Metabolomics in epidemiology: from metabolite concentrations to integrative reaction networks

Technical Report: A Comprehensive Comparison between Different Quantification Versions of Nightingale Health's 1H-NMR Metabolomics Platform

1H-NMR metabolomics data is increasingly used to track health and disease. Nightingale Health, a major supplier of 1H-NMR metabolomics, has recently updated the quantification strategy to further align with clinical standards. Such updates, however, might influence backward replicability, particularly affecting studies with repeated measures. Using data from BBMRI-NL consortium (~28,000 samples from 28 cohorts), we compared Nightingale data, originally released in 2014 and 2016, with a re-quantified version released in 2020, of which both versions were based on the same NMR spectra. Apart from two discontinued and twenty-three new analytes, we generally observe a high concordance between quantification versions with 73 out of 222 (33%) analytes showing a mean ρ > 0.9 across all cohorts. Conversely, five analytes consistently showed lower Spearman's correlations (ρ < 0.7) between versions, namely acetoacetate, LDL-L, saturated fatty acids, S-HDL-C, and sphingomyelins. Furthermore, previously trained multi-analyte scores, such as MetaboAge or MetaboHealth, might be particularly sensitive to platform changes. Whereas MetaboHealth replicated well, the MetaboAge score had to be retrained due to use of discontinued analytes. Notably, both scores in the re-quantified data recapitulated mortality associations observed previously. Concluding, we urge caution in utilizing different platform versions to avoid mixing analytes, having different units, or simply being discontinued.

The Essential Role of 17-Octadecynoic Acid in the Pathogenesis of Periapical Abscesses

INTRODUCTION

Periapical abscesses are 1 of the most frequent pathologic lesions in the alveolar bone. Recently, we have identified 17-octadecynoic acid (17-ODYA) as the highest unique metabolite in periapical abscesses. Therefore, the aim of this study was to investigate the immunologic and pathophysiological roles of this metabolite in the initiation and development of periapical abscesses.

METHODS

Periodontal ligament fibroblasts and peripheral blood mononuclear cells were treated with 17-ODYA. Gene expression analysis and interleukin (IL)-8 release were determined using quantitative real-time polymerase chain reaction and enzyme-linked immunosorbent assay. Macrophage polarization and cytokine release were also determined using flow cytometry and Luminex bioassay (R&D Systems, Minneapolis, MN), respectively.

RESULTS

In periodontal ligament fibroblasts, 17-ODYA caused significant (P < .0001) up-regulation of IL-1α, IL-1β, IL-6, matrix metalloproteinase-1, and monocyte chemoattractant protein-1 at 10 μmol/L after 6 days of treatment and up-regulation of platelet-derived growth factor alpha and vascular endothelial growth factor alpha at all tested concentrations after 2 days of treatment. In peripheral blood mononuclear cells, 17-ODYA significantly increased the expression of IL-1α, IL-1β, IL-6, matrix metalloproteinase-1, and monocyte chemoattractant protein-1 at 10 μmol/L (P < .0001) and vascular endothelial growth factor alpha and platelet-derived growth factor alpha at 1 μmol/L 17-ODYA (P < .0001). 17-ODYA polarized macrophages toward a proinflammatory phenotype (M1) and suppressed the release of pro- and anti-inflammatory cytokines. 17-ODYA significantly enhanced the release of IL-8.

CONCLUSIONS

This study was the first to identify the pathologic role of 17-ODYA in the development of periapical abscesses. The results of this study are important in shedding light on the pathogenesis of periapical abscesses in relation to microbial metabolites.

