair et al., 2015). Interestingly, plasma metabolic profile is also useful to distinguish other allergic diseases, such as food allergy. Individuals with asthma showed decreased serum fatty acid levels, a finding similar to that in children with food allergy (Crestani et al., 2020). On the other hand, some biomarkers help differentiate between asthma and other allergic disorders. For instance, sphingomyelins are reported to be decreased in food allergy while increased in asthma (Crestani et al., 2020).

Nevertheless, various pre-analytical factors should be addressed in serum or plasma metabolomics. Blood coagulation is an important factor to consider in sample selection between serum and plasma. Collection tubes and collection time may modify the metabolic profiles, including the abundance of arachidonic acid, arginine, sarcosine, and kynurenine (Cruickshank-Quinn et al., 2018; Paglia et al., 2018). Additional confounding factors such as temperature and centrifugation may also contribute to unique metabolic patterns, due to metabolites released from platelets or enzymes activated in thrombosis (Liu et al., 2018b). Lastly, sample transportation and storage conditions should also be carefully controlled for optimal reproducibility (Gonzalez-Dominguez et al., 2020; Torell et al., 2017).

4.3. Exhaled breath condensate (EBC) metabolomics

Exhaled breath condensate (EBC) is a biospecimen relevant to airways diseases (Horvath et al., 2005). A variety of biomarkers, including hydrogen peroxide, leukotrienes, and nitrogen oxides, can be measured from EBC to assess airway inflammation and oxidative stress. Volatile organic compounds (VOCs) generated from tissue level metabolic processes are the primary foci in EBC-based metabolomic studies (Filipiak et al., 2012; Rufo et al., 2016). Inflammation, a hallmark of asthma, provides an important source of endogenous VOCs. Reactive oxygen species (ROS) induced by airway inflammation can degrade polyunsaturated fatty acids in membrane lipids that in turn generates VOCs (Dallinga et al., 2010). Specific EBC-based VOCs, such as butanoic acid and 3-(1-methylethyl)-benzene, can be used to diagnose pediatric asthma with high sensitivity and specificity (Dallinga et al., 2010). Furthermore, EBC-based VOC markers, such as 3,7-dimethylnonane, nonanal, and 1-propanol, can potentially distinguish between neutrophilic and eosinophilic types of asthma (Schleich et al., 2019). This distinction is critical to identify severe, corticosteroid-insensitive patients with asthma such that novel therapeutics can be developed (Carr et al., 2018; Fahy, 2009; Wang et al., 2016). Collectively, these studies suggest that EBC-derived VOCs may serve as promising biomarker candidates for diagnosing, phenotyping and monitoring therapeutic responses (Rufo et al., 2016). In addition to VOCs, EBC is also a source of other macromolecules and secretory markers relevant to lung pathology. For instance, the Vitamin A metabolite retinoic acid, prostaglandin derivatives (i.e.: 20-hydroxy-PGF2a, thromboxane B2), and nucleosides such as adenosine, are also present in EBC and their levels are altered in severe asthma patients (Carraro et al., 2013; Leung et al., 2013; Sinha et al., 2017).

Despite being a convenient and physiologically relevant sample for metabolomic screening, EBC has its limitations. Cysteinyl leukotrienes are elevated in EBC collected from moderate and severe asthma patients (Samitas et al., 2009). However, factors such as saliva contamination in EBC results in inaccurate measurements of metabolites, including amino acids and eicosanoids (Cruickshank-Quinn et al., 2017). Also, the metabolites in the bronchoalveolar region may not be accurately represented in the exhaled air captured from the upper airways. For instance, nonanal and decanal, two metabolites involved in lipid peroxidation and inflammation, significantly increased in the BALF from OVA-treated mice but remained low in EBC (Neuhaus et al., 2011). Thus, BALF from murine asthma models remains a preferred sample to investigate metabolomic basis of asthma pathogenesis. Mirroring the findings from other sample types, metabolites associated with energy, amino acid, lipid, and sterol metabolisms are altered in BALF from murine models of asthma (Ho et al., 2013; Jun Peng et al., 2014).

4.4. Metabolomics in translational medicine

The aforementioned studies clearly establish that asthma is associated with metabolomic changes, at the pulmonary and systemic levels (Table 1). The altered levels of metabolites detected in urine, blood, and EBC samples in asthma patients underscore the potential uses of these metabolites as diagnostic biomarkers and predictors of treatment outcome. The diagnostic biomarker application of metabolites in asthma was demonstrated in a recent pediatric asthma study, in which plasma sphingolipid levels at 6 months of age predicted early-onset asthma at the age of 3 years (Rago et al., 2021). Similarly, a survey-based study reported increased baseline bromotyrosine as a reliable indicator of poorly controlled asthma in children (Wedes et al., 2011).

Table 1.

Systemic metabolomics studies in asthma.

