Skip to content
Publicly Available Published by De Gruyter August 31, 2018

Application of NMR metabolomics to search for human disease biomarkers in blood

  • Zikuan Song ORCID logo , Haoyu Wang , Xiaotong Yin , Pengchi Deng and Wei Jiang EMAIL logo

Abstract

Recently, nuclear magnetic resonance spectroscopy (NMR)-based metabolomics analysis and multivariate statistical techniques have been incorporated into a multidisciplinary approach to profile changes in small molecules associated with the onset and progression of human diseases. The purpose of these efforts is to identify unique metabolite biomarkers in a specific human disease so as to (1) accurately predict and diagnose diseases, including separating distinct disease stages; (2) provide insights into underlying pathways in the pathogenesis and progression of the malady and (3) aid in disease treatment and evaluate the efficacy of drugs. In this review we discuss recent developments in the application of NMR-based metabolomics in searching disease biomarkers in human blood samples in the last 5 years.

Introduction

Metabolomics is a new biological analysis method to provide precise analysis of small molecules (<1500 Da) associated with human metabolism, which has been widely used in the field of systems biology in recent years. Unlike DNA, RNA or proteins (including enzymes), these small molecules can accurately reflect the most direct metabolic changes in our body under a certain condition in a short time, which means metabolomics gives a new approach to help clinical disease research [1]. Furthermore, metabolomics can perfectly complement transcriptomics, genomics and proteomics on account of its ability to analyze the changes of small metabolites, which are totally different kinds of compounds from DNA, RNA and proteins [2]. Compared with other omics, metabolomics provide more direct and complete information about the biological system from an individual or a group [3], which can be used for disease diagnosis [4], [5], [6], [7], [8] and treatment [9], [10], drug toxicological mechanism analysis [11], [12], and precision medicine [13], etc.

Metabolomics must be based on powerful detection techniques. The sensitivity and specificity of detection techniques increased rapidly, allowing more precise characterization and quantification of numerous metabolites in a single sample [2]. The two most successful detection methods used in metabolomics are nuclear magnetic resonance (NMR) and mass spectrometry (MS, including liquid chromatography-mass spectroscopy [LC-MS] and gas chromatography-mass spectroscopy [GC-MS]). In addition, blood [14], urine [15], saliva [7] and diseased tissues [16] are commonly used as detection samples for metabolomics analysis. As a high-throughput method, NMR can be used to analyze hundreds of small molecules simultaneously and efficiently in human samples including blood, serum and tissues [17]. Almost every metabolite has their obvious and reproducible NMR signature, therefore NMR can be used to explore the metabolic processes of different organisms or screen metabolites. Moreover, NMR costs less than MS and it does not require any physical or chemical treatment to the sample before analysis, which means there is no destruction to the test sample, so that the same sample can be re-tested several times if desired [2], [17], [18]. Therefore, NMR can provide high-throughput and reproducible analysis of metabolites in a more economical manner [19] to determine the structure of unknown compounds, elucidate the mechanism of metabolite transformation and screen drugs [20]. Moreover, NMR-based metabolomics have been incorporated with corresponding statistical techniques and performed with independent samples to identify disease-specific changes of metabolites and further validate disease biomarkers [21]. Figure 1 illustrates the general workflow of NMR-based metabolomics studies in disease-related biomarkers discovery.

Figure 1: General workflow of NMR-based metabolomics studies in disease-related discovery of biomarkers.
Figure 1:

General workflow of NMR-based metabolomics studies in disease-related discovery of biomarkers.

Biomarkers are critical for disease research. The disturbances of glucose metabolism, lipid metabolism and protein metabolism occur in the onset and development of many diseases, and the representative changes of metabolites in patients’ samples can be screened by high-throughput NMR and statistical analysis. For clinical purposes, these metabolites can be used to diagnose or stage diseases [22], [23]. In addition, if metabolites that significantly change in the early stages of some diseases can be identified, they can be used as predictors to predict the risks of these diseases [24]. These changed metabolites can also serve as novel targets for drug therapy and may even be used in precision medicine [13]. For research purposes, specific metabolic changes occur during the progression of diseases and blood-based metabolite markers could improve disease diagnosis, and changes in the metabolites can be tracked to specific pathways and further combined with metabolomics, proteomics and genomics to provide mechanistic information about the pathogenesis of diseases.

It is worth mentioning that statistical analysis is a cornerstone of metabolomics, which can be used to differentiate metabolites between healthy and diseased tissues, and determine the degree of association between these metabolites and clinical phenotypes or variables [2]. Multiple statistical analysis methods were used to ensure that researchers can get satisfactory results from the NMR data. For example, most analytical techniques including NMR produce complex multi-dimensional data sets, so dimension reduction techniques such as principal components analysis (PCA) and various clustering methods should be used as a helpful tool for the preliminary survey of comprehensive properties of the data [2]. Besides, considering the large amount of metabolites measured in a single assay, multiple tests have to be carried out but it may increase the possibility of type I error. In order to solve this problem, test results should be adjusted by using family wise rate or FDR [25], [26].

As an integral part of the circulatory system, blood is closely related to every organs and tissues in the human body. When disease occurs, the metabolites are released from the disease tissues into blood, so the changes of metabolites in blood can reflect the metabolic changes in disease tissues in parallel. Besides, blood samples are readily available and the acquisition process is less invasive compared to tissue samples. Based on that, blood can be used as one of the most classic and common sample for NMR analysis, especially NMR-based metabolomics research. In addition, human serum metabolome included the complete complement of metabolites present in serum and was used to facilitate searching for disease biomarkers in human blood samples increasingly with the deepening of research in this field. As for the reasons, metabolites belonging to the human serum metabolome are the closest connection between the genotype and phenotype and are the end products of numerous metabolic pathways, so they have the great potential to reflect and indicate diseases as biomarkers. Moreover, the measurement of the entire serum metabolome rather than a single metabolite in a single assay made the process of searching biomarkers more convenient and the results more indicative [27], [28], [29]. Psychogios et al. established the Serum Metabolome Database that contained 4229 confirmed and possible serum/plasma compounds, which can absolutely facilitate future research into blood chemistry and blood metabolomics to search disease biomarkers [14].

Therefore, we will particularly review the NMR-based metabolomics analysis of human blood (including plasma and serum) samples for searching biomarkers of diseases in the last 5 years (i.e. since and including 2013) (Figure 2), trying to clarify that NMR-based metabolomics play an important role in searching disease biomarkers, as well as biomarkers have great potential in disease prediction, diagnosis and treatment.

Figure 2: Number of papers published in the last 5 years (2013–2017) on NMR-based metabolomics of blood for searching and identifying biomarkers, in relation to specific disease type and the year of publication.Source: PubMed databases.
Figure 2:

Number of papers published in the last 5 years (2013–2017) on NMR-based metabolomics of blood for searching and identifying biomarkers, in relation to specific disease type and the year of publication.

Source: PubMed databases.

Diseases studied by NMR-based metabolomics of blood samples for searching and identifying biomarkers

According to the International Classification of Diseases (ICD-10 Version: 2016, http://apps.who.int/classifications/icd10/browse/2016/en) issued by the World Health Organization, we classified diseases into several types, and all of these diseases will be addressed in decreasing order of the number of research papers we found, in the following text. In Table 1, we list the changed metabolites or biomarkers identified by NMR-based metabolomics analysis of blood samples (including plasma and serum), from 2013 to 2017.

Table 1:

Changed metabolites or biomarkers identified by NMR-based metabolomics analysis in blood samples (including plasma and serum), from 2013 to 2017.