Integrated 16S rDNA Gene Sequencing and Untargeted Metabolomics Analyses to Investigate the Gut Microbial Composition and Plasma Metabolic Phenotype in Calves With Dampness-Heat Diarrhea

Dampness-heat diarrhea (DHD), a common syndrome in Chinese dairy farms, is mainly resulted from digestive system disorders, and accompanied with metabolic disorders in some cases. However, the underlying mechanisms in the intestinal microbiome and plasma metabolome in calves with DHD remain unclear. In order to investigate the pathogenesis of DHD in calves, multi-omics techniques including the 16S rDNA gene sequencing and metabolomics were used to analyze gut microbial compositions and plasma metabolic changes in calves. The results indicated that DHD had a significant effect on the intestinal microbial compositions in calves, which was confirmed by changes in microbial population and distribution. A total of 14 genera were changed, including Escherichia-Shigella, Bacteroides, and Fournierella, in calves with DHD (P < 0.05). Functional analysis based on the Kyoto Encyclopedia of Genes and Genomes (KEGG) annotations indicated that 11 metabolic functions (level 2) were significantly enriched in DHD cases. The untargeted metabolomics analysis showed that 440 metabolites including bilineurin, phosphatidylcholine, and glutamate were significantly different between two groups (VIP > 1 and P < 0.05), and they were related to 67 signal pathways. Eight signal pathways including alpha-linolenic acid, linoleic acid, and glycerophospholipid metabolism were significantly enriched (P < 0.05), which may be potential biomarkers of plasma in calves with DHD. Further, 107 pairs of intestinal microbiota-plasma metabolite correlations were determined, e.g., Escherichia-Shigella was significantly associated with changes of sulfamethazine, butyrylcarnitine, and 14 other metabolites, which reflected that metabolic activity was influenced by the microbiome. These microbiota-metabolite pairs might have a relationship with DHD in calves. In conclusion, the findings revealed that DHD had effect on intestinal microbial compositions and plasma metabolome in calves, and the altered metabolic pathways and microorganisms might serve as diagnostic markers and potential therapeutic targets for DHD in calves.

Per- and polyfluoroalkyl substance (PFAS) exposure, maternal metabolomic perturbation, and fetal growth in African American women: A meet-in-the-middle approach

BACKGROUND

Prenatal exposures to per- and polyfluoroalkyl substances (PFAS) have been linked to reduced fetal growth. However, the detailed molecular mechanisms remain largely unknown. This study aims to investigate biological pathways and intermediate biomarkers underlying the association between serum PFAS and fetal growth using high-resolution metabolomics in a cohort of pregnant African American women in the Atlanta area, Georgia.

METHODS

Serum perfluorohexane sulfonic acid (PFHxS), perfluorooctane sulfonic acid (PFOS), perfluorooctanoic acid (PFOA), and perfluorononanoic acid (PFNA) measurements and untargeted serum metabolomics profiling were conducted in 313 pregnant African American women at 8-14 weeks gestation. Multiple linear regression models were applied to assess the associations of PFAS with birth weight and small-for-gestational age (SGA) birth. A high-resolution metabolomics workflow including metabolome-wide association study, pathway enrichment analysis, and chemical annotation and confirmation with a meet-in-the-middle approach was performed to characterize the biological pathways and intermediate biomarkers of the PFAS-fetal growth relationship.

RESULTS

Each log2-unit increase in serum PFNA concentration was significantly associated with higher odds of SGA birth (OR = 1.32, 95% CI 1.07, 1.63); similar but borderline significant associations were found in PFOA (OR = 1.20, 95% CI 0.94, 1.49) with SGA. Among 25,516 metabolic features extracted from the serum samples, we successfully annotated and confirmed 10 overlapping metabolites associated with both PFAS and fetal growth endpoints, including glycine, taurine, uric acid, ferulic acid, 2-hexyl-3-phenyl-2-propenal, unsaturated fatty acid C18:1, androgenic hormone conjugate, parent bile acid, and bile acid-glycine conjugate. Also, we identified 21 overlapping metabolic pathways from pathway enrichment analyses. These overlapping metabolites and pathways were closely related to amino acid, lipid and fatty acid, bile acid, and androgenic hormone metabolism perturbations.