Samples	Participants/Models	Metabolism Pathways	Distinct metabolites	Ref#
Urine	Children with non-stable asthma (n = 20), stable asthma (n = 73), or healthy controls (n = 42)	Citric Acid Cycle	2-oxoglutarate, succinate, threonine, cis-aconitate	Saude et al. (2011)
		Lipid metabolism	2-hydroxyisobutyrate, 3-hydroxybutyrate, 3-methyladipate
	Children with asthma acute exacerbation (n = 69)	Citric Acid Cycle and pyruvate metabolism	Cis-aconitate, lactate, formic acid, acetate, citrate	Quan-Jun et al. (2017)
		Arginine and proline metabolism	Citrulline, sarcosine, ornithine, creatine, creatinine
	Adult subjects with or without asthma (n = 57 in total)	Lipid metabolism	2,4-dimethylheptane, 4-methylheptane, octanal, heptano	Loureiro et al. (2016)
	Adult subjects with mild-to-moderate asthma (n = 86), severe asthma (n = 411), or healthy control (n = 100)	Eicosanoid metabolism	LTE4, TetranorPGDM, 2,3-dinor-11β-PGF2α.	Kolmert et al. (2021)
	Children with recurrent wheezing (n = 32) or healthy controls (n = 13)	Tryptophan metabolism	glutaric acid, 5-hydroxy-l-tryptophan, indole-3-acetamide, and 3-indoleacetic acid	Carraro et al. (2018)
		Lipid metabolism	N-acryloylglycine and tiglylglycine
Plasma/Serum	Adult subjects with non-severe asthma (n = 10), severe asthma (n = 10), and healthy controls (n = 10)	Taurine metabolism Bile acid metabolism	Taurine, Arachidonate, β alanine, Taurocholate, glycodeoxycholate, lathosterol	Comhair et al. (2015)
	Female, 4-6-week-old BALB/c mice, challenged with OVA or PBS (n = 16 in total)	Purine metabolism	Uric acid, inosine	Yu et al. (2016)
		Glycerophospholipid metabolism	Lysophosphatidylcholines, phosphatidylserine
	Patients with acute asthma exacerbations (n = 120), and healthy controls (n = 120)	Uric acid metabolism	Uric acid	Abdulnaby et al. (2016))
	Patients with acute asthma exacerbations (n = 217), and healthy controls (n = 142)	Uric acid metabolism	Uric acid	Li et al. 2014
	Adult subjects with asthma (n = 232), or healthy controls (n = 26)	Arginine metabolism	Methylarginine, ornithine, citrulline	Lara et al. (2008)
	Adult subjects with asthma (n = 39), or healthy controls (n = 26)	Methyl transfer pathway	Formate, choline, methionine, O-phosphocholine, methanol	Jung et al. (2013)
		Arginine metabolism	Methylarginine, arginine
	Children with asthma (n = 35), food allergy (n = 35), asthma and food allergy (n = 35), or healthy controls (n = 20)	Lipid metabolism Amino acid metabolism	Glycerophosphoinositol, sphingomyelin, DHEA-S threonine, valine, glutamate	Crestani et al. (2020)
	Adult subjects with mild asthma (n = 12), moderate asthma (n = 20), severe asthma (n = 22), or healthy controls (n = 22)	Lipid metabolism	Cortisone, cortisol, DHEA-S Oleoylethanolamide, sphingosine-1-phosphate, 22-hydroxycholesterol, α-linolenic acid, N-palmitoyltaurine, sphinganine-1-phosphate, ceramides, sphingomyelins	Reinke et al. (2017)
	Pregnant women (n = 738) and their children (n = 700); Pregnant women (n = 881) and their children (n = 810)	Sphingolipid metabolism		Rago et al. (2021)
	Adult subjects with asthma (n = 51)	Lipid metabolism	Lysophosphatidylglycerol, phosphatidylcholine	Gai et al. (2019)
	Adult subjects with asthma (n = 18) or healthy controls (n = 17)	Tryptophan metabolism	Tryptophan, kynurenine, quinolinic acid, anthranilic acid	van der Sluijs et al. (2013)
EBC	Children with asthma (n = 63), or healthy controls (n = 57)	Lipid metabolism	Branched hydrocarbon (C13H28), 1-penten-2-on, butanoic acid, 3-(1-methylethyl)-benzene	Dallinga et al. (2010)
	Adult subjects with eosinophil or neutrophil asthma (n = 521 in total)	Lipid metabolism	3,7-dimethylnonane, 1-propanol, nonanal	Schleich et al. (2019)
	Children with non-severe asthma (n = 31), severe asthma (n = 11), or healthy controls (n = 15)	Vitamin metabolism	retinoic acid, vitamin D	Carraro et al. (2013)
		Adenosine metabolism	Deoxyadenosine
	Children with asthma (n = 89), or healthy controls (n = 20)	Ketogenesis	Hydroxybutyrate, acetate, acetone	Sinha et al. (2017)
	Adult subjects with mild asthma (n = 16), moderate asthma (n = 12), severe asthma (n = 15), or healthy controls	Eicosanoid metabolism	8-isoprostane, cysteinyl-leukotrienes	Samitas et al. (2009)
	Adult subjects with asthma (n = 18) or healthy controls (n = 17)	Tryptophan metabolism	Kynurenine, quinolinic acid	van der Sluijs et al. (2013)
		Arginine metabolism	Arginine, ornithine
BALF	Female, 5- to 6-week-old BALB/c mice challenged with	Lipid metabolism	Nonanal, decanal	Neuhaus et al. (2011)
	Female, 6- to 8-week-old BALB/c mice from 4 groups (sensitized: n = 12, asthma model: n = 12, asthma with dexamethasone: n = 12, or naïve: n = 12)	Lipid metabolism	Choline, hexadecenoylcholine	Ho et al. (2013)
		Energy metabolism	Lactate, malate
	Male, 10-week-old Brown Norway inflamed (n = 18), sensitized (n = 9) or naïve (n = 9) rats	Arginine-proline pathway	Arginine, proline	(Jun Peng et al., 2014)
	Male, 8-week-old BALB/c mice, sensitized with OVA or saline aerosol (n = 8 in total)	Arachidonic acid metabolism	Prostaglandin D2, 6-keto-prostaglandin F1alpha, 2S-hydroxy-5Z,8E,10E-heptadecatrienoic acid	Chiba et al. (2018)