DiseasesaSources of samplebChanged metabolites/identified biomarkerscControl groupsdReference
Neoplasms
 Hepatocellular carcinomaHAscorbate (2.028)e, oxaloacetate (2.789), glycerol (1.540), isobutyrate (6.188), β-hydroxybutyrate (1.745) ↑HC[30]
Formate (0.071), tyrosine (0.759) ↓
Ascorbate (8.555), oxaloacetate (1.953), glycerol (1.472), isobutyrate (3.235), β-hydroxybutyrate (1.694) ↑Liver cirrhosis
Formate (0.200), tyrosine (0.398) ↓
 Pancreatic ductal adenocarcinomaRAscorbate, lipid, α-glucose, β-glucose ↑HC[31]
3-Hydroxybutyrate, asparagine, creatine, glutamate, glycerol, glycine, isoleucine, lactate, leucine, methylmalonate, phenylalanine, phosphocholine, pyruvate, succinate, tyrosine, valine ↓
 Poor differentiated pancreatic cancerMCholine ↑HC[32]
Lactate, glutamate, glutamine, amino acids ↓
 Malignant pancreatic and periampullary adenocarcinomasHGlutamate, urea, succinate, galactose ↑Benign[33]
 Benign pancreatic and periampullary adenocarcinomasHArginine, creatine, glutamine, ornithine, proline, alanine ↑Malignant
 Pancreatic islet β cell tumorMMethionine, citrate, choline ↑HC[34]
Acetate, taurine, glucose ↓
 Colorectal cancerHGlycine ↑HC[35]
Pyridoxine, orotidine, S-adenosylhomocysteine, pyridoxamine, glycocholic acid, β-leucine, 5-methylcytidine, taurocholic acid, 3-hydroxybutyric acid, 7-ketocholesterol, 3-hydroxyisovaleric acid, L-fucose, cholesterol, L-palmitoylcarnitine ↓
 Colorectal cancer (II–IV stage)HGlycine, cholesterol, taurocholic acid, cholesteryl ester, deoxyinosine ↑Colorectal cancer (0–Ⅰ stage)[36]
Pyridoxine ↓
 Lung cancerHGlucose ↑HC[37]
Lactate, phospholipid ↓
 Lung cancer (stage 3)H2-Hydroxybutyrate, 2-oxoisocaproate, acetate, carnitine, 3-hydroxyisovalerate, 2-hydroxyisovalerate, glycerol, glycineLung cancer (stage 1 and 2)[38]
 Non-small cell lung cancer (NSCLC)HMethanol (1.792), glutamate (1.379), creatine (1.333), lactate (1.315) etc. ↑HC[39]
Choline-N(CH3)3+ (0.747), histidine (0.784), HDL (CH3) (0.826), threonine (0.847) etc. ↓
HN-Acetylated glycoproteins, leucine, lysine, mannose, choline, lipid (CH3 (CH2)n) ↑HC[40]
Acetate, citrate, methanol ↓
Isoleucine, acetoacetate, creatine, N-acetylated glycoproteins, glycerolCOPD
 Advanced NSCLCHGlycerol, NAC1, NAC2 ↑Early NSCLC[40]
Isoleucine, acetoacetate ↓
 Renal cell carcinomaHAlanine, creatine, isoleucine, lactate, leucine, valine ↑HC[41]
Choline ↓
2-Oxoisocaproate, creatine, isoleucine, glutamate, ornithine ↑Benign disease[42]
Citrate, methanol, threonine glycine, histidine taurine, glutamine ↓
 Breast cancer (high IP3Rs expression)HLactate, lysine, alanine ↑HC[43]
Pyruvate, glucose ↓
 Metastatic breast cancerHAcetoacetate, glycerol, pyruvate, N-acetyl glycoproteins, mannose, glutamate, phenylalanine ↑Early breast cancer[44]
Histidine ↓
HGlucose, lactate, pyruvate, acetate, acetoacetate, β-hydroxybutyrate, urea, creatine, creatinine[45]
 Endometrial cancerH3-Hydroxybutyrate, C6 (C4:1DC), hexanoylcarnitine C14:2HC[46]
 Prostate cancerHAmino acids, glucose, glycerophospholipids (three lysophosphatidylcholines and five phosphatidylcholines), acylcarnitine ↑Benign prostatic hyperplasia[47]
Triglycerides ↓
 Prostate cancerHAlanine, pyruvate, sarcosine ↑HC[48]
Glycine ↓
 High-grade PCaHAlanine, pyruvate, glycineLow-grade PCa
 Abnormal prostateHGlycine, sarcosine, alanine, creatine, xanthine, hypoxanthineHC[49]
 Prostate cancerHAlanine, sarcosine, creatinine, glycine, citrateBenign prostatic hyperplasia
 Prostate cancer and benign prostatic hyperplasiaHGlycine, sarcosine, alanine, creatine, xanthine, hypoxanthineHC
 Nasopharyngeal carcinomaHMethionine (3.500), taurine (1.478), choline (1.790) ↑HC[50]
Lipids (0.178), isoleucine (0.543), LDL (0.626), trimethylamine oxidase (0.370), glucose (0.473) ↓
 Multiple myelomaHCarnitine, acetylcarnitineHC[51]
 Burkitt lymphomaMIsoleucine, leucine, pyruvate, lysine, α-ketoglutarate, betaine, etc.Wild-type[52]
Glutamate, glycerol, choline ↑
Diseases of the circulatory system
 Yin-deficiency and Yang-hyperactivity syndromeHBetaine, mevalonic acid, corticosterone, β-leucine ↑HC[53]
 Yin-Yang deficiency syndromeBetaine, mevalonic acid, corticosterone, β-leucine ↓HC
Betaine, mevalonic acid, corticosterone, β-leucine, propionic acid ↓Yin-deficiency/Yang-hyperactivity syndrome
 Left ventricular hypertrophy secondary to arterial hypertensionH–CH2–/–CH3 ratio ↑Hypertensive without LVH and HCs[54]
 Microvasclar diseaseH2-Hydroxybutirate, alanine, leucine, isoleucine, N-acetyl groups ↑HC[55]
Creatine/phosphocreatine, creatinine, glucose ↓
3-Hydroxybutirate, acetate ↑Stenotic disease
2-Hydroxybutirate, alanine, leucine, N-acetyl groups ↓
 Stenotic disease3-Hydroxybutirate, acetate ↑HC
2-Hydroxybutirate ↓
 Coronary heart diseaseHLeucine, N-acetyl glycoprotein, α-glucose, β-glucose, phenylalanine, acetone, HDL, glutamate, glutamine, methylamine, lysine, tyrosine, ornithine, taurine, proline, lactic acid, tryptophan, threonine, aspartic acid, valine, acetyl glutamic acid, isoleucine, lipids, histidine ↑HC[56]
β-Hydroxy isobutyric acid ↓
 Coronary heart disease Qi deficiencyAcetyl glutamic acid, lysine, valine, carnitine ↓Non-Qi deficiency
 Unstable anginaHLactic acid, m-I, lipid, VLDL, 3-HB, LDL ↑HC[57]
Threonine, Cr, Cho, PC/GPC, glucose, glutamine, lysine, HDL, isoleucine, leucine, valine ↓
HValine, alanine, glutamine, inosine, adenineHC[58]
 Heart failure (HF) with preserved ejection fraction (HFpEF)HAcylcarnitines, carnitine, creatinine, betaine, amino acids ↑Non-HF HC[59]
Phosphatidylcholines, lysophosphatidylcholines, sphingomyelins ↓
Medium and long-chain acylcarnitines, ketone bodies ↑HF with reduced ejection fraction (HFrEF)
 Myocardial energy expenditure elevation in heart failure patients elevationH3-Hydroxybutyrate, acetone, succinateHC[60]
 Short term prognostic of mortality in acute heart failureHLact/Chol ratioHC[61]
 Cardiovascular riskHPhenylalanine, monounsaturated, polyunsaturated fatty acidsHC[62]
Diseases of the digestive system
 Primary biliary cholangitisH3-Hydroxyisovalerate, 4-hydroxyproline, citraconate, pyruvateHC[63]
 HCV with liver fibrosisHCholine, acetoacetate, LDLHCV[64]
 Decompensated cirrhosis (non-survived)HTyrosine, phenylalanine, methionine, lactate ↑HC[65]
Choline, phosphatidylcholines, lipid (VLDL, LDL) ↓
 Liver fibrosisRPhenylalanine, N,N-dimethyl glycine, O-acetyl glycoprotein, N-acetyl glycoprotein, cholineHC[66]
 Acute-on-chronic liver failureHCreatinine, glutamine, ketone bodies, lactate, pyruvate, phenylalanine, tyrosine ↑Cirrhosis (stable liver function)[67]
HDL ↓
 Non-alcoholic steatohepatitisHAlanine (1.105), histidine (1.071), phenylalanine (1.103), tyrosine (1.183), leucine (1.222) ↑HC[68]
β-OHB (0.436), acetoacetate (0.406) ↓Simple steatosis
 Chronic cholecystitisHFormate, lactate, glutamate, acetate, 1,2-propanediol ↑HC[69]
BCAA, glutamine, histidine, alanine, tyrosine, LDL, VLDL ↓
 Inflammatory bowel diseaseMValine, lipid, glycerol, ω-3 fatty acid, lysine, phenylalanine ↑HC[70]
Acetate, glucose ↓
 Active inflammatory bowel diseasesHLeucine, isoleucine, 3-hydroxybutyric acid, N-acetylated compounds, acetoacetate, glycine, phenylalanine, lactate ↑HC[71]
Creatine, dimethyl sulfone, histidine, choline and its derivatives ↓
N-acetylated compounds, phenylalanine ↑IBD in remission
LDL, VLDL ↓
Certain infectious and parasitic diseases
 HIVHSarcosine, methyllmalonic acid (MMA), D-glucose, choline, L-aspartic acid ↑HC[72]
5β-Cholestanol, L-lysine, acetoacetate, L-threonine ↓
HCreatine (2.700), glutamine (1.468), glycine (1.204), lactate (1.262), glucose (1.334) ↑HC[73]
Alanine (0.893), choline (0.946), glutamate (0.888), glycerophosphocholine (0.909), phosphocholine (0.813), taurine (0.910), valine (0.775) ↓
 HBVHHistidine, phenylalanine, unsaturated lipid, citrate ↑HC[74]
Acetone, N-acetylglycoprotein ↓
Acetone, N-acetylglycoprotein, formate ↑HBV-LC
Unsaturated lipid, phenylalanine, glutamine, acetate, pyruvate, histidine, citrate ↓
 Viral hepatitis with schistosomiasis mansoniHLactate ↑Viral hepatitis[75]
HDL cholesterol ↓
 TuberculosisH1-Methylhistidine, acetoacetate, acetone, glutamate, glutamine, isoleucine, lactate, lysine, nicotinate, phenylalanine, pyruvate, tyrosine ↑HC[76]
Formate, alanineand glycine, glycerolphosphocholine, LDL ↓
H (children)L-valine, pyruvic acid, betaine ↓HC[77]
HKetone bodies, lactate, pyruvate ↑HC[78]
 Cerebral malariaMLipids/lipoproteins ↑HC[79]
HLactate ↑Sepsis and encephalitis[80]
Glycoprotein ↑Sepsis
Endocrine, nutritional and metabolic diseases
 Type 1 diabetes mellitus (children and adolescents)HGlucose ↑HC[81]
Triglycerides, phospholipids, cholinated phospholipids, serine, tryptophan, cysteine ↓
 Diabetic nephropathyMonVLDL/LDL (3.244), lipids (2.372), acetate (1.714), acetoacetate (1.750), unsaturated lipids (2.538) ↑HC[82]
Valine (0.586), alanine (0.483), glutamate (0.403), pyruvate (0.545), tyrosine (0.641), histidine (0.519) ↓
VLDL/LDL (1.486), lipids (1.599), unsaturated lipids (2.063) ↑Diabetes mellitus
Pyruvate (0.727), alanine (0.783), glutamate (0.543), HDL (0.684) ↓
 Polycystic ovary syndromeHAlanine (1.25), valine (1.34), leucine (1.23), threonine (2.53), lactate (1.76), acetate (1.32), histidine (3.1) ↑HC[83]
L-glutamine (0.79), proline (0.77), glutamate (0.87), α-glucose (0.77), β-glucose (0.86), 3-hydroxybutyric acid (0.91) ↓
 ObesityMAcetoacetate, acetone, succinate, carnitine, VLDL/LDL cholesterol, TMAO ↑HC[84]
Alanine, arginine, glycine, isoleucine, lysine, methionine, ornithine, serine, citrate, glucose, glycolate, lactate, pyruvate, creatine, choline, ethylene glycol ↓
 Niemann-Pick C1 disease (NPC1)HLipoprotein triacylglycerol, isoleucine ↑HC[85]
 NPC1 heterozygoteHLipoprotein triacylglycerol ↑HC
Diseases of respiratory system
 Ventilator-associated pneumoniaHGlycoproteins, phenylalanine, valine ↑Brain injuries[86]
Lipids, choline, lactate ↓
 COPDHGlycerolphosphocholine ↑HC[87]
Phenylalanine, tyrosine, alanine, valine, leucine, isoleucine, HDL ↓
 COPD with abnormal Savda syndromeHUnsaturated lipids, lactic acid, acetoacetic acid, carnitine, acetone ↑HC[88]
Amino acids, glycoproteins, LDL ↓
 AECOPDHGlycine (0.041), glutamine (0.750), alanine (0.636), proline (0.515), glutamate (0.410), mannitol (0.327), citrate (0.350), histidine (0.511), formate (0.258), creatine phosphate (0.417) ↓Stable COPD[89]
 ARDSHLipid 1 (1.633), lipid 2 (4.308), NAC (1.654), acetoacetate (1.947), creatinine (1.979), histidine (3.825), formate (7.653), lactate (1.792) ↑NARDS[90]
Mental and behavioral disorders
 DepressionHTrimethylamine oxide, glutamine, lactate ↑HC[91]
Phenylalanine, valine, alanine, glycine, leucine, citrate, choline, lipids, glucose ↓
RMyoinositol, glycerol, glycine, creatine, glutamine, glutamate, β-glucose, α-glucose, acetoacetate, 3-hydroxybutyrate, leucine, unsaturated lipidsYHTA-treated rat[92]
RLipid, lactate, alanine, N-acetyl-glycoproteins ↑HC[93]
TMAO, β-hydroxybutyric acid ↓
 Bipolar disorderHCholine, myo-inositol, L-glutamine, N-acetyl-L-phenyl alanine, amygdalin, N-acetyl-L-aspartylL-glutamic acid, α-ketoglutaric acid, lipoamideHC[94]
 Alzheimer’s diseaseMPyruvate, creatine, citrate, phenylalanine, tyrosine ↓AL mouse, HC[95]
Diseases of genitourinary system
 Acute kidney injuryHPropofol glucuronide (2.500), lactic acid (1.446), creatinine (1.556), D-glucose (1.160), Mg-EDTA2− (1.229) ↑Non-AKI[96]
Valine (0.850) ↓
 Chronic kidney injuryRAlanine (1.689, 1.395)f, glutamine (1.231, 1.031), glutamate (1.848, 1.109), citrate (2.846, 1.834), β-hydroxybutyrate (1.329, 1.857), lactate (1.164, 1.102), acetate (1.649, 1.314), acetoacetate (1.290, 1.332), formate (4.049, 1.214) ↑Sham-operated rats[97]
HAcetate, citrate, creatinine, dimethyl sulfone ↑HC[98]
Valine, lysine creatine, tyrosine, O-phosphocholine ↓
 Lupus nephritisHGlucose, lipids, lipoproteins ↑HC[99]
Lactate ↓
 Systemic lupus erythematosusGlucose ↑HC
Lactate, lipids, lipoproteins ↓
 Lupus nephritisCreatinine, lipid metabolites (including LDL/VLDL lipoproteins) ↑Systemic lupus erythematosus
Acetate ↓
 EndometriosisHValine, fucose, choline-containing metabolites, lysine/arginine ↑HC[100]
Creatinine ↓
Diseases of the nervous system
 Cerebral ischemia/reperfusionRMalonic acid, glycine, ornithine, L-lactic acid, pyruvic acid, betaine, glutamic acid, L-serine, L-alanine, succinic acid, acetic acid, citric acid, L-valineSham Control[101]
 Neuromyelitis opticaHAcetate, scylloinositolMultiple sclerosis[102]
 Secondary progressive multiple sclerosisMFatty acids, glucose, taurineRelapsing-remitting multiple sclerosis[103]
Pregnancy, childbirth and the puerperium
 PreeclampsiaHVLDL, LDL ↑HC[104]
HDL ↓
 Gestational hypertension and preeclampsiaHTriglycerides, phosphatidylcholines, etc.HC[105]
 Mild gestational diabetes mellitusHLDL/VLDL (1.21), pyruvatoxime (1.10), arginine (1.13), acetoacetate (1.28), triethanolamine (8.25) ↑Normal glucose-tolerant[106]
Leucine/isoleucine (0.75), valine (0.48), alanine (0.91), pyruvate (0.97), lactate (0.86), trimethylamine N-oxide (0.91), glycerophosphorylcholine (0.72), glycerol-3-phosphate (0.72), creatinine (0.90), 3-hydroxybutyrate ↓
 Gestational diabetesHCholesterol, lipoproteins, fatty acid, triglycerides ↑HC[107]
Other diseases
 Hemorrhagic shockRAcetate, tyrosine, lactate, lysine ↑HC[108]
 Severe burnHButyric acid, dihydrobiopterin, aldosterone, 7-dehydrocholesterol, biotin, odotyrosine, α-ketoisovaleric acid, deoxycorticosterone, 2-methoxyestrone, 2-hydroxybutyric acid, 1,3-diaminopropane, 3-methylhistidineHC[109]
α-Ketoisovaleric acid, 3-methylhistidine, β-hydroxybutyric acid ↑
 Age-related macular degenerationHCoimbra subjects: unsaturated fatty acids, acetate, creatine, dimethyl sulfone, C18 cholesterol, HDL-choline resonances boston subjects: albumin, histidine, glutamine, unsaturated fatty acidsHC[110]
 Rheumatoid arthritisH3-Hydroxyisobutyrate, acetate, nac, acetoacetate, acetone ↑HC[111]
Valine, isoleucine, lactate, alanine, creatinine, gpc, apc, histidine ↓
  1. aDiseases are classified by ICD-10 Version: 2016 issued by the World Health Organization. All of these diseases are addressed in decreasing order of the number of research papers we found. bSources of samples of the experimental groups. H represents human; M represents mouse; R represents Rat; Mon represents monkey. cThe arrows indicate the levels of metabolites or biomarkers increased (↑) or decreased (↓), compared to control groups. dHC represents healthy controls. eThe numerical values in parentheses indicated the fold-change of the increase and/or decrease of the identified metabolites between experimental groups and control groups, and all numerical calculations were based on the average concentrations of every metabolite which could be accurately found in those articles. fIn this research, the first value in parentheses was calculated from the experimental data in 4th week, and the second value was calculated from the experimental data in 8th week.

Neoplasms

As shown in Figure 3, we can see the number of papers published in last 5 years (2013–2017) on NMR-based metabolomics of blood for searching and identifying biomarkers in different tumors with different purposes.

Figure 3: Heat map showing the number of papers published in the last 5 years (2013–2017) on NMR-based metabolomics of blood for searching and identifying biomarkers in different tumors with different purposes (including screening, diagnosis, distinguishing, staging, treatment, prognosis).a‘Screening’ represents that the identified biomarkers can be used for the preliminarily screening test before invasive examination for patients with tumors. b‘Diagnosis’ represents that these biomarkers can probably be used to diagnose the neoplasms. c‘Distinguishing’ represents that the identified biomarkers have the potential to distinguish a tumor from other diseases. d‘Staging’ represents that biomarkers can be used to stage neoplasms. e‘Treatment’ represents that biomarkers can serve as targets for treatment of neoplasms. f‘Prognosis’ represents that biomarkers can be used to evaluate the prognosis of neoplasms (such as postoperative recovery). Source: PubMed databases.
Figure 3:

Heat map showing the number of papers published in the last 5 years (2013–2017) on NMR-based metabolomics of blood for searching and identifying biomarkers in different tumors with different purposes (including screening, diagnosis, distinguishing, staging, treatment, prognosis).

a‘Screening’ represents that the identified biomarkers can be used for the preliminarily screening test before invasive examination for patients with tumors. b‘Diagnosis’ represents that these biomarkers can probably be used to diagnose the neoplasms. c‘Distinguishing’ represents that the identified biomarkers have the potential to distinguish a tumor from other diseases. d‘Staging’ represents that biomarkers can be used to stage neoplasms. e‘Treatment’ represents that biomarkers can serve as targets for treatment of neoplasms. f‘Prognosis’ represents that biomarkers can be used to evaluate the prognosis of neoplasms (such as postoperative recovery). Source: PubMed databases.

Hepatocellular carcinoma (HCC) usually occurs after chronic liver diseases, especially liver cirrhosis (LC), so early HCC needs to be distinguished from cirrhosis. Thirty-two potential biomarkers were found by NMR-based metabolomics analysis of serum samples from HCC and LC patients, and these makers involved a variety of metabolic processes such as synthesis of ketone bodies, citrate cycle, phospholipid metabolism, sphingolipid metabolism, fatty acid oxidation, amino acid catabolism and bile acid metabolism. To be specific, ketone body (β-hydroxybutyrate and oxaloacetate) is higher in HCC patients (early stage to late stage) than LC patients and healthy individuals, indicating disturbance of energy metabolism in HCC patients which may be related to the rapid growth of a tumor. Other metabolisms also had more or less changes in HCC patients compared to HC, such as the higher levels of ascorbate, glycerol, isobutyrate, whereas they had lower levels of formate and tyrosine. To sum up, these potential biomarkers are vitally important in differentiating HCC from LC and have significantly diagnostic and prognostic values for HCC patients [30].

Due to the poor prognosis of pancreatic cancer (PC), early diagnosis becomes very important. NMR metabolomics analysis of serum samples from pancreatic intraepithelial neoplasia (PanIN) and pancreatic ductal adenocarcinoma (PDAC) rat models revealed that there are meaningful metabolic differences between PanIN groups and control groups (16 metabolites changed significantly), PDAC groups and control groups (20 metabolites changed significantly), PanIN groups and PDAC groups (26 metabolites changed significantly). Furthermore, dramatic metabolic changes were found in the precursor lesion, including increased levels of ketone compounds, amino acids, glycoproteins, lipoproteins and some metabolites related to mutagenicity and cancer promotion (deoxyguanosine, cytidine). All of these metabolites may be essential in pathogenesis and progression of pancreatic cancer and have the potential to be an early biomarkers for diagnosis [31]. By implanting PC cell strains Panc-1, Bxpc-3 and SW1990 in situ, Wen et al. established three different orthotopic xenograft PDAC mice models with different differentiations. The serum metabonomic profiling of these mice by using NMR demonstrated that concentrations of specific serum metabolites were different in these three models, for example, the concentrations of choline and its derivatives are higher in the Panc-1 group, whereas alcohols, amino acids and the metabolites belonged to glycolysis and glutaminolysis are lower in SW1990 group. What is more, poorly differentiated PC showed increased levels of choline and decreased levels of lactate, glutamate, glutamine and amino acids, so these metabolites be used to identify differentiation of tumors [32]. Another study demonstrated that pancreatic and periampullary adenocarcinomas patients can be distinguished from patients with benign lesions by using NMR-based metabolomics analysis, and the mainly metabolic difference between malignant and benign patients were found in the metabolic pathways of carbohydrate and amino acids (such as the arginine/proline pathway, the alanine/aspartate/glutamate pathway, the galactose metabolism pathway). Specifically, up-regulated concentrations of arginine, creatine, ornithine, proline, alanine and glutamine were found in benign samples, whereas concentrations of glutamate, urea, succinate and galactose increased in malignant samples. These metabolites have the potential to distinguish between benign and malignant PC patients [33].