CONCLUSION

In this cohort of pregnant African American women, higher serum concentrations of PFOA and PFNA were associated with reduced fetal growth. Perturbations of biological pathways involved in amino acid, lipid and fatty acid, bile acid, and androgenic hormone metabolism were associated with PFAS exposures and reduced fetal growth, and uric acid was shown to be a potential intermediate biomarker. Our results provide opportunities for future studies to develop early detection and intervention for PFAS-induced fetal growth restriction.

Investigating metabolism by mass spectrometry: From steady state to dynamic view

Metabolism is the set of life-sustaining reactions in organisms. These biochemical reactions are organized in metabolic pathways, in which one metabolite is converted through a series of steps catalyzed by enzymes in another chemical compound. Metabolic reactions are categorized as catabolic, the breaking down of metabolites to produce energy, and/or anabolic, the synthesis of compounds that consume energy. The balance between catabolism of the preferential fuel substrate and anabolism defines the overall metabolism of a cell or tissue. Metabolomics is a powerful tool to gain new insights contributing to the identification of complex molecular mechanisms in the field of biomedical research, both basic and translational. The enormous potential of this kind of analyses consists of two key aspects: (i) the possibility of performing so-called targeted and untargeted experiments through which it is feasible to verify or formulate a hypothesis, respectively, and (ii) the opportunity to run either steady-state analyses to have snapshots of the metabolome at a given time under different experimental conditions or dynamic analyses through the use of labeled tracers. In this review, we will highlight the most important practical (e.g., different sample extraction approaches) and conceptual steps to consider for metabolomic analysis, describing also the main application contexts in which it is used. In addition, we will provide some insights into the most innovative approaches and progress in the field of data analysis and processing, highlighting how this part is essential for the proper extrapolation and interpretation of data.

Opportunities for Gene and Environment Research in Cancer: An Updated Review of NCI's Extramural Grant Portfolio

BACKGROUND

The study of gene-environment (GxE) interactions is a research priority for the NCI. Previously, our group analyzed NCI's extramural grant portfolio from fiscal years (FY) 2007 to 2009 to determine the state of the science in GxE research. This study builds upon our previous effort and examines changes in the landscape of GxE cancer research funded by NCI.

METHODS

The NCI grant portfolio was examined from FY 2010 to 2018 using the iSearch application. A time-trend analysis was conducted to explore changes over the study interval.

RESULTS

A total of 107 grants met the search criteria and were abstracted. The most common cancer types studied were breast (19.6%) and colorectal (18.7%). Most grants focused on GxE using specific candidate genes (69.2%) compared with agnostic approaches using genome-wide (26.2%) or whole-exome/whole-genome next-generation sequencing (NGS) approaches (19.6%); some grants used more than one approach to assess genetic variation. More funded grants incorporated NGS technologies in FY 2016-2018 compared with prior FYs. Environmental exposures most commonly examined were energy balance (46.7%) and drugs/treatment (40.2%). Over the time interval, we observed a decrease in energy balance applications with a concurrent increase in drug/treatment applications.

CONCLUSIONS

Research in GxE interactions has continued to concentrate on common cancers, while there have been some shifts in focus of genetic and environmental exposures. Opportunities exist to study less common cancers, apply new technologies, and increase racial/ethnic diversity.

IMPACT

This analysis of NCI's extramural grant portfolio updates previous efforts and provides a review of NCI grant support for GxE research.

Mass spectrometry-based metabolomics for an in-depth questioning of human health

Today, metabolomics is becoming an indispensable tool to get a more comprehensive analysis of complex living systems, providing insights on multiple aspects of physiology. Although its application in large scale population-based studies is very challenging due to the processing of large sample sets as well as the complexity of data information, its potential to characterize human health is well recognized. Technological advances in metabolomics pave the way for the efficient biomarker discovery of disease etiology, diagnosis and prognosis. Here, different steps of the metabolomics workflow, particularly mass spectrometry-based approaches, are discussed to demonstrate the potential of metabolomics to address biological questioning in human health. First an overview of metabolomics is provided with its interest in human health studies. Analytical development and advances in mass spectrometry instrumentation and computational tools are discussed regarding their application limits. Advancing metabolomics for applicability in human health and large-scale studies is presented and discussed in conclusion.