Another potential application of asthma metabolomics is identifying distinct endotypes and metabotypes that would aid in personalized treatments. Based on its etiology and diverse pathological features, asthma is classified into different phenotypes and endotypes (Kaur and Chupp, 2019). Recent studies in asthma metabolomics seek to subclassify these asthma phenotypes into metabotypes. Studies suggest that EBC-based VOC markers and serum glycerophospholipids can distinguish between neutrophilic and eosinophilic types of asthma (Gai et al., 2019; Schleich et al., 2019). Other important parameters used for classification of asthma, such as FEV₁ and serum IgE levels, correlate well with the altered lipid and nucleotide metabolites (Kelly et al., 2017b; Loureiro et al., 2016). For instance, an untargeted metabolomic screening in urine clearly distinguished early-onset asthma from transient wheezing in preschool children, providing a valuable tool for personalized clinical management of these conditions (Carraro et al., 2018). Similarly, severity of rhinovirus-induced cold was positively correlated with systemic tryptophan catabolites in allergic asthma, suggesting blood screening for these catabolites has a prognostic value in rhinoviral infection and asthma exacerbations (van der Sluijs et al., 2013) Together, these studies support that metabolomics could be potentially used for the identification of asthma endotypes in practice.

Additionally, as previously mentioned, conventional asthma treatments can change urinary metabolomic profiles, suggesting that non-invasive metabolic assays can be developed to assess response to treatment (Kolmert et al., 2021; Quan-Jun et al., 2017). Supporting this potential, VOC profile measured by electronic nose distinguished steroid-responsive and steroid-unresponsive asthma, with relatively higher accuracy than traditional methods such as sputum eosinophils and FeNO (van der Schee et al., 2013). Therefore, further investigations integrating treatment-induced and disease-induced metabolomic profiles in asthma would benefit personalized and precision medicine in asthma.

While the aforementioned studies significantly contributed to biomarker discovery with potential diagnostic applications, elucidating the mechanistic link between asthma and metabolomic signature remains a challenge. In order to understand how metabolites modulate clinical and pathological features of asthma, metabolomic investigations at the cellular levels are essential.

5. Metabolomics in cells

As any cellular process, metabolism is a response to physiological and environmental cues. Metabolomic studies in homogenous cell populations represent a common investigative platform in a wide range of research areas, including drug discovery, diagnostic markers, toxicological assessment and signaling pathway identification (Hayton et al., 2017). Study of metabolomics in homogeneous cell cultures help identify signatures specific to a particular cell type, with minimal interference from other tissue components (Hayton et al., 2017). Single-cell metabolomics also facilitates the investigation of molecular mechanisms underlying the specific disease phenotypes manifested in the resident tissues. Therefore, cell-specific approaches to metabolomic screening offers opportunities to explore disease states driven by phenotypic modifications in single cell types (Table 2).

Table 2.

Cellular metabolomics studies related to asthma.