The pancreatic islet β cell tumor is the most common islet cell tumor and its progression in the Rip1-Tag2 mice model includes five stages: normal (1~3 weeks), hyperplasia (4~5 weeks), angiogenic islets (6~8 weeks), tumorigenesis (9~10 weeks) and invasive carcinoma (11–14 weeks). The levels of acetate, dimethylamine, taurine, glucose and myo-inositol decreased and the level of lactate increased compared to healthy controls (HCs) by using NMR-based metabolomics analysis at 8 weeks (angiogenic islets stage), and the changes of lactate, dimethylamine and myo-inositol have statistical significance among these metabolites. In the invasive carcinoma stage, the concentrations of alanine and glutamate increased, whereas the concentration of glycine decreased. These changes are metabolic features for this stage and will aggravate in the angiogenic islets stage. Furthermore, the authors suggested higher levels of methionine, citrate, choline and lower levels of acetate, taurine and glucose could act as potential biomarkers for early pancreatic islet β cell tumors [34].

Colorectal cancer (CRC) is usually diagnosed by biopsy in current clinical practice, but Zamani et al. suggested that metabolomics analysis of patients’ serum may act as the preliminarily screening test before invasive examination. Based on 1H-NMR serum analysis of CRC patients and HC, they found 15 metabolites changed meaningfully, such as pyridoxine, orotidine and taurocholic acid which are mainly involved in bile acid biosynthesis, vitamin B6 metabolism, methane metabolism and glutathione metabolism. They considered that the differences of these metabolites are sufficient for primary screening [35]. In addition, metabolomics can also be used to determine the stage of CRC. NMR-based metabolomics analysis of 16 patients’ serum samples (eight at stage 0–I, eight at stage II–IV) revealed that concentrations of glycine, cholesterol, taurocholic acid, cholesteryl ester and deoxyinosine are higher and the level of pyridoxine is lower in patients at stage II–IV compared to patients at stage 0 and I, suggesting that these metabolic changes may be associated with the progression of CRC [36]. Therefore, NMR-based analysis of serum samples can be used for early diagnosing and staging CRC.

1H-NMR metabolic profiles of plasma can be used to distinguish lung cancer (LC) from HC and the metabolic changes can be detected in the very early stage of LC. Higher levels of glucose and lower levels of lactate, phospholipid were found in LC patients, and the combination of these metabolites can classify 92% HC and 78% of LC patients correctly, suggesting that this model can serve as a valuable tool to diagnose LC (sensitivity, 71%; specificity, 81%) [37]. Based on NMR metabolomics analysis data, eight metabolites (2-hydroxybutyrate, 2-oxoisocaproate, acetate, carnitine, 3-hydroxyisovalerate, 2-hydroxyisovalerate, glycerol and glycine) were identified to have the potential to distinguish LC at stages 1–2 from stage 3. Moreover, the non-small cell lung cancer (NSCLC) patients can be differentiated into subtypes of squamous and adenocarcinoma by using NMR metabolomics profiles [38].

NSCLC is the most malignant LC with a high lethal rate, whereas chronic obstructive pulmonary disease (COPD) is a common disease that has the potential to increase the risk of LC. In another study, nine metabolites were found that changed significantly in NSCLC patients’ serum compared to HC: higher levels of N-acetylated glycoproteins, leucine, lysine, mannose, choline, lipid (CH3(CH2)n) and lower levels of acetate, citrate and methanol by using NMR-based metabolomics analysis. Further analysis confirmed that isoleucine, acetoacetate, creatine, N-acetylated glycoproteins and glycerol can be used to distinguish NSCLC from COPD. The authors also found that the metabolism of NSCLC patients will change with the progression of NSCLC: concentrations of glycerol, NAC1, NAC2 are increased and concentrations of isoleucine, acetoacetate are decreased in the serum of advanced NSCLC patients compared to early NSCLC patients, so these metabolites have the potential to stage NSCLC. Interestingly, if NSCLC is derived from COPD, the glycerol, NAC1 and NAC2 are still positively correlated with NSCLC progression [40]. In another study, compared to HC, the NMR-based metabolic profile of NSCLC patients’ blood demonstrated 18 metabolite changes compared to HC, including the higher levels of methanol (78.81%), glutamate (37.65%), creatine (33.82%), lactate (31.53) etc. and the lower levels of choline-N(CH3)3+ (−25.30%), histidine (−20.76%), high density lipoprotein (HDL) (CH3) (−17.39%), threonine (−15.34%) etc. Moreover, another 17 metabolites were found that involved in metabolic changes in different disease stages. Besides, the authors also revealed that benign pulmonary diseases patients were different from NSCLC patients and healthy individuals in the NMR metabolic profile [39]. These discoveries may contribute to the early diagnosis, staging and prognosis of NSCLC.

By using NMR metabolomics analysis of serum from renal cell carcinoma (RCC) patients and healthy people, one study demonstrated that seven metabolites changed significantly: the levels of alanine, creatine, isoleucine, lactate, leucine, valine increased and the level of choline decreased. Interestingly, the turbulence of these metabolites in RCC patients changed to normal levels after nephrectomy, suggesting all of the changes of metabolites are specific for RCC and these metabolites may serve as a biomarker to diagnose early RCC and evaluate postoperative recovery [41]. It is difficult to distinguish benign renal tumors from RCC by routine examination, while NMR-based metabolomics are expected to solve this problem. Higher concentrations of 2-oxoisocaproate, creatine, isoleucine, glutamate, ornithine and lower concentrations of citrate, methanol, threonine glycine, histidine taurine, glutamine were observed in RCC patients’ serum samples, compared to HC, and these metabolites are closely related to glycolytic and the tricarboxylic acid cycle, amino acid and fatty acid metabolism. Moreover, orthogonal partial least squares discriminant analysis (PLS-DA) plots based on NMR metabolic data can be used to distinguish benign from pT1 RCC group (R2=0.46, Q2=0.28; AUC=0.83), benign from pT3 RCC group (R2=0.58, Q2=0.37; AUC=0.87). In conclusion, NMR metabolomics of serum can contribute to distinguish benign renal tumors from RCC and stage RCC [42].

In a recent study, the high expression of inositol 1, 4, 5 trisphosphate receptor (IP3R, a receptor regulates cellular bioenergetics) type 2 and type 3 were found in diseased tissue of breast cancer and the authors also found it is related to metabolic changes in breast cancer. By using NMR-based metabolomics analysis, they demonstrated that serum from patients with high IP3Rs expression had higher levels of lactate, lysine, alanine and lower levels of pyruvate, glucose compared to healthy individuals, which indicated that all of these metabolites could be identified as biomarkers for BC and the metabolism regulated by IP3Rs may serve as a potential therapeutic target for BC [43]. Jobard et al. used an NMR-based metabolomics approach to analyze serum from metastatic BC (MBC) and early BC (EBC), and they observed that the concentrations of acetoacetate, glycerol, pyruvate, N-acetyl glycoproteins, mannose, glutamate and phenylalanine increased, while the level of histidine decreased in MBC compared to EBC, suggesting that NMR metabolic profiles of serum can be used as an advantageous tool to diagnose breast cancer metastasis, with a sensitivity of 89.8% and a specificity of 79.3% [44]. In another study, NMR-based plasma metabolic analysis of EBC and MBC patients demonstrated that the levels of glucose, lactate, pyruvate, acetate, acetoacetate, β-hydroxybutyrate, urea, creatine, creatinine and a variety of amino acids are modulated across patient clusters, but after Bonferroni adjustment only the differences of lactate, pyruvate, glucose glutamine and lysine still have statistically significance [45]. These results suggested that BC cells could affect the whole body’s metabolism and change with the progression of BC. The metabolic changes may serve as a new approach to diagnose BC, but the specifically metabolic pathways still need further study. Hart et al. found that metabolic changes of serum due to micrometastasis in EBC patients have metastatic signatures in their NMR metabolomics profiles, and they found several biomarkers to discriminate MBC from EBC and predict a relapse of EBC. As for predicting the relapse of EBC, a random forest (RF) classification model based on preoperative metabolomics profiles of serum biomarkers was proposed as a new tool to predict the recurrence of ER-positive EBC, and the accuracy of this “RF recurrence risk score” correlated with relapse is maximized at 71.3% (sensitivity, 70.8%; specificity, 71.4%) [112].

Extensive changes of metabolites in endometrial cancer (EC) patients’ serum samples were found (32 metabolites come from NMR data) compared to HC by using NMR metabolomics analysis, and some metabolites could be used to accurately predict early-stage EC and EC patients. Based on their results, the authors suggested that 3-hydroxybutyrate, C6 (C4:1DC) or hexanoylcarnitine C14:2 or tetradecadienyl-l-carnitine could serve as the biomarkers to predict EC [46].

In recent years, NMR-based metabolomics analysis of serum has been used to distinguish prostate cancer (PCa) patients from healthy individuals. Increased levels of alanine, pyruvate, sarcosine and decreased levels of glycine are able to differentiate 90.2% of PCa patients from healthy people with 84.4% sensitivity and 92.9% specificity. Moreover, high-grade PCa and low-grade PCa can also be distinguished by the combination of biomarkers (alanine, pyruvate and glycine) with 92.5% sensitivity and 93.3% specificity [48]. In another study, by using the same NMR-based metabolomics approach, Giskeødegård et al. found the serum of PCa patients has higher levels of acylcarnitine, glucose, amino acids and glycerophospholipids (three lysophosphatidylcholines and five phosphatidylcholines), and a lower level of triglycerides compared to benign prostatic hyperplasia (BPH) patients, and the changes of these metabolites can reflect the abnormal fatty acid metabolism, choline metabolism and amino acids metabolism in PCa patients. Based on the results, the authors recommend that the changes in fatty acid (acylcarnitines), choline (glycerophospholipids) and amino acid metabolism (arginine) can be used as biomarkers to identify BPH from PCa [47]. Furthermore, another linear multivariate discriminant function analysis based on NMR data of serum analysis showed the concentrations of alanine, sarcosine, creatinine, glycine and citrate are different between BPH and PC, and these metabolites can be used to distinguish BPH from PC, with an accuracy of 88.3%. Moreover, glycine, sarcosine, alanine, creatine, xanthine and hypoxanthine were used to distinguish abnormal prostate (BPH and PCa) from healthy individuals, with an accuracy of 86.2%. It is worth mentioning that the diagnostic accuracy of NMR analysis approach is higher than the current clinical laboratory methods [49].

By comparing the serum samples from nasopharyngeal carcinoma (NPC) patients with normal controls through NMR-based metabolomics analysis, Wang et al. observed that the levels of methionine, taurine and choline-like metabolites increased, while the levels of lipids, isoleucine, LDL, trimethylamine oxidase and carbohydrates (glucose) decreased in the NPC patients compared to healthy individuals. These results were helpful to understand the turbulence of metabolic pathways in NPC and provide potential serum biomarkers to differentiate NPC patients from healthy people [50].

The NMR-based metabolomics profiling analysis of 71 serum samples from multiple myeloma (MM) and health controls demonstrated that carnitine and/or acetylcarnitine were accumulated in the MM patients at the active stage and after relapse, and it may be associated with the progression of MM, in addition, carnitine and acetylcarnitine have the potential to be biomarkers of MM [51].

Burkitt lymphoma (BL) is a highly aggressive non-Hodgkin lymphoma. At present, one study searching for BL serum biomarkers has made some progress in a mouse model by using NMR-based metabolomics analysis. Compared to the wild-type mice, the BL mice had higher levels of glutamate, glycerol and choline (AUC greater than 0.9), these three metabolites were demonstrated to have the highest diagnosis accuracy. Other metabolites also changed, such as isoleucine, leucine, pyruvate, lysine, α-ketoglutarate, betaine and so on (AUC are in the range of 0.7–0.9), these metabolites may have the potential to be diagnostic biomarkers but still require further study [52].

Diseases of the circulatory system

NMR metabolomics has been used to study the different circulatory system diseases. Hypertension is a very common disease worldwide, with one in four people in the United States having hypertension. Aiming at investigating the metabolic mechanism of essential hypertension and its Chinese medicine subtypes which are “Yin-deficiency and Yang-hyperactivity syndrome” (YDYHS) and “Yin-Yang deficiency syndrome” (YYDS), Li et al. analyzed plasma samples by 1H-NMR and GC-MS spectroscopy and identified several potential biomarkers of YDYHS and YYDS including betaine, mevalonic acid, corticosterone, β-leucine, propionic acid, methionine, D-glucose, glycine, tyrosine and malic acid (the first five metabolites were identified by NMR analysis) which indicated that abnormal glucose metabolism might be the main common pathway from YDYHS to YYDS [53]. Moreover, 1H-NMR was also used to identify blood metobolomic biomarkers of left ventricular hypertrophy (LVH) secondary to arterial hypertension, and they found that the -CH2-/-CH3 ratio which is an indicator of the mean length of the plasma aliphatic lipid chains might be a biomarker to give brand new diagnostic methods for rapid LVH detection in hypertensive patients [54].

Deidda et al. used 1H-NMR-based metabolomics to analyze 32 coronary blood samples from normal people and myocardial ischemia patients including stenotic disease (SD) and microvascular disease (“Micro”). Compared to the control group, Micro patients showed a lower content of creatine/phosphocreatine, creatinine and glucose and higher levels of 2-hydroxybutirate, alanine, leucine, isoleucine and N-acetyl groups, however, SD patients showed a lower content of 2-hydroxybutirate and higher levels of 3-hydroxybutirate and acetate. In addition, compared to Micro patients, SD patients had a higher content of 3-hydroxybutirate and acetate and lower levels of 2-hydroxybutirate, alanine, leucine and N-acetyl groups. Therefore, these metabolites mentioned above might be potential biomarkers to differentiate and diagnose different development and pathological stages of coronary artery diseases [55]. In addition, regarding coronary heart disease (CHD), one study used 1H-NMR to identify plasma metabolomics biomarkers in CHD patients and CHD Qi deficiency syndrome patients (a Chinese medicine subtype). The results showed that 25 typical metabolites including leucine, N-acetyl glycoprotein, α-glucose, β-glucose, β-hydroxy isobutyric acid, phenylalanine, acetone, HDL, glutamate, glutamine, methylamine, lysine, tyrosine, ornithine, taurine, proline, lactic acid, tryptophan, threonine, aspartic acid, valine, acetyl glutamic acid, isoleucine, lipids and histidine changed meaningfully in CHD patients in comparison with the control group, and four identifiable variables including acetyl glutamic acid, lysine, valine and carnitine had most important differences between Qi deficiency and other patients which meant these metabolites might become potential biomarkers for CHD and its Qi deficiency syndrome [56].

NMR metabolomics profiling of serum has been proved to have a promising future in identifying biomarkers of unstable angina (UA) and screening UA patients. One study found the UA patients’ serum concentrations of lactic acid, m-I, lipid, VLDL, 3-HB, and LDL had higher levels but the concentrations of threonine, Cr, Cho, PC/GPC, glucose, glutamine, lysine, HDL, Isoleucine, leucine and valine were lower compared to healthy people [57]. Furthermore, in another study Ameta et al. identified five biomarkers with extremely high sensitivity and specificity, precisely including valine, alanine, glutamine, inosine and adenine, with statistical significance between UA and HC [58].