Current Concepts in Pharmacometabolomics, Biomarker Discovery, and Precision Medicine

Pharmacometabolomics (PMx) studies use information contained in metabolic profiles (or metabolome) to inform about how a subject will respond to drug treatment. Genome, gut microbiome, sex, nutrition, age, stress, health status, and other factors can impact the metabolic profile of an individual. Some of these factors are known to influence the individual response to pharmaceutical compounds. An individual’s metabolic profile has been referred to as his or her “metabotype.” As such, metabolomic profiles obtained prior to, during, or after drug treatment could provide insights about drug mechanism of action and variation of response to treatment. Furthermore, there are several types of PMx studies that are used to discover and inform patterns associated with varied drug responses (i.e., responders vs. non-responders; slow or fast metabolizers). The PMx efforts could simultaneously provide information related to an individual’s pharmacokinetic response during clinical trials and be used to predict patient response to drugs making pharmacometabolomic clinical research valuable for precision medicine. PMx biomarkers can also be discovered and validated during FDA clinical trials. Using biomarkers during medical development is described in US Law under the 21st Century Cures Act. Information on how to submit biomarkers to the FDA and their context of use is defined herein.

Toward a Standardized Strategy of Clinical Metabolomics for the Advancement of Precision Medicine

Despite the tremendous success, pitfalls have been observed in every step of a clinical metabolomics workflow, which impedes the internal validity of the study. Furthermore, the demand for logistics, instrumentations, and computational resources for metabolic phenotyping studies has far exceeded our expectations. In this conceptual review, we will cover inclusive barriers of a metabolomics-based clinical study and suggest potential solutions in the hope of enhancing study robustness, usability, and transferability. The importance of quality assurance and quality control procedures is discussed, followed by a practical rule containing five phases, including two additional "pre-pre-" and "post-post-" analytical steps. Besides, we will elucidate the potential involvement of machine learning and demonstrate that the need for automated data mining algorithms to improve the quality of future research is undeniable. Consequently, we propose a comprehensive metabolomics framework, along with an appropriate checklist refined from current guidelines and our previously published assessment, in the attempt to accurately translate achievements in metabolomics into clinical and epidemiological research. Furthermore, the integration of multifaceted multi-omics approaches with metabolomics as the pillar member is in urgent need. When combining with other social or nutritional factors, we can gather complete omics profiles for a particular disease. Our discussion reflects the current obstacles and potential solutions toward the progressing trend of utilizing metabolomics in clinical research to create the next-generation healthcare system.

Application of metabolomics to characterize environmental pollutant toxicity and disease risks

The increased incidence of non-communicable human diseases may be attributed, at least partially, to exposures to toxic chemicals such as persistent organic pollutants (POPs), air pollutants and heavy metals. Given the high mortality and morbidity of pollutant exposure associated diseases, a better understanding of the related mechanisms of toxicity and impacts on the endogenous host metabolism are needed. The metabolome represents the collection of the intermediates and end products of cellular processes, and is the most proximal reporter of the body's response to environmental exposures and pathological processes. Metabolomics is a powerful tool for studying how organisms interact with their environment and how these interactions shape diseases related to pollutant exposure. This mini review discusses potential biological mechanisms that link pollutant exposure to metabolic disturbances and chronic human diseases, with a focus on recent studies that demonstrate the application of metabolomics as a tool to elucidate biochemical modes of actions of various environmental pollutants. In addition, classes of metabolites that have been shown to be modulated by multiple environmental pollutants will be discussed with an emphasis on their use as potential early biomarkers of disease risks. Taken together, metabolomics is a useful and versatile tool for characterizing the disease risks and mechanisms associated with various environmental pollutants.