Cell types	Metabolism Pathways	Distinct metabolites	Affected cellular functions	Ref#
Dendritic cells	Glucose metabolism	Glucose	Motility and migration	Guak et al. (2018)
		Glucose, lactate	Activation and cytokine production	Krawczyk et al. (2010)
Macrophages	Glycolysis and citric acid cycle Purine and amino acid metabolism	D-Glucose, glyceraldehyde-3-P, citrate, Malate Carnitine, AMP, cysteine, allantoate, leucine	Proliferation and cytokine production	Mould et al. (2017)
	Glucose metabolism	Itaconate, GABA, 2-hydroxyglutarate	Polarization	Puchalska et al. (2018)
	Amino sugar and nucleotide sugar metabolism	UDP-GlcNAc, UDP-glucose, and UDP-glucuronate	Polarization	Jha et al. (2015)
	Glutamine/glutamate metabolism	Glutamine, glutamate, AKG, ornithine
T-lymphocytes	Glycolysis	Lactate	Activation and proliferation	Ostroukhova et al. (2012)
Eosinophils	Lipid metabolism	Protectin D1, 4-hydroxy DHA, arachidonic acid, docosahexaenoic acid	Biosynthetic activities, migration, and adhesion	Miyata et al. (2013)
Platelets	Glycolysis and citric acid cycle	ATP	Energy production	Xu et al. (2015)
Airway epithelial cells	Glycolysis	Glucose, lactate, pantothenate, ADP	Cytokine production	Qian et al. (2018)
	Citrate cycle	L-Malic acid, cis-aconitic acid	Energy production	(Qingyu Huang et al., 2015)
	Amino acid metabolism	Phenylalanine, tryptophan	Biosynthetic activities
	Glutathione metabolism	Oxidized glutathione	Oxidative stress
	Lipid metabolism	Phosphatidylcholine	Membrane remodeling	Nguyen et al. (2018)
Type II pneumocytes	Lipid metabolism	Phosphatidylcholine	Biosynthetic activities	Liu et al. (1997)
Airway smooth muscle cells	Lipid metabolism	Sphingosine 1-phosphate	Proliferation, contraction, and cytokine secretion	(Ammit et al., 2001; Liu et al., 2018a)
	Vitamin metabolism	Retinoic acid	Migration	Day et al. (2006)

Airway hyperresponsiveness (AHR), inflammation, and airway remodeling are the three pivotal characteristics of asthma and a plethora of pulmonary cells contribute to these functions (Busse and Lemanske, 2001). The phenotypic heterogeneity of asthma presents a challenge for universally successful treatments. Mammalian lungs contain more than 40 different cell types, including airway smooth muscle (ASM) cells, immune cells, epithelial cells, and vascular smooth muscle cells (Franks et al., 2008). Therefore, we posit that single-cell metabolomics studies on select cell types may shed light on mechanisms of the three pivotal characteristics of asthma (Fig. 2).

Fig. 2. Cell types and their metabolic roles in asthma.

Each tissue in lungs plays its unique role in asthma pathology. Each tissue is composed of differentiated cell types, homogeneous or heterogeneous, with critical functions. In these cells, altered mitochondrial functions and metabolism of macro- and micro-nutrients -elaborated in the main text- modulate the specialized cellular functions. Airway smooth muscle (ASM) cells primarily dirve bronchial spasm in asthma, which is charaterized by increased cytosolic Ca²⁺ and myosin light chain (MLC) phosphorylation. Asthma is characterized by airway inflammtion, remodeling, and hyperresponsiveness. DC - dendritric cells, T_reg - regulatory T-lymphocytes, Ca²⁺ - cytosolic Ca²⁺, pMLC - phosphorylated myosin light chain.

5.1. Inflammatory cells in asthma

Airway inflammation, a salient feature in asthma, contributes to airway hyperresponsiveness and remodeling (Chapman and Irvin, 2015; Murdoch and Lloyd, 2010). Stimulated by the allergens, antigen-presenting cells (APCs), such as dendritic cells (DCs), process and present the allergens to activate allergen-specific T cells (Gordon, 2002). Activated T helper cells, mostly T helper 2 (Th2) cells, release cytokines and chemokines to elicit cellular responses (Zimmermann et al., 2003). The key T_H2 cytokines, IL-4, IL-13, and IL-5, regulate the pivotal events of allergic inflammation, such as B cell activation (IgE synthesis), eosinophil recruitment, and goblet cell metaplasia (Ingram and Kraft, 2012; Lorentz et al., 1999; Maes et al., 2012). Other cell types such as mast cells, regulatory T (T_reg) lymphocytes, natural killer (NK) cells, and basophils are also actively involved (Murdoch and Lloyd, 2010). Compelling evidence shows that nutrient-derived metabolites regulate the functions of these inflammatory and immune cells (Ganeshan and Chawla, 2014; Veldhoen and Ferreira, 2015).