Recent studies have pointed out many pathophysiological differences between heart failure (HF) with preserved ejection fraction (HFpEF) and HF with reduced ejection fraction (HFrEF). Zordoky et al. carried out metabolomic analysis of patients’ serum by LC-MS/MS and 1H-NMR spectroscopy, the results showed that HFpEF patients had higher levels of acylcarnitines, carnitine, creatinine, betaine and amino acids whereas lower serum concentrations of phosphatidylcholines, lysophosphatidylcholines and sphingomyelins, compared to non-HF controls. Besides, medium and long-chain acylcarnitines and ketone bodies were lower in HFrEF than HFpEF patients, which will help identify novel biomarkers for diagnosis and distinguishing [59]. Increased myocardial energy expenditure (MEE) is associated with HFrEF and is also used to predict cardiovascular mortality independently. 1H-NMR-based metabolic analysis of human serum revealed that patients with MEE elevation in HF showed obvious changes in metabolomics’ profiling results especially the level of 3-hydroxybutyrate, acetone and succinate which might be identified as potential biomarkers to diagnose MEE elevation and study energy change in HF [60]. In another study, 1H-NMR-based metabonomics analysis of plasma samples identified the lactate to cholesterol ratio as a useful, simple and independent biomarker which can predict short-term mortality and decide treating methods in acute heart failure [61].

Zhang et al. used two different analysis methods to identify potential circulating metabolic biomarkers from diastolic left ventricular function patients’ serum by NMR spectroscopy. In multivariable-adjusted regression analysis, significant changes of circulating tyrosine, high-density lipoprotein apolipoproteins, glucose+glutamine, glucose+2-aminobutyrate, glucose+2-phosphoglycerate and an unidentified molecule were found. However, glucose+glutamine, glucose+2-aminobutyrate, glucose+2-phosphoglycerate, 4-aminobutyrate, 4-hydroxybutyrate, creatinine, and phosphocholine were identified in PLS-DA, to be specific, the first three metabolites were associated with normal diastolic LV function and the last four were related to abnormal function [113].

Quantitative NMR metabolite profiling of three population-based cohorts in a large prospective demonstrated that phenylalanine, monounsaturated and polyunsaturated fatty acids could be biomarkers for cardiovascular risk and potentially result in earlier and more precise prediction of patients at high risk for cardiovascular disease [62].

Diseases of the digestive system

As an idiopathic inflammatory disease involving the ileum, rectum and colon. Inflammatory bowel disease (IBD) has two forms: ulcerative colitis (UC) and Crohn’s disease (CD). By using NMR-based metabolic profiles, Dong et al. analyzed the plasma samples from acute UC mice model (establishing by providing water containing dextran sulfate sodium), and they found eight metabolites changed significantly: reduced levels of acetate and glucose, and increased levels of valine, lipid, glycerol, ω-3 fatty acid, lysine and phenylalanine. These changes reflected lipid dysregulation and increased energy consumption in acute UC mice, and these metabolites identified by NMR metabonomics analysis could be biomarkers to assess the development of acute UC and discover novel therapeutic targets for the treatment of IBD [70]. Another study used 1H-NMR-based metabolomics to analyze serum samples from normal people and IBD patients including UC and CD. Compared to a control group, active IBD patients showed lower levels of creatine, dimethyl sulfone, histidine, choline and its derivatives, and higher levels of leucine, isoleucine, 3-hydroxybutyric acid, N-acetylated compounds, acetoacetate, glycine, phenylalanine and lactate. In addition, compared to IBD patients in remission, active IBD patients had higher contents of N-acetylated compounds and phenylalanine and lower contents of low-density lipoproteins and very low-density lipoproteins. These metabolites are the most effective biomarkers to differentiate and diagnose different stages of IBD disease [71].

In a study related to non-alcoholic steatohepatitis (NASH), Mannisto et al. found that both simple steatosis and NASH patients’ blood had higher levels of alanine, histidine, phenylalanine, tyrosine and leucine than controls by using NMR metabolic analysis. Furthermore, levels of ketone bodies including β-hydroxybutyrate and acetoacetate were significantly lower in NASH patients compared to the individuals with simple steatosis, suggesting ketone bodies could possibly be biomarkers to distinguish NASH from simple steatosis [68].

NMR-based metabolomics analysis of blood samples from primary biliary cholangitis (PBC) patients indicated that 33 metabolites changed significantly compared to HC, and four metabolites among them (3-hydroxyisovalerate, 4-hydroxyproline, citraconate and pyruvate) were identified as potential biomarkers for PBC diagnosis. Moreover, 4-hydroxyproline is associated with collagen stability and reflects the liver fibrosis stage or collagen content, and 3-hydroxyisovalerate (3-HIVA) is a side product of increased ketogenesis. Furthermore, their results showed this diagnostic model with the combination of four biomarkers had high clinical value for the diagnosis of PBC [63].

Chronic cholecystitis (CC) is a benign disease caused by recurrent acute or subacute cholecystitis, whereas xanthogranulomatous cholecystitis (XGC) is a devastating inflammatory disease which is based on CC and accompanied by the formation of yellow granulomas. Although XGC is benign, it is very difficult to differentiate from gallbladder cancer (GBC). Therefore, one study found obvious differences in serum metabolic characterizations among healthy individuals, CC, XGC and GBC patients by using 1H-NMR-based metabonomics analysis. Importantly, they emphasized that higher levels of lactate and pyruvate as well as lower levels of glucose, some amino acids and low density lipoprotein/very low density lipoprotein (LDL/VLDL) were found, and these metabolites from different metabolic pathways could be identified as the biomarkers of the GBC to discriminate from other gallbladder diseases [114]. In another study, Sharma et al. used the NMR metabolomics approach to analyze serum samples from CC patients and HC, and they found increased levels of formate, lactate, 1,2-propanediol, glutamate and acetate whereas they found decreased levels of BCAA, glutamine, histidine, alanine, tyrosine, LDL and VLDL in CC patients. Furthermore, among these metabolites, lactate, glutamine, formate, 1,2-propanediol, histidine and acetate were suggested to be identified as biomarkers to early diagnosis and monitoring CC progression [69].

1H-NMR metabolomics analysis was used in a study to detect plasma samples from the non-surviving patients with decompensated cirrhosis (DC), stable cirrhosis patients and HC, the results showed elevated levels of lactate, tyrosine, phenylalanine and methionine and decreased levels of choline, phosphatidylcholines, lipid (VLDL, LDL) in non-surviving DC patients. Moreover, their study provided several potential biomarkers of non-surviving DC and a new method to effectively predict mortality in DC and to distinguish non-surviving from surviving patients, even had high reference values for liver transplantation [65]. In another research, Wei et al. used 1H-NMR-based metabolomics approach to profile serum samples from a liver fibrosis rat model (established by chronic and low-dose exposure of thioacetamide), and they identified five metabolites precisely including phenylalanine, N,N-dimethyl glycine, O-acetyl glycoprotein, N-acetyl glycoprotein and choline which could become biomarkers of liver fibrosis [66].

Any healing process of liver injury is accompanied by liver fibrosis, and liver fibrosis may develop into cirrhosis. One study used the NMR-based metabolomics method to analyze serum samples obtained from HCV patients with cirrhosis and HCV patients without detectable fibrosis. Based on the PLS-DA test and ROC analysis, they found choline, acetoacetate and LDL1 can be identified to be significant biomarkers to distinguish two groups of HCV patients and predict advanced liver fibrosis [64].

Acute-on-chronic liver failure (ACLF) is an acute or sub-acute decompensation of liver function resulting from chronic liver disease. Compared with compensated or decompensated stable cirrhosis, the NMR-based metabolic signature of ACLF demonstrated reduced levels of high-density lipids, and increased levels of creatinine, glutamine, ketone bodies, lactate, pyruvate, phenylalanine and tyrosine. In addition, all of these potential biomarkers associated with variation of lipid metabolism, amino-acid, lactate and urea metabolism can effectively distinguish chronic liver failure from ACLF [67].

Chemotherapy is one of the most effective methods to treat cancer, but chemotherapy-induced gastrointestinal toxicity caused various adverse reactions and side effects. One study used the NMR-based metabolomics approach to analyze blood and urine samples from a chemotherapy rat model (established by injecting 5-fluorouracil, oxaliplatin and irinotecan), and they found increased levels of PUFAs and N(CH3)3 moieties, and reduced levels of tryptophan, glutamine, arginine compared to controls. Moreover, these changes indicated activation of intestinal inflammatory processes, and these metabolites could be identified as biomarkers helping to optimize treatment efficacy of chemotherapy and reduce adverse reactions [115].

Certain infectious and parasitic diseases

Chronic HCV carriers are at risk of chronic hepatitis, liver fibrosis, HCC and many other serious diseases, and one study used the NMR metabolic approach to analyze blood samples from three groups of HCV patients (divided by liver histopathology characteristics): advanced fibrosis (METAVIR F3-4), necroinflammation (METAVIR A2-3) and moderate to severe steatosis (≥33%). In addition, their results demonstrated that 21 metabolites were in connection with advanced fibrosis, 17 metabolites were related to necroinflammation and 16 metabolites were associated with moderate to severe steatosis, all of these metabolites may be significant potential biomarkers of these diseases in patients with chronic HCV [116]. Hepatitis B virus (HBV) infection is one of the main causes of LC. By using NMR-based metabolomics analysis, compared to healthy people, a study observed that patients with HBV showed higher levels of histidine, phenylalanine, unsaturated lipid, citrate in serum and lower levels of N-acetylglycoprotein, acetone and especially serum histidine which was considered as a potential biomarker for patients with HBV. In addition, compared to HBV patients, serum from HBV-LC ones had higher levels of unsaturated lipid, phenylalanine, glutamine, acetate, pyruvate, histidine and citrate, and lower levels of formate, acetone and N-acetylglycoprotein. Among these metabolites, glutamine, formate, pyruvate and acetate were were suggested as potential biomarkers to monitor the progression from HBV to HBC-LC [74]. Schistosomiasis mansoni coinfection with HBV or HCV can bring a series of liver diseases which is different from HBV or HCV monoinfection, and the periportal fibrosis diagnosis is very challenging under this situation. In one study, by using the NMR blood analysis approach, Gouvei et al. found that the coinfected group had higher levels of lactate, while the monoinfected group had higher levels of HDL cholesterol, and these metabolites can be identified as biomarkers to distinguish these two groups’ patients. They provided a metabolomics model based on changed metabolites to diagnose periportal fibrosis of coinfected patients with schistosomiasis and viral hepatitis (HBV or HCV), which is very beneficial for clinical diagnosis [75].

AIDS remains a devastating infectious disease caused by infection by the HIV virus. The NMR-based metabolic signature of plasma from normal controls, treatment naïve HIV/AIDS patients and patients receiving anti-retroviral therapy (ART) showed: the levels of sarcosine, methyllmalonic acid (MMA), D-glucose, choline, L-aspartic acid increased in HIV/AIDS patients compared to controls, and they decreased in patients receiving ART compared to treatment naïve patients but still higher than controls; the levels of 5β-cholestanol, L-lysine, acetoacetate and L-threonine were lower in HIV/AIDS patients than controls, and higher in patients receiving ART than treatment naïve ones but did not reach the levels found in controls. The authors suggested that plasma sarcosine and choline could be used as biomarkers of HIV/AIDS infection, providing a new clue of mechanism research and treatment [72]. Another study used 1H-NMR-based metabolomics to analyze serum samples from the same three groups mentioned in last study. The results showed that, compared to control group, HIV-infected patients had a lower content of alanine, choline, glutamate, glycerophosphocholine, phosphocholine, taurine, valine and a higher content of creatine, glutamine, glycine, lactate, glucose, which could be identified as potential biomarkers with further research. Moreover, they also found the reduced level of alanine and the enhanced level of glutamine and glucose were directly related to the CD4 count [73].

By using the NMR-based metabolomics’ approach, one study analyzed serum samples obtained from early- and late-stage breast cancer patients with tuberculosis (TB). Compared to HC, the results showed reduced levels of low-density lipoproteins, formate, glycerolphosphocholine, alanine and glycine, and increased levels of acetone, acetoacetate, isoleucine, nicotinate, lysine, glutamate, 1-methylhistidine, glutamine, lactate, phenylalanine, pyruvate and tyrosine. The authors suggested that NMR-based metabolomics is a valuable tool for TB diagnosis as well as monitoring, and the detailed measurement of biomarkers can also improve our understanding of the disease mechanism [76]. Another study focusing on pediatric TB indicated the metabolic changes in plasma between TB children and non-TB controls (RTIs [respiratory tract infection] and healthy children) by using NMR-based metabolomics. They found metabolic differences in 17 compounds, among them, the levels of L-valine, pyruvic acid and betaine decreased in the active TB children group, which could potentially be validated as biomarkers for diagnosis of pediatric TB [77]. Zhou et al. used 1H-NMR-based metabolomics to identify plasma biomarkers in TB, diabetes, malignancy and community-acquired pneumonia (CAP) patients. Their results showed the levels of ketone bodies, lactate and pyruvate in plasma of TB, malignancy and CAP patients were higher than HC, but the levels of TB patients were lower than the other two diseases. In addition, they identified 26 metabolites that can be used to distinguish diabetes from TB patients, 27 metabolites to separate CAP from TB patients and 24 metabolites to discriminate malignancy from TB patients. Therefore, the authors claimed that these potential biomarkers can be used to differentiate TB from the other three diseases and reflect Mycobacterium tuberculosis infection [78].

Ghosh et al. used 1H-NMR-based metabolomics method to analyze serum samples from a mouse cerebral malaria (CM) model, and they found the CM animal started to have metabolic changes (increased levels of serum lipids/lipoproteins) on the 4th day after infection, although it is not distinguishable from non-CM by physical observation at this time. In addition, all of these metabolites could become potential biomarkers of CM for early diagnosis [79]. Another study used the same method to analyze plasma samples obtained from patients of mild malaria (MM), severe non-cerebral malaria (SNCM), CM, and patients suffering from cerebral malaria with multiple organ dysfunctions (CMMOD). Combining with the OPLSDA method, the authors found these four groups showed significantly elevated levels of lipoprotein and lactate compared to uninfected controls. Moreover, this study indicated that the concentration of lactate is higher in the MM, SNCM, CM groups compared to sepsis and encephalitis patients which are symptomatically similar diseases of CM. In addition, MM, SNCM and CM groups showed higher concentration of glycoprotein compared to the sepsis group, but CMMOD had lower concentration of that compared to any other groups. Therefore, all of these metabolites could be identified to potential biomarkers to distinguish CM from its symptomatically similar diseases and separate different malaria subtypes efficiently [80].

Endocrine, nutritional and metabolic diseases

As a common pediatric disease, the incidence of type 1 diabetes mellitus (T1DM) is increasing year by year in many countries. 1H-NMR-based metabolic profiling of plasma from T1DM children and adolescents were compared to healthy control, and the results showed significantly lower concentrations of lipids (triglycerides, phospholipids and cholinated phospholipids) and slightly lower levels of amino acids (serine, tryptophan and cysteine) and expected higher levels of glucose which indicated disorders in metabolic pathways (choline reduction, increased gluconeogenesis and glomerular hyperfiltration) in T1DM patients [81]. Diabetic nephropathy (DN) is a very serious and common complication of diabetes mellitus. Liu et al. performed 1H-NMR-based metabolomics to analyze serum acquired from a nonhuman primate model of DN. In comparison with HC, DN monkeys had not only higher concentrations of VLDL/LDL, lipids, acetate, acetoacetate and unsaturated lipids, but also lower concentrations of alanine, valine, glutamate, pyruvate, tyrosine and histidine. Furthermore, DN monkeys demonstrated increased concentrations of VLDL/LDL, lipids and unsaturated lipids, as well as decreased levels of pyruvate, alanine, glutamate and HDL compared with the diabetes mellitus group. All of these metabolites mentioned above might become potential biomarkers of DN and provide new clues in the pathogenesis of DN [82].