5.1.1. Antigen presenting cells (APCs)

As the first responders for injury, APCs, including DCs and macrophages, play fundamental roles in inflammatory responses. Studies showed that a bioenergetic switch from oxidative phosphorylation to glycolysis characterizes migration and differentiation of DCs (Guak et al., 2018; Mishra, 2017). The migration and differentiation of DCs are critical for the downstream interaction of DCs with other immune cells. Also, the inflammatory stimulation of DCs regulates subsequent NO production and reduces coupled respiration, which in turn promotes metabolic reprogramming towards glycolysis in DCs (Guak et al., 2018). This metabolic reprogramming of DCs is primarily supported by glycogen metabolism at the early stage, which can further increase inflammatory cytokine production, including TNF-α. (Krawczyk et al., 2010; Thwe et al., 2017). Another prominent APC in the lungs is the alveolar macrophage, which surveys and provides innate immune defense on alveolar surface. Key metabolic pathways, such as glycolysis, amino acid synthesis, and lipid metabolism have regulatory impact on proliferative and secretory response of the alveolar macrophages (Viola et al., 2019). Increased glycolysis and amino acid catabolism drive macrophages to adopt a pro-inflammatory M1 phenotype, while the pro-resolution M2 macrophages are dependent on oxidative phosphorylation in mitochondria (Puchalska et al., 2018; Viola et al., 2019). These differences may be due to the dependence of M2 macrophages on glutamine synthesis and N-glycosylation for their differentiation and polarization, which requires intermediates supplied with intact TCA cycles (Jha et al., 2015). Meanwhile, aerobic glycolysis is upregulated to meet elevated ATP demand of M1 macrophages during proliferation and differentiation (Koo and Garg, 2019; Van den Bossche et al., 2015; Yu et al., 2020). Resident alveolar macrophages showed increased proliferation that correlated with increase in TCA cycle and amino acid metabolism (Mould et al., 2017). A similar association between metabolism and function is demonstrated in macrophages in extra-pulmonary tissues as well (Liu et al., 2016). Clinical and epidemiological studies show that obesity, a clear outcome of metabolic perturbation, is linked to late-onset asthma and decreased response to mainstream asthma treatment (Kim et al., 2015; Orfanos et al., 2018; Peters et al., 2018). In obesity, elevated accumulation of pro-inflammatory M1 macrophages in adipose tissue is linked to low-grade chronic inflammation, which contributes to various obesity-related co-morbidities (Chawla et al., 2011; Weisberg et al., 2003).

5.1.2. T- lymphocytes

Various subsets of T-lymphocytes have pivotal roles in asthma. Studies have revealed unique metabolic signatures of T cells among asthma patients. Aerobic glycolysis was elevated in T-cells from asthma patients, characterized by increased lactate production, higher expression of hexokinase (HK) and phosphofructokinase (PFKFB) in T cells accompanied with elevated T cell proliferation (Ostroukhova et al., 2012). Pharmacological inhibition of aerobic glycolysis also attenuated airway inflammation and hyperreactivity in a murine model of asthma, suggesting that elevated glycolic pathways play a causal role in AHR (Ostroukhova et al., 2012). These observations resemble the Warburg effect, where neoplastic cells switch their metabolism from oxidative phosphorylation to aerobic glycolysis to enhance proliferation (Warburg et al., 1927). Functions of T regulatory (T_reg) cells, which negatively regulate airway inflammation, are also be modulated by metabolites. 12, 13-diHOME, a metabolite of the unsaturated long chain fatty acid linoleic acid, reduced secretion of anti-inflammatory cytokines from DCs, via activation of peroxisome proliferator-activated receptor-γ (PPARγ). Subsequently, the PPARγ activation decreased the number of T_reg cells and enhanced lung-resident T cells, neutrophils, and macrophages in an allergen-induced mouse model of asthma (Levan et al., 2019). In pediatric asthma, the serum levels of 12,13-diHOME were positively correlated with airway inflammation and atopy (Levan et al., 2019).

Downstream to APCs and T cells, granulocytic inflammatory cells drive airway inflammation. Eosinophils, a key granulocyte in T_h2-driven inflammation, exhibited altered metabolic status in asthma. Protectin D1 (PD1) and 4-hydroxy DHA, metabolites of docosahexaenoic acid (DHA), was lower in asthma patients. Attenuated eosinophilic PD1 level appears to decrease its negative regulation on eosinophils and restrict the pro-resolving functions of eosinophils, which in turn exacerbate eosinophil inflammation in asthma (Miyata et al., 2013). Studies, although limited in number, show that even low-abundant cell types such as natural killer (NK) T cells adopt a unique metabolic profile in inflammatory diseases like asthma (O’Brien and Finlay, 2019). Evidence also shows that platelets, with their innate immune functions, show a metabolic shift characterized by increased TCA cycle in asthma (Li et al., 2012; Xu et al., 2015).

5.1.3. Airway epithelial cells

Airway epithelial cells serve as the first line of defense in lungs by providing physical barrier and their secretory and synthetic roles. Evidence shows metabolic dysfunctions in airway epithelial cells in human asthma and animal models of the disease. In asthma patients and HDM-induced murine models of asthma, airway epithelial cells manifested elevated glycolysis, characterized by activated IL-1/inhibitory kappa B kinase ε (IκBε) signaling (Qian et al., 2018). In a ragweed pollen extract-induced mouse model of allergic asthma, levels of carbonylated proteins were increased in mitochondrial respiratory complexes, leading to mitochondrial dysfunction and ROS generation, which further enhanced epithelial inflammation and airway hyperresponsiveness (Aguilera-Aguirre et al., 2009). This mitochondrial dysfunction has also been widely reported as a pivotal mechanism in lung diseases. Frequent exacerbation is a salient clinical feature of severe asthma and ambient air pollutant exposures are well known causes of this exacerbation. Human airway epithelial cells, following fine particulate matter (PM_2.5) exposure, exhibited altered TCA cycle and amino acid biosynthesis, characterized by decreased levels of cis-aconitate and malate, and increased levels of phenylalanine and tryptophan (Qingyu Huang et al., 2015). Importantly, PM_2.5 exposure also elevated oxidized GSH (GSSG), disturbed glutathione metabolism, and promoted oxidative damage of mitochondria in airway epithelial cells (Qingyu Huang et al., 2015). Taken together, these metabolic changes imply abnormal oxidative stress status in airway epithelial cells. Rhinovirus, another common cause of asthma exacerbation, modulates lipid metabolism in bronchial epithelium. In primary human bronchial epithelial cells, rhinovirus infection modified the activities of a multiple lipid-modifying enzymes, to alter fatty acid elongation and desaturation, suggesting a regulatory role for these long chain fatty acids in airway epithelial cell function in response to viral infection (Nguyen et al., 2018). Obesity, a risk factor for refractory asthma, is also associated with frequent exacerbations and is positively correlated with metabolic disorders (Baffi et al., 2015; Peters et al., 2018). The surfactant layer, composed of lipids and proteins, lines alveolar epithelia and is critical for normal lung function. Phospholipids are major components of alveolar surfactants, which are essential for surface-tension lowering and innate immunity in the lungs (Agassandian and Mallampalli, 2013). Compared to the lean subjects, those with higher BMI showed unique lipidomic compositions and increased sputum phospholipid concentrations, such as phosphocholines (PC) and phosphoglycerols (PG), in lung epithelial lining fluid (Brandsma et al., 2018). Increased concentrations of PG and phosphatidylethanolamine (PE) species are also found in the lungs from mice fed with obesogenic diet, suggesting a link between obesity and airway epithelial lining lipid composition in the context of obesity-associated asthma (Showalter et al., 2018). Due to their juxtaposition to a variety of other lung cells, airway epithelial cells are also regulated by external metabolites through paracrine signaling. For instance, 12/15-lipoxygenase, an enzyme in prostaglandin synthesis cascade, catalyzes arachidonic acid in alveolar macrophages and fibroblasts (Bryant et al., 1982; Mabalirajan et al., 2013). The metabolites generated from this catabolism, such as 12-S-HETE, act on surrounding airway structural cells and mediate epithelial injury response in airway diseases (Mabalirajan et al., 2013).

The type II pneumocytes lining alveoli secrete surfactants and play mechanical and defense roles in lungs (Andreeva et al., 2007). Phospholipid metabolism, including phosphatidylcholine synthesis, was attenuated in type II pneumocytes isolated from OVA-challenged guinea pigs (Liu et al., 1997). Investigations using gene knockout mouse models suggested a critical role for ATP-binding cassette transporter A1 (ABCA1) in maintaining the lipid homeostasis, including phospholipid and cholesterol content, in the type II pneumocytes (Bates et al., 2005). Repair and regeneration of airway epithelium are critical steps in lung injury response and a variety of airway progenitor cells drive these functions. A recent study showed that glycolysis was critical for proliferation of the airway progenitor cells, Club cells and variant Club cells. Blocking glycolysis attenuated proliferation of club cells but promoted their differentiation into goblet cells and ciliated epithelial cells, suggesting a critical role for metabolic signals in lung epithelial injury response (Li et al., 2019a). As far as we are aware, the metabolic profiles of goblet cells, a key cell type linked to innate defense mechanism, have not been investigated. Further investigations are needed to understand the role of macronutrient metabolism in these cell types in airway epithelium.

5.1.4. Airway smooth muscle cells

Airway smooth muscle (ASM) cells are the pivotal structural cells contributing to AHR. ASM cells isolated from asthma donors show amplified excitation-contraction coupling and agonist-induced contractile response, thus ASM remains as an important therapeutic target in asthma. Studies on ASM cell metabolism have focused on mitochondrial bioenergetics and described that ASM mitochondrial dysfunction is linked to airway diseases such as COPD and asthma (Delmotte and Sieck, 2019; Prakash et al., 2017). For instance, a cigarette smoke (CS) exposure model showed that mitochondrial fission/fusion proteins modulate the proliferation and apoptosis of ASM cells (Aravamudan et al., 2017). In another study to elucidate excitation-energy coupling, proinflammatory cytokine TNF-α attenuated mitochondrial mobility in human ASM cells to modulate cytosolic Ca²⁺ buffering functions (Delmotte et al., 2017). These observations posit a hypothesis that metabolic alterations in ASM cells contribute to AHR and asthma onset.