Focusing on Indian women with polycystic ovary syndrome (PCOS) particularly, a study used 1H-NMR coupled with the metabolomics’ approach to analyze serum samples from Indian PCOS patients, and the results showed that significant changes were found in 12 metabolites which might be identified as potential biomarkers. Compared with the controls, five amino acids (alanine, valine, leucine, histidine and threonine), lactate and acetate increased significantly, whereas another three amino acids (L-glutamine, proline and glutamate), glucose and 3-hydroxybutyric acid obviously decreased [83].

In another study, a time-related 1H-NMR-based metabonomic profiling of plasma from hyperliqidemic hamsters found discriminating changes of 40 endogenous metabolites, and the analysis of these potential hyperlipidemia biomarkers demonstrated that obvious lipid and amino acid metabolism disturbances, inflammation and oxidative stress, as well as intestinal microflora metabolites changes happened after cholesterol overload [117]. Won et al. carried out gender-specific 1H-NMR-based metabonomic profiling of obesity by using serum from leptin-deficient ob/ob mice, they identified 22 metabolites as potential biomarkers which are involved in amino acid metabolism, tricarbocylic acid cycle and glucose metabolism, lipid metabolism, creatine metabolism and intestinal microflora metabolites. Particularly, their results showed that metabolism of obese male mice was associated with insulin signaling, however, in the obese female mice it was specifically related to lipid metabolism [84]. In regard to the 1H-NMR-based metabolic profiles of blood serum samples from patients with nodular thyroid diseases, decreased alanine and increased glucose are only slightly changed, but have no statistical significance, so it is hard to identify these metabolites as biomarkers right now [118]. Niemann-Pick type C (NPC) disease is one kind of sphingolopodosis due to a neurodegenerative lysosomal storage disorder. High-resolution 1H-NMR spectroscopy metabolomic profiling coupled with advanced multivariate analysis methods revealed significant difference in the plasma of Niemann-Pick C1 (NPC1) patients and heterozygous carriers compared with HC. Lipoprotein triacylglycerol and isoleucine changed significantly between NPC1 and HC, and heterozygote samples also showed higher levels of lipoprotein triacylglycerol vs. HC. Therefore, these identified metabolites could represent potential biomarkers for NPC1 [85].

Diseases of the respiratory system

Ventilator-associated pneumonia (VAP) is an important type of hospital-acquired pneumonia (HAP). Antcliffe et al. used NMR to analyze serum samples obtained from pneumonia patients and brain injury patients (seven of them went on to develop VAP after ventilating treatment). They found that the levels of formate, glycoproteins and phenylalanine in pneumonia patients’ serum changed significantly, whereas the concentrations of phospholipids, glutamine and alanine in brain injuries patients’ samples altered meaningfully, compared to each other. Moreover, as for the comparison between simple brain injury patients and VAP ones, they found the increased levels of lipids, choline and lactate in brain injured ones, as well as the elevated levels of glycoproteins, phenylalanine and valine in VAP ones, but when univariate comparisons were used in spectral data all changes of these metabolites had no statical significance. Therefore, they successfully identified potential biomarkers to distinguish pneumonia from brain injuries, but markers for separating brain injuries from VAP still need further research [86].

COPD is a common chronic disease with airflow obstruction, which can further cause pulmonary heart disease and respiratory failure. One study based on 1H-NMR spectroscopy analyzed serum obtained from COPD patients, and they found decreased levels of phenylalanine, tyrosine, alanine, valine, leucine, isoleucine, HDL and elevated level of glycerolphosphocholine in COPD patients compared to HC. What is more, all of these markers they identified were beneficial to the early diagnosis and treatment of COPD [87]. In another study, Xu et al. used the same approach to analyze plasma collected from COPD patients with abnormal and non-abnormal Savda syndrome (traditional Uyghur medicine subtypes) and HC. Performing NMR data with OPLS-DA, they found the concentrations of lactic acid, carnitine, acetone, acetoacetic acid and unsaturated lipids were higher, and the concentrations of glycoprotein, LDL and amino acids were lower in COPD patients with abnormal Savda syndrome compared with the controls. Furthermore, these changes may serve as the potential diagnostic biomarkers of COPD with abnormal Savda syndrome [88].

Fortis et al. attempted to use the NMR-metabolomics method to distinguish severe acute exacerbations of COPD (AECOPD) from other coexisting diseases such as heart failure and pneumonia, which often happen together with AECOPD and have similar symptoms. Compared to stable COPD patients, AECOPD ones had lower levels of glycine, glutamine, proline, citrate, histidine, formate, creatine, phosphate, alanine and mannitol, and the first six metabolites changed significantly statistically, which can potentially be used as markers to separate AECOPD from COPD patients. However, the various respiratory failure groups could not be discriminated by this approach, they did not find unique metabolic changes of AECOPD to differentiate it from heart failure and pneumonia [89].

1H-NMR-based metabolic profiling of serum from acute respiratory distress syndrome (ARDS) patients was compared to non-ARDS (NARDS) controls, and the results showed the significantly enhanced level of lipid 1, lipid 2, NAC, acetoacetate, creatinine, histidine, formate and lactate. Moreover, the authors suggested that lipids and lipoproteins are especially important biomarkers in the discrimination between ARDS and NARDS, and these findings also provided a new diagnostic and treatment method for ARDS [90].

Mental and behavioral disorders

Depression is one of the most common and serious mental disorders. One study used 1H-NMR-based metabolomics approach to analyze serum from depressed patients and healthy volunteers, the results showed the concentrations of trimethylamine oxide, glutamine and lactate were up-regulated, as well as the levels of phenylalanine, valine, alanine, glycine, leucine, citrate, choline, lipids and glucose were down-regulated significantly in depressed patients compared with HC [91]. Another study used the same approach to analyze plasma from a chronic unpredictable mild stress (CUMS) rat model, and they identified 13 biomarkers of CUMS-induced depression including myoinositol, glycerol, glycine, creatine, glutamine, glutamate, β-glucose, α-glucose, acetoacetate, 3-hydroxybutyrate, leucine and unsaturated lipids (L7, L9), which means abnormal changes occurred in amino acid metabolism, glycometabolism, lipid metabolism and energy metabolism in the CUMS rat mode. Expect glycine, glutamine, glucose, unsaturated lipids (L7, L9) and 3-hydroxybutyrate, other biomarkers are reported for the first time in this study [92]. Three different studies tested three different drugs’ anti-depressant effect by using 1H-NMR-based metabolomics approach. Wu et al. found the concentrations of all the 13 identified biomarkers tended to return to normal after using the total alkaloid of Corydalis rhizome in a CUMS rat model [92]. In addition, Liu et al. found that alanine, choline, trimethylamine oxide, glutamine, lactate and glucose were restored to standard concentrations after Xiaoyaosan treatment in a depressed patient [91]. Moreover, Tian et al. found that trimetlylamine oxide (TMAO) and β-hydroxybutyric acid (β-HB) decreased in serum, whereas lipid, lactate, alanine and N-acetyl-glycoproteins increased after genipin treatment in a CUMS rat model [93]. All of these markers may be beneficial to optimize the diagnosis and even the prediction of depression and to assess the efficacy of different drugs in clinical practice.

1H-NMR, 1H-NMR T2-edited and 2D-NMR metabolomic profiling in serum samples from bipolar disorder patients showed that serum lipids, lipid metabolism-related molecules (choline, myo-inositol), some amino acids (N-acetyl-L-phenyl alanine, N-acetyl-L-aspartyl-L-glutamic acid, L-glutamine), amygdalin, α-ketoglutaric acid and lipoamide were significantly changed compared to HC, and these metabolites may be identified as biomarkers of bipolar disorders [94].

Kim et al. used NMR-based metabolomics analysis approach to analyze samples from an Alzheimer’s-like (AL) model mice, compared with a control group and gallate-treated AL mice, both whole blood and plasma analysis showed that several metabolites significantly decreased, especially pyruvate and creatine, which could be confirmed as novel plasma biomarkers for AD in further research [95].

Diseases of the genitourinary system

Several blood biomarkers for genitourinary system disorders were identified by using a NMR-based metabolomics’ approach. Acute kidney injury (AKI) is a common postoperative complication after cardiac surgery with cardiopulmonary bypass. The NMR-based metabolomics analysis of 85 patients’ plasma samples found that the concentrations of propofol glucuronide, lactic acid, creatinine, D-glucose and Mg-EDTA2− increased and the level of valine decreased in AKI patients compared to HC, and all of them had statistically significant differences. Furthermore, the combination of the plasma metabolites Mg-EDTA2−, lactate and creatinine showed the potential to predict the prognostic of AKI [96]. In the chronic kidney injury (CKD), an NMR-based profiling of plasma metabolites revealed that compared to sham-operated rats the levels of alanine, glutamine, glutamate and organic anions (citrate, b-hydroxybutyrate, lactate, acetate, acetoacetate and formate) significantly increased in CKD rats (5/6 nephrectomy [5/6 Nx]), suggesting that these metabolites are related to the metabolic changes in CKD and had the potential to be biomarkers of CKD [97]. In another recent study, Mika et al. used the same analysis approach to analyze 55 human plasma samples from CKD patients and healthy people, and they found that the levels of acetate, citrate, creatinine and dimethyl sulfone were significantly higher in CKD patients than in controls, whereas concentrations of valine, lysine creatine, tyrosine and O-phosphocholine were significantly lower, and the content of these plasma metabolites would change in the progress of CKD. Moreover, the MetPA analysis demonstrated five potential target pathways and 14 metabolites may serve as the potential therapeutic targets and potential biomarkers of CKD [98].

Lupus nephritis (LN) is a kidney disease caused by systemic lupus erythematosus (SLE). SLE, LN and HC can be identified by the NMR metabolomics’ analysis of serum samples. Compared to HC, SLE and LN samples both had higher levels of glucose and lower level of lactate, besides, SLE samples had lower levels of lipids and lipoproteins, LN samples had higher levels of lipids and lipoproteins. In addition, LN samples had increased levels of creatinine, lipid metabolites (including LDL/VLDL lipoproteins) and decreased level of acetate compared to SLE. In conclusion, these metabolic changes may serve as potential metabolic markers to provide a novel way to distinguish SLE from LN, and also have the potential to be used in diagnosis and the clinical treatment of LN [99].

Endometriosis is an estrogen-dependent chronic condition, usually occurring in women at reproductive age. The NMR-based plasma metabolomics analysis of samples collected at the follicular phase and luteal phase revealed that the concentrations of valine, fucose, choline-containing metabolites and lysine/arginine increased, and the concentration of creatinine decreased compared to control, and these changed metabolites may become potential early biomarkers of endometriosis [100].

Diseases of the nervous system

1H-NMR spectroscopy-based dynamic metabolomic profile on rats with focal cerebral ischemia/reperfusion (I/R) injury demonstrated metabolite changes in the early stages of I/R injury, and the study showed that 13 metabolites had obvious changed especially malonic acid and glycine which could potentially be identified as biomarkers [101].

In order to discriminate relapsing-remitting (RR) from secondary progressive (SP) disease stages in multiple sclerosis (MS) patients, Dickens et al. combined NMR metabolomics with PLS-DA of serum from SPMS and RRMS patients, and they found a significant type II biomarker for the RR to SP transition in MS patients which is beneficial in classifying patients for treatment [21]. In another mice model study, they found fatty acids, glucose and taurine were the key metabolites in distinguishing RRMS from SPMS. In addition, it is worth mentioning that fatty acids and glucose were previously reported as biomarkers in separating RRMS from SPMS in a human study [103]. As for separating neuromyelitis optica (NMO) from MS, the 1H-NMR-assisted metabolomic serum analysis identified acetate and scylloinositol as hopeful serum biomarkers of NMO and MS, separately. They can distinguish MS from NMO patients by using these two metabolites at the same time, with a sensitivity of 94.3% and a specificity of 90.2% [102].

Pregnancy, childbirth and the puerperium

NMR metabolic analysis of serum samples from pregnant women with preeclampsia (PE), pregnant controls (PC) and non-pregnant controls (NP) revealed that there are differences in metabolites between PE and PC (nine metabolites), PC and NP (15 metabolites). Moreover, total serum lipid increased in PE compared to NP, and were elevated higher in PC than PE. In addition, compared to the PC groups, PE had a lower level of HDL and higher levels of VLDL and LDL. Furthermore, all the metabolites mentioned above could become biomarkers for the early diagnosis of preeclampsia [104]. In another study, the possibility of metabolomics to predict gestational hypertension and preeclampsia from first trimester serum and urine samples were evaluated and the metabolic changes associated with these diseases were clarified by using NMR spectroscopy. NMR metabolomics’ profiles of serum predicted gestational hypertension and preeclampsia at 33% and 15% sensitivity in their prediction models, separately. In addition, they also found the levels of lipids especially triglycerides elevated in both diseases compared to healthy people, and an atherogenic lipid profile elucidated phosphatidylcholines could be the most important biomarkers for prediction [105].

Mild gestational diabetes mellitus (GDM) is a common diabetes mellitus arising in pregnancy. NMR-assisted metabolomic analysis of plasma from GDM patients showed significantly higher levels of fasting blood glucose, insulin, and homeostasis model assessment-2 for insulin resistance than those in normal pregnant women, which indicated significant perturbations in glucose, fatty acid and amino acid metabolism, as well as activation of inflammatory response in mild GDM patients [106]. By NMR metabolomics analyzing maternal blood, increased levels of cholesterol, lipoproteins, fatty acids, triglycerides and several significant changes of metabolites including glucose, amino acids, betaine, urea, creatine and other metabolites related to gut microflora were found which could be used as biomarkers to predict gestational diabetes [107].

Other diseases

Hemorrhagic shock (HS) often occurs due to the fast, massive blood loss without timely replenishment, and the following reperfusion process during medical treatment causes ischemia/reperfusion (I/R) injury which can lead to poor prognosis. One study used SD rats and Arctic ground squirrels (AGS) to establish HS and I/R model by withdrawing blood and then reperfusing with Ringers, performing 1H-NMR-based metabolomics’ analysis to detect plasma samples taken instantly before hemorrhage and 3 h after reperfusion from rat model and AGS (as a negative control because their ability to resist I/R without decreased body temperature as a hibernator). Compared to negative controls, the SD group showed increased levels of acetate, tyrosine, lactate and lysine, these metabolites were identified as biomarkers related to poor outcome after HS in rats and they were absent in AGS [108].

The metabolic signature of serum samples from severe burn patients was characterized by NMR analysis. Zhang et al. reported a set of biomarkers which offered a new approach to improve the diagnosis and reduce the mortality of severe burn patients, including 12 metabolites: butyric acid, dihydrobiopterin, aldosterone, 7-dehydrocholesterol, biotin, odotyrosine, α-ketoisovaleric acid, deoxycorticosterone, 2-methoxyestrone, 2-hydroxybutyric acid, 1,3-diaminopropane and 3-methylhistidine. Among these biomarkers, significantly increased levels of α-ketoisovaleric acid, 3-methyl histidine and β-hydroxybutyric acid were observed in the early stages of burn injury. Moreover, α-ketoisovaleric acid is a new biomarker related to mitochondrial dysfunction, 3-methyl histidine level is associated with skeleton catabolism after severe burns and β-hydroxybutyric acid reflects the increasing level of ketogenic metabolism [109].