AHR, a salient clinical sign in asthma and chronic obstructive pulmonary disease (COPD), manifests as an amplified contractile response of ASM cells to contractile agonists (Chapman and Irvin, 2015). A variety of endogenous mediators, including cytokines and metabolites, are involved in amplifying the contractile response of ASM cells (Black et al., 2012). Products of arachidonic acid metabolism act as key mediators of AHR. Prostaglandin D2 (PGD₂) is known as an inflammatory mediator functioning in several diseases, including asthma (Matsuoka et al., 2000; Mohri et al., 2006). In an OVA-induced allergic asthma mouse model, arachidonic acid (AA) metabolism is driven towards PGD₂ production pathway by elevated prostaglandin D synthase expression in bronchial smooth muscle tissue, which elicited ASM hyperreactivity (Chiba et al., 2018). Cytosolic Ca²⁺ mobilization is a key component of excitation-contraction coupling in ASM cells. Agonist-induced cytosolic Ca²⁺ mobilization initiates a signaling cascade to activate myosin light chain kinase (MLCK) and MLC phosphorylation resulting in cell shortening in ASM cells (Kamm and Stull, 1985, 2001; Somlyo and Somlyo, 2003). In bronchial smooth muscle cells derived from asthma patients, calcium/calmodulin-dependent protein kinase IV (CaMK-IV) is activated to enhance calcium influx and increase mitochondrial biogenesis (Trian et al., 2007). The motility of mitochondria in ASM cells can module calcium uptake to modulate the contractile responses to agonists, such as histamine and acetylcholine (Delmotte et al., 2017). Cytosolic calcium levels are also modulated by various signaling cascades driven by a variety of metabolic by-products (Berridge, 2008; Somlyo and Himpens, 1989). Particularly, lipid metabolites regulate both calcium mobilization and sensitization pathways to modulate ASM cell shortening. Sphingosine 1-phosphate, the sphingolipid metabolite, is elevated in BALF following allergen challenge (Ammit et al., 2001), which further amplifies calcium mobilization and sensitization and evokes ASM hyperreactivity via inhibition of MLC phosphatase (Ammit et al., 2001; Kume et al., 2007; Rosenfeldt et al., 2003). In addition, there are multiple lines of evidence to suggest that dysfunctional metabolic status is associated with increased asthma incidence, severity, and resistance to treatment. Our studies showed that ASM cells obtained from obese human donors showed enhanced agonist-induced cell shortening accompanied with amplified MLC phosphorylation and calcium mobilization (Orfanos et al., 2018). Other investigators have reported distinct metabolic phenotypes among obese asthma patients, based on energy, amino acid, and lipid metabolisms (Maniscalco et al., 2017). Particularly, vitamin D deficiency has been identified as a risk factor of obesity-associated asthma (Mirzakhani et al., 2017; Vo et al., 2015). In ASM cells, vitamin D metabolism plays a role in decreasing AHR and airway remodeling through the inhibition of ASM cell growth and cytokine secretion (Damera et al., 2009; Hall et al., 2016; Himes et al., 2015).

Airway remodeling involves increased thickness of airway wall, characterized by myocyte hypertrophy, hyperplasia, angiogenesis, and increased extra-cellular matrix (ECM) deposition (Hirota and Martin, 2013; Lazaar and Panettieri, 2005). Mitochondrial energy metabolism is directly linked to cell proliferation in normal and cancer cell lines (Bhatti et al., 2017; Ganapathy-Kanniappan and Geschwind, 2013; Lunt and Vander Heiden, 2011; Vyas et al., 2016). In ASM cells, findings showed that increased mitochondrial mass and activity drive bronchial ASM cell proliferation in vitro (Trian et al., 2007). In fetal human ASM cells, hyperoxia decreased proliferation via modulating mitochondrial function, along with decreased fusion protein Mfn2 expression and increased fission protein Drp1 (Hartman et al., 2012). Furthermore, inhibition of glycolysis inhibits cell cycle progression and attenuates LPS-induced ASM cell proliferation (Zhang et al., 2019). Supporting a similar role for mitochondria in ASM cell proliferation, ASM cells from COPD patients showed reduced mitochondrial respiration with enhanced glycolysis. Inhibition of glycolysis attenuated proliferation of ASM cells obtained from COPD patients (Michaeloudes et al., 2017). Similar findings were reported in vascular smooth muscle cells obtained from subjects with pulmonary arterial hypertension (Hernandez-Saavedra et al., 2020). The upregulated glycolytic activity is likely an adaptive mechanism elicited by the mitochondrial dysfunction in highly proliferative ASM cells (Sutendra et al., 2010). These results underscore the metabolic basis for the phenotypic changes in smooth muscle cells in pulmonary disorders.