Age-related macular degeneration (AMD) is a common eyes condition and causes vision loss among people in their 50s and older, and the early and intermediate stages of AMD usually do not have obvious symptoms. One study used the NMR-based metabolomics’ method to analyze blood samples from 396 patients (Coimbra Group and Boston Group) with AMD, and researchers found the variation of some amino acids, specific lipid moieties, dimethyl sulfone and organic acids in different stages of AMD. Comparison between the different groups showed that nutrition and lifestyle affected the metabolites’ variations of AMD. Coimbra subjects (unsaturated fatty acids, acetate, creatine, dimethyl sulfone, C18 cholesterol and HDL-choline resonances) and Boston subjects (albumin, histidine, glutamine and also unsaturated fatty acids) had their different metabolic changes in early AMD stage. Furthermore, all of these potential biomarkers could possibly be used to distinguish different stages of AMD [110].

By analyzing serum samples from rheumatoid arthritis (RA) patients with NMR-based metabolic profiles, 12 potential biomarkers were found and these metabolites involved a variety of the networks of the synthesis and degradation of ketone bodies, valine, leucine and isoleucine degradation, glycolysis or gluconeogenesis, propanoate metabolism, pyruvate metabolism and glycerophospholipid metabolism. Notably, levels of valine, isoleucine, lactate, alanine, creatinine, GPC, APC and histidine were lower in RA patients, whereas concentrations of 3-hydroxyisobutyrate, acetate, NAC, acetoacetate and acetone were higher compared to controls. Furthermore, these potential biomarkers have diagnostic and prognostic value for RA [111].

Discussion

In order to search for human disease biomarkers, numerous metabolomics research focuses on identifying the changes of small-molecule metabolites in samples, but the determination of metabolites’ changes can be affected by so many factors including individual differences, interactions with other small molecules or the environment in the human body, methods of sample collection, preparation and storage, experiment procedures, data analysis, data description, database standardization and so on [119], [120], [121]. Experimental data will present great differences under the influence of these factors, which is absolutely not conducive to creating comparable data in the study of metabolomics. Therefore, the standardizations of the whole process especially metabolomics experiments, data analyses and reports are desperately needed to make data from different laboratories in the metabolomics’ societies comparable, which helps to find more accurate disease biomarkers in a standard condition and allow data to be efficiently applied, shared and reused. The standardization of experimental protocols allow other laboratories to verify the data through standard experimental procedures, at the same time the results obtained by standard procedures can be shared and compared among different laboratories and groups. In addition, different laboratories can share data to achieve larger sample size and get more comprehensive metabolite information on change. Moreover, disease-related biomarkers obtained in standard processes have higher application values and can be better translated into the clinic [120]. Based on the need for standardization of metabolomics’ studies, the metabolomics standards initiative (MSI) was published in 2007 and mainly focused on recognizing reporting standards which aim to provide a common mechanism for describing the metabolomics studies so that the data can be evaluated, replicated and published in a public repository [122]. In addition, as a part of MSI, the Chemical Analysis Working Group is generally responsible for identifying, developing and disseminating consistent descriptions related to the chemical analysis practices of metabolomics [123]. Besides, MetaboLights is the first general-purpose, open-access repository for metabolomics’ studies which is cross-species and cross-techniques, and it adheres to the standards of MSI for metadata reporting. The MetaboLights repository includes the structures of metabolites and their reference spectra as well as their biological functions, locations and concentrations, what is more, it spans over eight different species and covers a variety of techniques [124]. Although MSI and MetaboLights have made great contributions to the standardization in terms of data acquisition, data processing, data reporting and database establishment in the last decade, a lot of work about standardization still needs to be done in the future to help create comparable data and make the metabolomics research more scientific and convenient.

Metabolomics has been widely used to search for disease-related biomarkers in recent years, however, limitations and challenges still exist in many aspects of metabolomics studies. As mentioned in past reviews, different analytical techniques (NMR [125] or MS [126]) have their own weakness and limitation in sensitivity, reproducibility or equipment costs. Apart from this, no matter which analytical method is used, metabolomics research is faced with the following common problems [119], [127]: (1) The composition of metabolites in samples is very complicated, so the accurate measurement of the entire metabolome in a single analysis is currently impossible; (2) The analysis of metabolomics produces massive amounts of data, while it is difficult to comprehensively identify all the changed metabolites with the available metabolite databases. In addition, the functions and mechanisms of the identified metabolite changes are also difficult to interpret; (3) The metabolome is very sensitive to many internal and external factors, so some minor differences may affect the metabolic profile of samples and cause challenges to data interpretation; (4) Compared with the large amounts of laboratory data, the number of patients’ samples is not enough, and this imbalance makes the translation of these laboratory achievements into clinical applications difficult; (5) Standardization is also a big problem, as discussed in the previous paragraph. Therefore, the continued development of metabolomics’ research techniques will help overcome the above limitations, and the standardization of metabolomics’ studies will make results from different studies more comparable and aid in the identification of biomarkers.

Conclusions

In this paper, we review the application of NMR-based metabolomics of human blood samples for searching and identifying biomarkers of diseases over the last 5 years. Aiding in the diagnosis and treatment, NMR metabolomics significantly contributes to the validation and quantification of biomarkers to numerous human diseases including neoplasms, digestive, nervous, respiratory, mental, infectious, parasitic diseases and so on. Obviously, blood is one of the most classic and common samples, but recent researches placed much emphasis upon other body fluid samples such as urine, saliva, sweat and tears due to the recognized significance of developing and optimizing non-invasive strategies. In fact, there are still many problems that need to be solved in sample preparation, experimental and statistical techniques, as well as the validation, quantification and clinical application of biomarkers. Furthermore, overcoming the obstacles mentioned above will absolutely trigger much more developments in NMR metabolomics’ analysis and solve more clinical challenges.


Corresponding author: Wei Jiang, PhD, Molecular Medicine Research Center, West China Hospital, Sichuan University, Chengdu 610041, Sichuan, P.R. China, Phone: +86-028-85164103 (Office), +85164092 (Lab)

  1. Author contributions: All the authors have accepted responsibility for the entire content of this submitted manuscript and approved submission.

  2. Research funding: This work was supported by the National Natural Science Foundation of China (81670249, 31271226 and 31071001 to Dr. W Jiang).

  3. Employment or leadership: None declared.

  4. Honorarium: None declared.

  5. Competing interests: The funding organization(s) played no role in the study design; in the collection, analysis, and interpretation of data; in the writing of the report; or in the decision to submit the report for publication.

References

1. Wishart DS. Metabolomics: the principles and potential applications to transplantation. Am J Transplant 2005;5:2814–20.10.1111/j.1600-6143.2005.01119.xSearch in Google Scholar PubMed

2. Stringer KA, McKay RT, Karnovsky A, Quemerais B, Lacy P. Metabolomics and its application to acute lung diseases. Front Immunol 2016;7:44.10.3389/fimmu.2016.00044Search in Google Scholar PubMed PubMed Central

3. Wang M, Wang F, Wang Y, Ma X, Zhao M, Zhao C. Metabonomics study of the therapeutic mechanism of Gynostemma pentaphyllum and atorvastatin for hyperlipidemia in rats. PLoS One 2013;8:e78731.10.1371/journal.pone.0078731Search in Google Scholar PubMed PubMed Central

4. Chan AW, Mercier P, Schiller D, Bailey R, Robbins S, Eurich DT, et al. (1)H-NMR urinary metabolomic profiling for diagnosis of gastric cancer. Br J Cancer 2016;114:59–62.10.1038/bjc.2015.414Search in Google Scholar PubMed PubMed Central

5. Hunter WG, Kelly JP, McGarrah RW, Kraus WE, Shah SH. Metabolic dysfunction in heart failure: diagnostic, prognostic, and pathophysiologic insights from metabolomic profiling. Curr Heart Fail Rep 2016;13:119–31.10.1007/s11897-016-0289-5Search in Google Scholar PubMed PubMed Central

6. Liang Q, Liu H, Zhang T, Jiang Y, Xing H, Zhang A-H. Discovery of serum metabolites for diagnosis of progression of mild cognitive impairment to Alzheimer’s disease using an optimized metabolomics method. RSC Adv 2016;6:3586–91.10.1039/C5RA19349DSearch in Google Scholar

7. Mikkonen JJ, Singh SP, Herrala M, Lappalainen R, Myllymaa S, Kullaa AM. Salivary metabolomics in the diagnosis of oral cancer and periodontal diseases. J Periodontal Res 2016;51:431–7.10.1111/jre.12327Search in Google Scholar PubMed

8. Ussher JR, Elmariah S, Gerszten RE, Dyck JR. The emerging role of metabolomics in the diagnosis and prognosis of cardiovascular disease. J Am Coll Cardiol 2016;68:2850–70.10.1016/j.jacc.2016.09.972Search in Google Scholar PubMed

9. Lin W, Zhang J, Liu Y, Wu R, Yang H, Hu X, et al. Studies on diagnostic biomarkers and therapeutic mechanism of Alzheimer’s disease through metabolomics and hippocampal proteomics. Eur J Pharm Sci 2017;105:119–26.10.1016/j.ejps.2017.05.003Search in Google Scholar PubMed

10. Min L, Choy E, Tu C, Hornicek F, Duan Z. Application of metabolomics in sarcoma: from biomarkers to therapeutic targets. Crit Rev Oncol Hematol 2017;116:1–10.10.1016/j.critrevonc.2017.05.003Search in Google Scholar PubMed PubMed Central

11. Dinis-Oliveira RJ. Metabolomics of methylphenidate and ethylphenidate: implications in pharmacological and toxicological effects. Eur J Drug Metab Pharmacokinet 2017;42:11–6.10.1007/s13318-016-0362-1Search in Google Scholar PubMed

12. Ramirez T, Strigun A, Verlohner A, Huener HA, Peter E, Herold M, et al. Prediction of liver toxicity and mode of action using metabolomics in vitro in HepG2 cells. Arch Toxicol 2018;92: 893–906.10.1007/s00204-017-2079-6Search in Google Scholar PubMed PubMed Central

13. Trifonova O, Knight RA, Lisitsa A, Melino G, Antonov AV. Exploration of individuality in drug metabolism by high-throughput metabolomics: the fast line for personalized medicine. Drug Discov Today 2016;21:103–10.10.1016/j.drudis.2015.07.011Search in Google Scholar PubMed

14. Psychogios N, Hau DD, Peng J, Guo AC, Mandal R, Bouatra S, et al. The human serum metabolome. PLoS One 2011;6:e16957.10.1371/journal.pone.0016957Search in Google Scholar PubMed PubMed Central

15. Bouatra S, Aziat F, Mandal R, Guo AC, Wilson MR, Knox C, et al. The human urine metabolome. PLoS One 2013;8:e73076.10.1371/journal.pone.0073076Search in Google Scholar PubMed PubMed Central

16. Eritja N, Jove M, Fasmer KE, Gatius S, Portero-Otin M, Trovik J, et al. Tumour-microenvironmental blood flow determines a metabolomic signature identifying lysophospholipids and resolvin D as biomarkers in endometrial cancer patients. Oncotarget 2017;8:109018–26.10.18632/oncotarget.22558Search in Google Scholar PubMed PubMed Central

17. Monleon D, Wang L, Tang Y, Liu S, Mao S, Ling Y, et al. Metabonomic profiling of serum and urine by 1H NMR-based spectroscopy discriminates patients with chronic obstructive pulmonary disease and healthy individuals. PLoS One 2013;8:e65675.10.1371/journal.pone.0065675Search in Google Scholar PubMed PubMed Central

18. Nicholson JK, Lindon JC. Systems biology: metabonomics. Nature 2008;455:1054–6.10.1038/4551054aSearch in Google Scholar PubMed

19. Soininen P, Kangas AJ, Wurtz P, Suna T, Ala-Korpela M. Quantitative serum nuclear magnetic resonance metabolomics in cardiovascular epidemiology and genetics. Circ Cardiovasc Genet 2015;8:192–206.10.1161/CIRCGENETICS.114.000216Search in Google Scholar PubMed

20. Markley JL, Bruschweiler R, Edison AS, Eghbalnia HR, Powers R, Raftery D, et al. The future of NMR-based metabolomics. Curr Opin Biotechnol 2017;43:34–40.10.1016/j.copbio.2016.08.001Search in Google Scholar PubMed PubMed Central

21. Dickens AM, Larkin JR, Griffin JL, Cavey A, Matthews L, Turner MR, et al. A type 2 biomarker separates relapsing-remitting from secondary progressive multiple sclerosis. Neurology 2014;83:1492–9.10.1212/WNL.0000000000000905Search in Google Scholar PubMed PubMed Central

22. Jimenez B, Mirnezami R, Kinross J, Cloarec O, Keun HC, Holmes E, et al. 1H HR-MAS NMR spectroscopy of tumor-induced local metabolic “field-effects” enables colorectal cancer staging and prognostication. J Proteome Res 2013;12:959–68.10.1021/pr3010106Search in Google Scholar PubMed

23. Loomba R, Quehenberger O, Armando A, Dennis EA. Polyunsaturated fatty acid metabolites as novel lipidomic biomarkers for noninvasive diagnosis of nonalcoholic steatohepatitis. J Lipid Res 2015;56:185–92.10.1194/jlr.P055640Search in Google Scholar PubMed PubMed Central

24. Li Y, Ju L, Hou Z, Deng H, Zhang Z, Wang L, et al. Screening, verification, and optimization of biomarkers for early prediction of cardiotoxicity based on metabolomics. J Proteome Res 2015;14:2437–45.10.1021/pr501116cSearch in Google Scholar PubMed

25. Storey JD. A direct approach to false discovery rates. J R Stat Soc Series B Stat Methodol 2002;64:479–98.10.1111/1467-9868.00346Search in Google Scholar

26. Storey JD. The positive false discovery rate: a Bayesian interpretation and the q-value. Ann Stat 2003;31:2013–35.10.1214/aos/1074290335Search in Google Scholar

27. Holmes E, Loo RL, Stamler J, Bictash M, Yap IK, Chan Q, et al. Human metabolic phenotype diversity and its association with diet and blood pressure. Nature 2008;453:396–400.10.1038/nature06882Search in Google Scholar PubMed PubMed Central

28. Holmes E, Wilson ID, Nicholson JK. Metabolic phenotyping in health and disease. Cell 2008;134:714.10.1016/j.cell.2008.08.026Search in Google Scholar PubMed

29. Spratlin JL, Serkova NJ, Eckhardt SG. Clinical applications of metabolomics in oncology: a review. Clin Cancer Res 2009;15:431–40.10.1158/1078-0432.CCR-08-1059Search in Google Scholar PubMed PubMed Central

30. Liu Y, Hong Z, Tan G, Dong X, Yang G, Zhao L, et al. NMR and LC/MS-based global metabolomics to identify serum biomarkers differentiating hepatocellular carcinoma from liver cirrhosis. Int J Cancer 2014;135:658–68.10.1002/ijc.28706Search in Google Scholar PubMed

31. Lin X, Zhan B, Wen S, Li Z, Huang H, Feng J. Metabonomic alterations from pancreatic intraepithelial neoplasia to pancreatic ductal adenocarcinoma facilitate the identification of biomarkers in serum for early diagnosis of pancreatic cancer. Mol Biosyst 2016;12:2883–92.10.1039/C6MB00381HSearch in Google Scholar PubMed