Lipid mediators also play a multitude of roles in airway remodeling. For instance, sphingosine 1 phosphate induces ASM cell proliferation via the activation or upregulation of the transcription factors, yes-associated protein (YAP) and forkhead box M1 (FOXM1) (Liu et al., 2018a). Sphingosine analog pre-treatment inhibited the proliferative capacity of ASM cells by interfering with oxidative phosphorylation and other metabolic processes (Blais-Lecours et al., 2020). Free fatty acids (FFAs) also elicit mitogenic effects in ASM cells. These free fatty acids act on nutrient-sensing G protein-coupled receptors (GPCRs) on target cells to elicit downstream effects (Thorburn et al., 2014). Both short-chain and long-chain FFAs, are associated with the activation of MAP kinases (MAPK), PI3 kinase (PI3K)/mTOR pathways downstream to the GPCRs (Thorburn et al., 2014; Yasuda et al., 2014). Free fatty acid receptor 1 (FFAR1) is reported to induce the proliferation of ASM cells via MAPK and PI3K pathways (Matoba et al., 2018).

ASM cell migration is another area where metabolic signaling is known to play a role. Cell migration is a multi-step process, coordinated by actin-driven elongation, adhesion, and contraction (Friedl and Wolf, 2009). Several endogenous mediators and exogenous substances, including lipid derivatives, vitamin metabolites and β₂-adrenergic receptor agonists are associated with the migration of ASM cells (Day et al., 2006; Goncharova et al., 2012; Salter et al., 2017).

5.1.5. Other cells and lung tissue models

A small number of studies have characterized lung tissue to understand the metabolic regulation of asthma. L-ornithine-derived polyamines (PAs), derived from impaired L-arginine catabolism, are of particular interest in certain airway diseases. Ornithine-derived PAs are elevated in the sputum of patients with allergic asthma and in allergen-induced murine models of asthma (North et al., 2013). Altered metabolism of a multitude of PAs was reported in lung tissues isolated from murine models of asthma, characterized by increased levels of putrescine, N1-acetylputrescine, and N8-acetylspermidine (Lee et al., 2019). Since lung tissue provides an integrated, multi-cellular platform almost similar to the in vivo environment, metabolomic investigations in lung tissues will have significant translational value.

In addition to asthma, metabolomic profiles of human lung tissues and other cell types have also been studied in other lung diseases. In certain types of lung cancers, metabolic dysfunctions were characterized in human lung tissues and identified changes such as increased glycolysis and TCA cycle metabolites (Fan et al., 2009; Hensley et al., 2016). Similar hyper-glycolytic phenotypes were also identified in pleural effusions with a single-cell on-chip metabolic cytometry assay (Li et al., 2019b). Pulmonary hypertension (PH), although considered a vascular disease by etiology, is a disease with clinical consequences in lungs. Pulmonary artery endothelial cells (PAECs) obtained from PH patients showed elevated levels of fructose bisphosphate and succinate, suggesting increased glycolysis and glutamine-derived anaplerosis, respectively (Hernandez-Saavedra et al., 2020). Idiopathic pulmonary arterial hypertension impairs NO production, which in turn impairs mitochondrial biogenesis, resulting in a metabolic switch to glycolysis-dominant bioenergetics in PAECs (Xu et al., 2007). In summary, complex platforms such as lung tissue have been used in a small number of metabolomic studies. Ex vivo platforms such as precision-cut lung slices and organoids are yet to be utilized as metabolomic platforms to study pulmonary diseases such as asthma.

6. Conclusion

Over the recent years, metabolomics has emerged as a promising tool in multiple areas of biomedicine. A complex disease with multi-etiology, asthma is well-suited to be investigated by metabolomic approach to understand its pathophysiology. Currently, a vast majority of these investigations focus on distinct metabolite biomarkers for asthma diagnosis. We propose a highly focused metabolomic approach by targeting individual cell populations in lungs. ASM cells are primary drivers of bronchial spasm and show altered contractile phenotype in asthma. Therefore, investigating the distinct metabolic profile of ASM cells in asthma and obesity will be a fruitful endeavor. Finding from such studies will not only improve our understanding of obesity-associated asthma, but also potentially facilitate the optimization of asthma therapy with precision medicine (Wishart, 2016).

Arguably, there exists limitations in the metabolomics studies in asthma. First, there are high variabilities in findings among different studies due to confounding external environmental factors and methodological variations. To address this, internal and external validations should be used to confirm the findings (Kelly et al., 2017a; Kim H. Esbensen, 2010). Secondly, in order to minimize artificial inferences, one must ensure the “homogeneity” of the human samples used in the study. For instance, metabolites from xenobiotics may confound metabolomic profiles. Common asthma treatments include short-acting β-agonists (SABAs), long-acting β-agonists (LABAs), and glucocorticoids. Studies have suggested that albuterol administration has an impact on sphingolipid metabolism and associated changes in gene expressions (McGeachie et al., 2015). Thirdly, in disease states, compensatory mechanisms may drive certain metabolic resulting in findings that are not actually reflective of the pathological process. Most of these challenges can be rectified through an experimental design to address gender, ethnicity, and BMI variabilities of the donor samples. Despite limitations, cell- and tissue-based metabolomics may offer novel platforms to explore new therapeutic targets and or pathways to understand the pathogenesis of airways disease.