32. Wen S, Zhan B, Feng J, Hu W, Lin X, Bai J, et al. Non-invasively predicting differentiation of pancreatic cancer through comparative serum metabonomic profiling. BMC Cancer 2017;17:708.10.1186/s12885-017-3703-9Search in Google Scholar PubMed PubMed Central

33. McConnell YJ, Farshidfar F, Weljie AM, Kopciuk KA, Dixon E, Ball CG, et al. Distinguishing benign from malignant pancreatic and periampullary lesions using combined use of (1)H-NMR spectroscopy and gas chromatography-mass spectrometry. Metabolites 2017;7:3.10.3390/metabo7010003Search in Google Scholar PubMed PubMed Central

34. Yang Y, Liu Y, Zheng L, Zhang Q, Gu Q, Wang L, et al. 1H NMR based serum metabolic profiles associated with pathological progression of pancreatic islet β cell tumor in Rip1-Tag2 Mice. Int J Biol Sci 2015;11:595–603.10.7150/ijbs.11058Search in Google Scholar PubMed PubMed Central

35. Zamani Z, Arjmand M, Vahabi F, Eshaq Hosseini SM, Fazeli SM, Iravani A, et al. A metabolic study on colon cancer using (1)h nuclear magnetic resonance spectroscopy. Biochem Res Int 2014;2014:348712.10.1155/2014/348712Search in Google Scholar PubMed PubMed Central

36. Vahabi F, Sadeghi S, Arjmand M, Mirkhani F, Hosseini E, Mehrabanfar M, et al. Staging of colorectal cancer using serum metabolomics with (1)HNMR Spectroscopy. Iran J Basic Med Sci 2017;20:835–40.Search in Google Scholar

37. Louis E, Adriaensens P, Guedens W, Bigirumurame T, Baeten K, Vanhove K, et al. Detection of lung cancer through metabolic changes measured in blood plasma. J Thorac Oncol 2016;11: 516–23.10.1016/j.jtho.2016.01.011Search in Google Scholar PubMed

38. Hao D, Sarfaraz MO, Farshidfar F, Bebb DG, Lee CY, Card CM, et al. Temporal characterization of serum metabolite signatures in lung cancer patients undergoing treatment. Metabolomics 2016;12:58.10.1007/s11306-016-0961-5Search in Google Scholar PubMed PubMed Central

39. Puchades-Carrasco L, Jantus-Lewintre E, Perez-Rambla C, Garcia-Garcia F, Lucas R, Calabuig S, et al. Serum metabolomic profiling facilitates the non-invasive identification of metabolic biomarkers associated with the onset and progression of non-small cell lung cancer. Oncotarget 2016;7:12904–16.10.18632/oncotarget.7354Search in Google Scholar PubMed PubMed Central

40. Deja S, Porebska I, Kowal A, Zabek A, Barg W, Pawelczyk K, et al. Metabolomics provide new insights on lung cancer staging and discrimination from chronic obstructive pulmonary disease. J Pharm Biomed Anal 2014;100:369–80.10.1016/j.jpba.2014.08.020Search in Google Scholar PubMed

41. Zheng H, Ji J, Zhao L, Chen M, Shi A, Pan L, et al. Prediction and diagnosis of renal cell carcinoma using nuclear magnetic resonance-based serum metabolomics and self-organizing maps. Oncotarget 2016;7:59189–98.10.18632/oncotarget.10830Search in Google Scholar PubMed PubMed Central

42. Falegan OS, Ball MW, Shaykhutdinov RA, Pieroraio PM, Farshidfar F, Vogel HJ, et al. Urine and serum metabolomics analyses may distinguish between stages of renal cell carcinoma. Metabolites 2017;7:6.10.3390/metabo7010006Search in Google Scholar PubMed PubMed Central

43. Singh A, Sharma RK, Chagtoo M, Agarwal G, George N, Sinha N, et al. 1H NMR metabolomics reveals association of high expression of inositol 1, 4, 5 trisphosphate receptor and metabolites in breast cancer patients. PLoS One 2017;12:e0169330.10.1371/journal.pone.0169330Search in Google Scholar PubMed PubMed Central

44. Jobard E, Pontoizeau C, Blaise BJ, Bachelot T, Elena-Herrmann B, Tredan O. A serum nuclear magnetic resonance-based metabolomic signature of advanced metastatic human breast cancer. Cancer Lett 2014;343:33–41.10.1016/j.canlet.2013.09.011Search in Google Scholar PubMed

45. Richard V, Conotte R, Mayne D, Colet JM. Does the 1H-NMR plasma metabolome reflect the host-tumor interactions in human breast cancer? Oncotarget 2017;8:49915–30.10.18632/oncotarget.18307Search in Google Scholar PubMed PubMed Central

46. Bahado-Singh RO, Lugade A, Field J, Al-Wahab Z, Han B, Mandal R, et al. Metabolomic prediction of endometrial cancer. Metabolomics 2017;14:6.10.1007/s11306-017-1290-zSearch in Google Scholar PubMed

47. Giskeodegard GF, Hansen AF, Bertilsson H, Gonzalez SV, Kristiansen KA, Bruheim P, et al. Metabolic markers in blood can separate prostate cancer from benign prostatic hyperplasia. Br J Cancer 2015;113:1712–9.10.1038/bjc.2015.411Search in Google Scholar PubMed PubMed Central

48. Kumar D, Gupta A, Mandhani A, Sankhwar SN. Metabolomics-derived prostate cancer biomarkers: fact or fiction? J Proteome Res 2015;14:1455–64.10.1021/pr5011108Search in Google Scholar PubMed

49. Kumar D, Gupta A, Mandhani A, Sankhwar SN. NMR spectroscopy of filtered serum of prostate cancer: a new frontier in metabolomics. Prostate 2016;76:1106–19.10.1002/pros.23198Search in Google Scholar PubMed

50. Wang Y, Luo X, Zhang GH, Li SL. Nuclear magnetic resonance-based study reveals the metabolomics profile of nasopharyngeal carcinoma. Genet Mol Res 2016;15:1–9.10.4238/gmr.15027796Search in Google Scholar PubMed

51. Lodi A, Tiziani S, Khanim FL, Gunther UL, Viant MR, Morgan GJ, et al. Proton NMR-based metabolite analyses of archived serial paired serum and urine samples from myeloma patients at different stages of disease activity identifies acetylcarnitine as a novel marker of active disease. PLoS One 2013;8:e56422.10.1371/journal.pone.0056422Search in Google Scholar PubMed PubMed Central

52. Yang F, Du J, Zhang H, Ruan G, Xiang J, Wang L, et al. Serum metabolomics of burkitt lymphoma mouse models. PLoS One 2017;12:e0170896.10.1371/journal.pone.0170896Search in Google Scholar PubMed PubMed Central

53. Li Y, Nie L, Jiang H, Lin J, Zhou H, Xie J, et al. Metabonomics study of essential hypertension and its Chinese medicine subtypes by using gas chromatography-mass spectrometry and nuclear magnetic resonance spectroscopy. Evid Based Complement Alternat Med 2013;2013:625906.10.1155/2013/625906Search in Google Scholar PubMed PubMed Central

54. Evaristi MF, Caubere C, Harmancey R, Desmoulin F, Peacock WF, Berry M, et al. Increased mean aliphatic lipid chain length in left ventricular hypertrophy secondary to arterial hypertension: a cross-sectional study. Medicine (Baltimore) 2016;95:e4965.10.1097/MD.0000000000004965Search in Google Scholar PubMed PubMed Central

55. Deidda M, Piras C, Cadeddu Dessalvi C, Congia D, Locci E, Ascedu F, et al. Blood metabolomic fingerprint is distinct in healthy coronary and in stenosing or microvascular ischemic heart disease. J Trans Med 2017;15:112.10.1186/s12967-017-1215-7Search in Google Scholar PubMed PubMed Central

56. Shi Q, Zhao H, Chen J, Li Y, Li Z, Wang J, et al. Study on Qi deficiency syndrome identification modes of coronary heart disease based on metabolomic biomarkers. Evid Based Complement Alternat Med 2014;2014:1–15.10.1155/2014/281829Search in Google Scholar PubMed PubMed Central

57. Yao W, Gao Y, Wan Z. Serum metabolomics profiling to identify biomarkers for unstable angina. Biomed Res Int 2017;2017:7657306.10.1155/2017/7657306Search in Google Scholar PubMed PubMed Central

58. Ameta K, Gupta A, Ameta D, Sethi R, Kumar D, Ahmad I, et al. 1 H NMR-derived metabolomics of filtered serum of myocardial ischemia in unstable angina patients. Clin Chim Acta 2016;456:56–62.10.1016/j.cca.2016.02.020Search in Google Scholar PubMed

59. Zordoky BN, Sung MM, Ezekowitz J, Mandal R, Han B, Bjorndahl TC, et al. Metabolomic fingerprint of heart failure with preserved ejection fraction. PLoS One 2015;10:e0124844.10.1371/journal.pone.0124844Search in Google Scholar PubMed PubMed Central

60. Du Z, Shen A, Huang Y, Su L, Lai W, Wang P, et al. 1H-NMR-based metabolic analysis of human serum reveals novel markers of myocardial energy expenditure in heart failure patients. PLoS One 2014;9:e88102.10.1371/journal.pone.0088102Search in Google Scholar PubMed PubMed Central

61. Abbate A, Desmoulin F, Galinier M, Trouillet C, Berry M, Delmas C, et al. Metabonomics analysis of plasma reveals the lactate to cholesterol ratio as an independent prognostic factor of short-term mortality in acute heart failure. PLoS One 2013;8:e60737.10.1371/journal.pone.0060737Search in Google Scholar PubMed PubMed Central

62. Wurtz P, Havulinna AS, Soininen P, Tynkkynen T, Prieto-Merino D, Tillin T, et al. Metabolite profiling and cardiovascular event risk: a prospective study of 3 population-based cohorts. Circulation 2015;131:774–85.10.1161/CIRCULATIONAHA.114.013116Search in Google Scholar PubMed PubMed Central

63. Hao J, Yang T, Zhou Y, Gao GY, Xing F, Peng Y, et al. Serum metabolomics analysis reveals a distinct metabolic profile of patients with primary biliary cholangitis. Sci Rep 2017;7:784.10.1038/s41598-017-00944-9Search in Google Scholar PubMed PubMed Central

64. Embade N, Marino Z, Diercks T, Cano A, Lens S, Cabrera D, et al. Metabolic characterization of advanced liver fibrosis in HCV patients as studied by serum 1H-NMR spectroscopy. PLoS One 2016;11:e0155094.10.1371/journal.pone.0155094Search in Google Scholar PubMed PubMed Central

65. McPhail MJ, Shawcross DL, Lewis MR, Coltart I, Want EJ, Antoniades CG, et al. Multivariate metabotyping of plasma predicts survival in patients with decompensated cirrhosis. J Hepatol 2016;64:1058–67.10.1016/j.jhep.2016.01.003Search in Google Scholar PubMed PubMed Central

66. Wei DD, Wang JS, Wang PR, Li MH, Yang MH, Kong LY. Toxic effects of chronic low-dose exposure of thioacetamide on rats based on NMR metabolic profiling. J Pharm Biomed Anal 2014;98:334–8.10.1016/j.jpba.2014.05.035Search in Google Scholar PubMed

67. Amathieu R, Triba MN, Nahon P, Bouchemal N, Kamoun W, Haouache H, et al. Serum 1H-NMR metabolomic fingerprints of acute-on-chronic liver failure in intensive care unit patients with alcoholic cirrhosis. PLoS One 2014;9:e89230.10.1371/journal.pone.0089230Search in Google Scholar PubMed PubMed Central

68. Männistö VT, Simonen M, Hyysalo J, Soininen P, Kangas AJ, Kaminska D, et al. Ketone body production is differentially altered in steatosis and non-alcoholic steatohepatitis in obese humans. Liver Int 2015;35:1853–61.10.1111/liv.12769Search in Google Scholar PubMed

69. Sharma RK, Mishra K, Farooqui A, Behari A, Kapoor VK, Sinha N. (1)H nuclear magnetic resonance (NMR)-based serum metabolomics of human gallbladder inflammation. Inflamm Res 2017;66:97–105.10.1007/s00011-016-0998-ySearch in Google Scholar PubMed

70. Dong F, Zhang L, Hao F, Tang H, Wang Y. Systemic responses of mice to dextran sulfate sodium-induced acute ulcerative colitis using 1H NMR spectroscopy. J Proteome Res 2013;12:2958–66.10.1021/pr4002383Search in Google Scholar PubMed

71. Dawiskiba T. Serum and urine metabolomic fingerprinting in diagnostics of inflammatory bowel diseases. World J Gastroenterol 2014;20:163.10.3748/wjg.v20.i1.163Search in Google Scholar PubMed PubMed Central

72. Motta A, Munshi SU, Rewari BB, Bhavesh NS, Jameel S. Nuclear magnetic resonance based profiling of biofluids reveals metabolic dysregulation in HIV-infected persons and those on anti-retroviral therapy. PLoS One 2013;8:e64298.10.1371/journal.pone.0064298Search in Google Scholar PubMed PubMed Central

73. McKnight TR, Yoshihara HA, Sitole LJ, Martin JN, Steffens F, Meyer D. A combined chemometric and quantitative NMR analysis of HIV/AIDS serum discloses metabolic alterations associated with disease status. Mol Biosyst 2014;10:2889–97.10.1039/C4MB00347KSearch in Google Scholar

74. Zheng H, Chen M, Lu S, Zhao L, Ji J, Gao H. Metabolic characterization of hepatitis B virus-related liver cirrhosis using NMR-based serum metabolomics. Metabolomics 2017;13:121.10.1007/s11306-017-1260-5Search in Google Scholar

75. Hsieh MH, Gouveia LR, Santos JC, Silva RD, Batista AD, Domingues AL, et al. Diagnosis of coinfection by schistosomiasis and viral hepatitis B or C using 1H NMR-based metabonomics. PLoS One 2017;12:e0182196.10.1371/journal.pone.0182196Search in Google Scholar PubMed PubMed Central

76. Zhou A, Ni J, Xu Z, Wang Y, Lu S, Sha W, et al. Application of 1H NMR spectroscopy-based metabolomics to sera of tuberculosis patients. J Proteome Res 2013;12:4642–9.10.1021/pr4007359Search in Google Scholar PubMed PubMed Central

77. Sun L, Li JQ, Ren N, Qi H, Dong F, Xiao J, et al. Utility of novel plasma metabolic markers in the diagnosis of pediatric tuberculosis: a classification and regression tree analysis approach. J Proteome Res 2016;15:3118–25.10.1021/acs.jproteome.6b00228Search in Google Scholar PubMed

78. Zhou A, Ni J, Xu Z, Wang Y, Zhang H, Wu W, et al. Metabolomics specificity of tuberculosis plasma revealed by 1H NMR spectroscopy. Tuberculosis 2015;95:294–302.10.1016/j.tube.2015.02.038Search in Google Scholar PubMed PubMed Central

79. Ghosh S, Sengupta A, Sharma S, Sonawat HM. Early prediction of cerebral malaria by 1H NMR based metabolomics. Malar J 2016;15:198.10.1186/s12936-016-1256-zSearch in Google Scholar PubMed PubMed Central

80. Sengupta A, Ghosh S, Das BK, Panda A, Tripathy R, Pied S, et al. Host metabolic responses to Plasmodium falciparum infections evaluated by 1H NMR metabolomics. Mol Biosyst 2016;12: 3324–32.10.1039/C6MB00362ASearch in Google Scholar PubMed

81. Bervoets L, Massa G, Guedens W, Louis E, Noben J-P, Adriaensens P. Metabolic profiling of type 1 diabetes mellitus in children and adolescents: a case–control study. Diabetol Metab Syndr 2017;9:48.10.1186/s13098-017-0246-9Search in Google Scholar PubMed PubMed Central

82. Liu J, Wang D, Chen Y, Sun H, He S, Wang C, et al. 1H NMR-based metabonomic analysis of serum and urine in a nonhuman primate model of diabetic nephropathy. Mol Biosyst 2013;9:2645–52.10.1039/c3mb70212jSearch in Google Scholar PubMed

83. Roy Choudhury S, Mishra BP, Khan T, Chattopadhayay R, Lodh I, Datta Ray C, et al. Serum metabolomics of Indian women with polycystic ovary syndrome using 1H NMR coupled with a pattern recognition approach. Mol Biosyst 2016;12:3407–16.10.1039/C6MB00420BSearch in Google Scholar

84. Won EY, Yoon MK, Kim SW, Jung Y, Bae HW, Lee D, et al. Gender-specific metabolomic profiling of obesity in leptin-deficient ob/ob mice by 1H NMR spectroscopy. PLoS One 2013;8:e75998.10.1371/journal.pone.0075998Search in Google Scholar PubMed PubMed Central

85. Probert F, Ruiz-Rodado V, Vruchte DT, Nicoli ER, Claridge TD, Wassif CA, et al. NMR analysis reveals significant differences in the plasma metabolic profiles of Niemann Pick C1 patients, heterozygous carriers, and healthy controls. Sci Rep 2017;7:6320.10.1038/s41598-017-06264-2Search in Google Scholar PubMed PubMed Central

86. Antcliffe D, Jimenez B, Veselkov K, Holmes E, Gordon AC. Metabolic profiling in patients with pneumonia on intensive care. EBioMedicine 2017;18:244–53.10.1016/j.ebiom.2017.03.034Search in Google Scholar PubMed PubMed Central

87. Wang L, Tang Y, Liu S, Mao S, Ling Y, Liu D, et al. Metabonomic profiling of serum and urine by (1)H NMR-based spectroscopy discriminates patients with chronic obstructive pulmonary disease and healthy individuals. PLoS One 2013;8:e65675.10.1371/journal.pone.0065675Search in Google Scholar PubMed PubMed Central

88. Xu W-F, Upur H, Wu Y-H, Mamtimin B, Yang J, Ga Y-J, et al. Metabolomic changes in patients with chronic obstructive pulmonary disease with abnormal Savda syndrome. Exp Ther Med 2015;9:425–31.10.3892/etm.2014.2085Search in Google Scholar PubMed PubMed Central

89. Fortis S, Lusczek ER, Weinert CR, Beilman GJ. Metabolomics in COPD acute respiratory failure requiring noninvasive positive pressure ventilation. Can Respir J 2017;2017:9480346.10.1155/2017/9480346Search in Google Scholar PubMed PubMed Central

90. Singh C, Rai RK, Azim A, Sinha N, Ahmed A, Singh K, et al. Metabolic profiling of human lung injury by 1 H high-resolution nuclear magnetic resonance spectroscopy of blood serum. Metabolomics 2015;11:166–74.10.1007/s11306-014-0688-0Search in Google Scholar

91. Liu CC, Wu YF, Feng GM, Gao XX, Zhou YZ, Hou WJ, et al. Plasma-metabolite-biomarkers for the therapeutic response in depressed patients by the traditional Chinese medicine formula Xiaoyaosan: a (1)H NMR-based metabolomics approach. J Affect Disord 2015;185:156–63.10.1016/j.jad.2015.05.005Search in Google Scholar PubMed

92. Wu H, Wang P, Liu M, Tang L, Fang J, Zhao Y, et al. A 1H-NMR-based metabonomic study on the anti-depressive effect of the total alkaloid of corydalis rhizoma. Molecules 2015;20: 10047–64.10.3390/molecules200610047Search in Google Scholar PubMed PubMed Central

93. Rudolph U, Tian J-S, Shi B-Y, Xiang H, Gao S, Qin X-M, et al. 1H-NMR-based metabonomic studies on the anti-depressant effect of genipin in the chronic unpredictable mild stress rat model. PLoS One 2013;8:e75721.10.1371/journal.pone.0075721Search in Google Scholar PubMed PubMed Central

94. Sethi S, Pedrini M, Rizzo LB, Zeni-Graiff M, Mas CD, Cassinelli AC, et al. (1)H-NMR, (1)H-NMR T2-edited, and 2D-NMR in bipolar disorder metabolic profiling. Int J Bipolar Disord 2017;5:23.10.1186/s40345-017-0088-2Search in Google Scholar PubMed PubMed Central

95. Kim E, Jung YS, Kim H, Kim JS, Park M, Jeong J, et al. Metabolomic signatures in peripheral blood associated with Alzheimer’s disease amyloid-beta-induced neuroinflammation. J Alzheimers Dis 2014;42:421–33.10.3233/JAD-132165Search in Google Scholar PubMed

96. Zacharias HU, Hochrein J, Vogl FC, Schley G, Mayer F, Jeleazcov C, et al. Identification of plasma metabolites prognostic of acute kidney injury after cardiac surgery with cardiopulmonary bypass. J Proteome Res 2015;14:2897–905.10.1021/acs.jproteome.5b00219Search in Google Scholar PubMed

97. Kim JA, Choi HJ, Kwon YK, Ryu DH, Kwon TH, Hwang GS. 1H NMR-based metabolite profiling of plasma in a rat model of chronic kidney disease. PLoS One 2014;9:e85445.10.1371/journal.pone.0085445Search in Google Scholar PubMed PubMed Central

98. Mika A, Wojtowicz W, Zabek A, Mlynarz P, Chmielewski M, Sledzinski T, et al. Application of nuclear magnetic resonance spectroscopy for the detection of metabolic disorders in patients with moderate kidney insufficiency. J Pharm Biomed Anal 2018;149:1–8.10.1016/j.jpba.2017.10.037Search in Google Scholar PubMed

99. Guleria A, Pratap A, Dubey D, Rawat A, Chaurasia S, Sukesh E, et al. NMR based serum metabolomics reveals a distinctive signature in patients with Lupus Nephritis. Sci Rep 2016;6:35309.10.1038/srep35309Search in Google Scholar PubMed PubMed Central

100. Vicente-Muñoz S, Morcillo I, Puchades-Carrasco L, Payã V, Pellicer A, Pineda-Lucena A. Pathophysiologic processes have an impact on the plasma metabolomic signature of endometriosis patients. Fertil Steril 2016;106:1733–41.10.1016/j.fertnstert.2016.09.014Search in Google Scholar PubMed

101. Wang Y, Wang YG, Ma TF, Li M, Gu SL. Dynamic metabolites profile of cerebral ischemia/reperfusion revealed by (1)H NMR-based metabolomics contributes to potential biomarkers. Int J Clin Exp Pathol 2014;7:4067–75.Search in Google Scholar

102. Moussallieh FM, Elbayed K, Chanson JB, Rudolf G, Piotto M, De Seze J, et al. Serum analysis by 1H nuclear magnetic resonance spectroscopy: a new tool for distinguishing neuromyelitis optica from multiple sclerosis. Mult Scler 2013;20:558–65.10.1177/1352458513504638Search in Google Scholar PubMed

103. Dickens AM, Larkin JR, Davis BG, Griffin JL, Claridge TD, Sibson NR, et al. NMR-based metabolomics separates the distinct stages of disease in a chronic relapsing model of multiple sclerosis. J Neuroimmune Pharmacol 2015;10:435–44.10.1007/s11481-015-9622-0Search in Google Scholar PubMed

104. Austdal M, Skrastad RB, Gundersen AS, Austgulen R, Iversen AC, Bathen TF. Metabolomic biomarkers in serum and urine in women with preeclampsia. PLoS One 2014;9:e91923.10.1371/journal.pone.0091923Search in Google Scholar PubMed PubMed Central

105. Austdal M, Tangeras LH, Skrastad RB, Salvesen K, Austgulen R, Iversen AC, et al. First trimester urine and serum metabolomics for prediction of preeclampsia and gestational hypertension: a prospective screening study. Int J Mol Sci 2015;16:21520–38.10.3390/ijms160921520Search in Google Scholar PubMed PubMed Central

106. Jin B, Liu L, Zhang S, Cao X, Xu Y, Wang J, et al. Nuclear magnetic resonance-assisted metabolic analysis of plasma for mild gestational diabetes mellitus patients. Metab Syndr Relat Disord 2017;15:439–49.10.1089/met.2017.0065Search in Google Scholar PubMed

107. Pinto J, Almeida LM, Martins AS, Duarte D, Barros AS, Galhano E, et al. Prediction of gestational diabetes through NMR metabolomics of maternal blood. J Proteome Res 2015;14:2696–706.10.1021/acs.jproteome.5b00260Search in Google Scholar PubMed

108. Bogren LK, Murphy CJ, Johnston EL, Sinha N, Serkova NJ, Drew KL. 1H-NMR metabolomic biomarkers of poor outcome after hemorrhagic shock are absent in hibernators. PLoS One 2014;9:e107493.10.1371/journal.pone.0107493Search in Google Scholar PubMed PubMed Central

109. Zhang Y, Cai B, Jiang H, Yan H, Yang H, Peng J, et al. Use of 1H-nuclear magnetic resonance to screen a set of biomarkers for monitoring metabolic disturbances in severe burn patients. Crit Care 2014;18:R159.10.1186/cc13999Search in Google Scholar PubMed PubMed Central

110. Lains I, Duarte D, Barros AS, Martins AS, Gil J, Miller JB, et al. Human plasma metabolomics in age-related macular degeneration (AMD) using nuclear magnetic resonance spectroscopy. PLoS One 2017;12:e0177749.10.1371/journal.pone.0177749Search in Google Scholar PubMed PubMed Central

111. Zabek A, Swierkot J, Malak A, Zawadzka I, Deja S, Bogunia-Kubik K, et al. Application of (1)H NMR-based serum metabolomic studies for monitoring female patients with rheumatoid arthritis. J Pharm Biomed Anal 2016;117:544–50.10.1016/j.jpba.2015.10.007Search in Google Scholar PubMed

112. Hart CD, Vignoli A, Tenori L, Uy GL, Van To T, Adebamowo C, et al. Serum metabolomic profiles identify er-positive early breast cancer patients at increased risk of disease recurrence in a multicenter population. Clin Cancer Res 2017;23:1422–31.10.1158/1078-0432.CCR-16-1153Search in Google Scholar PubMed PubMed Central

113. Zhang ZY, Marrachelli VG, Thijs L, Yang WY, Wei FF, Monleon D, et al. Diastolic left ventricular function in relation to circulating metabolic biomarkers in a general population. J Am Heart Assoc 2016;5:e002681.10.1161/JAHA.115.002681Search in Google Scholar PubMed PubMed Central

114. Sonkar K, Behari A, Kapoor VK, Sinha N. 1H NMR metabolic profiling of human serum associated with benign and malignant gallstone diseases. Metabolomics 2013;9:515–28.10.1007/s11306-012-0468-7Search in Google Scholar

115. Forsgard RA, Marrachelli VG, Korpela K, Frias R, Collado MC, Korpela R, et al. Chemotherapy-induced gastrointestinal toxicity is associated with changes in serum and urine metabolome and fecal microbiota in male Sprague-Dawley rats. Cancer Chemother Pharmacol 2017;80:317–32.10.1007/s00280-017-3364-zSearch in Google Scholar PubMed PubMed Central

116. Sarfaraz MO, Myers RP, Coffin CS, Gao ZH, Shaheen AA, Crotty PM, et al. A quantitative metabolomics profiling approach for the noninvasive assessment of liver histology in patients with chronic hepatitis C. Clin Transl Med 2016;5:33.10.1186/s40169-016-0109-2Search in Google Scholar PubMed PubMed Central

117. Jiang CY, Yang KM, Yang L, Miao ZX, Wang YH, Zhu HB. A (1)H NMR-based metabonomic investigation of time-related metabolic trajectories of the plasma, urine and liver extracts of hyperlipidemic hamsters. PLoS One 2013;8:e66786.10.1371/journal.pone.0066786Search in Google Scholar PubMed PubMed Central

118. Wojtowicz W, Zabek A, Deja S, Dawiskiba T, Pawelka D, Glod M, et al. Serum and urine 1H NMR-based metabolomics in the diagnosis of selected thyroid diseases. Sci Rep 2017;7:9108.10.1038/s41598-017-09203-3Search in Google Scholar PubMed PubMed Central

119. Monteiro MS, Carvalho M, Bastos ML, Guedes de Pinho P. Metabolomics analysis for biomarker discovery: advances and challenges. Curr Med Chem 2013;20:257–71.10.2174/092986713804806621Search in Google Scholar PubMed

120. Salek RM, Steinbeck C, Viant MR, Goodacre R, Dunn WB. The role of reporting standards for metabolite annotation and identification in metabolomic studies. Gigascience 2013;2:13.10.1186/2047-217X-2-13Search in Google Scholar PubMed PubMed Central

121. Emwas AH, Luchinat C, Turano P, Tenori L, Roy R, Salek RM, et al. Standardizing the experimental conditions for using urine in NMR-based metabolomic studies with a particular focus on diagnostic studies: a review. Metabolomics 2015;11:872–94.10.1007/s11306-014-0746-7Search in Google Scholar PubMed PubMed Central

122. Fiehn O, Robertson D, Griffin J, van der Werf M, Nikolau B, Morrison N, et al. The metabolomics standards initiative (MSI). Metabolomics 2007;3:175–8.10.1007/s11306-007-0070-6Search in Google Scholar

123. Sumner LW, Amberg A, Barrett D, Beale MH, Beger R, Daykin CA, et al. Proposed minimum reporting standards for chemical analysis Chemical Analysis Working Group (CAWG) Metabolomics Standards Initiative (MSI). Metabolomics 2007;3: 211–21.10.1007/s11306-007-0082-2Search in Google Scholar PubMed PubMed Central

124. Haug K, Salek RM, Conesa P, Hastings J, de Matos P, Rijnbeek M, et al. MetaboLights–an open-access general-purpose repository for metabolomics studies and associated meta-data. Nucleic Acids Res 2013;41:D781–6.10.1093/nar/gks1004Search in Google Scholar PubMed PubMed Central

125. Emwas A-H, Salek RM, Griffin JL, Merzaban J. NMR-based metabolomics in human disease diagnosis: applications, limitations, and recommendations. Metabolomics 2013;9:1048–72.10.1007/s11306-013-0524-ySearch in Google Scholar

126. Emwas AH. The strengths and weaknesses of NMR spectroscopy and mass spectrometry with particular focus on metabolomics research. Methods Mol Biol 2015;1277:161–93.10.1007/978-1-4939-2377-9_13Search in Google Scholar PubMed

127. Scalbert A, Brennan L, Fiehn O, Hankemeier T, Kristal BS, van Ommen B, et al. Mass-spectrometry-based metabolomics: limitations and recommendations for future progress with particular focus on nutrition research. Metabolomics 2009;5:435–58.10.1007/s11306-009-0168-0Search in Google Scholar PubMed PubMed Central

Received: 2018-04-14
Accepted: 2018-07-16
Published Online: 2018-08-31
Published in Print: 2019-03-26

©2019 Walter de Gruyter GmbH, Berlin/Boston

Downloaded on 1.3.2024 from https://www.degruyter.com/document/doi/10.1515/cclm-2018-0380/html
Scroll to top button