Integrating -Omics Approaches into Human Population-Based Studies of Prenatal and Early-Life Exposures

PURPOSE OF REVIEW

We present the study design and methodological suggestions for population-based studies that integrate molecular -omics data and highlight recent studies that have used such data to examine the potential impacts of prenatal environmental exposures on fetal health.

RECENT FINDINGS

Epidemiologic studies have observed numerous relationships between prenatal exposures (smoking, toxic metals, endocrine disruptors) and fetal and early-life molecular profiles, though such investigations have so far been dominated by epigenomic association studies. However, recent transcriptomic, proteomic, and metabolomic studies have demonstrated their promise for the identification of exposure and response biomarkers. Molecular -omics have opened new avenues of research in environmental health that can improve our understanding of disease etiology and contribute to the development of exposure and response biomarkers. Studies that incorporate multiple -omics data from different molecular domains in longitudinally collected samples hold particular promise.

Linking of metabolomic biomarkers with cardiometabolic health in Chinese population

Due to rapid nutrition transitions, the prevalence of cardiometabolic diseases, such as metabolic syndrome, type 2 diabetes, and cardiovascular diseases, has been increasing at an alarming rate in the Chinese population. Moreover, Asians, including Chinese, have been hypothesized to have a higher susceptibility to cardiometabolic diseases than Caucasians. Early prediction and prevention are key to controlling this epidemic trend; to this end, the identification of novel biomarkers is critical to reflect environmental exposure, as well as to reveal endogenous metabolic and pathophysiologic mechanisms. The emerging "omics" technologies, especially metabolomics, offer a unique opportunity to provide novel signatures or fingerprints to understand the effects of genetic and non-genetic factors on cardiometabolic health. During the past two decades, metabolomic approaches have been increasingly used in various epidemiological studies, primarily in Western populations. Although the field is still in its early stages, some studies have tried to identify novel compounds or confirm their metabolites and associations with cardiometabolic diseases in Chinese populations, including amino acids, fatty acids, acylcarnitines and other metabolites. Despite major efforts to discover novel biomarkers for disease prediction or intervention, the limits in current study design, analytical platforms, and data processing approaches are challenges in metabolomic research worldwide. Therefore, future research with more advanced technologies, rigorous study designs, standardized detection and analytic approaches, and integrated data from multiomics approaches are essential to evaluate the feasibility of using metabolomics in clinical settings. Finally, the functional roles and underlying biological mechanisms of metabolomic biomarkers should be elucidated by future mechanistic research.

Metabolomic correlates of central adiposity and earlier-life body mass index

BMI is correlated with circulating metabolites, but few studies discuss other adiposity measures, and little is known about metabolomic correlates of BMI from early life. We investigated associations between different adiposity measures, BMI from childhood through adulthood, and metabolites quantified from serum using ¹H NMR spectroscopy in 900 British men and women aged 60-64. We assessed BMI, waist-to-hip ratio (WHR), android-to-gynoid fat ratio (AGR), and BMI from childhood through adulthood. Linear regression with Bonferroni adjustment was performed to assess adiposity and metabolites. Of 233 metabolites, 168; 126; and 133 were associated with BMI, WHR, and AGR at age 60-64, respectively. Associations were strongest for HDL, particularly HDL particle size-e.g., there was 0.08 SD decrease in HDL diameter (95% CI: 0.07-0.10) with each unit increase in BMI. BMI-adjusted AGR or WHR were associated with 31 metabolites where there was no metabolome-wide association with BMI. We identified inverse associations between BMI at age 7 and glucose or glycoprotein at age 60-64 and relatively large LDL cholesteryl ester with postadolescent BMI gains. In summary, we identified metabolomic correlates of central adiposity and earlier-life BMI. These findings support opportunities to leverage metabolomics in early prevention of cardiovascular risk attributable to body fatness.

Review of recent developments in GC-MS approaches to metabolomics-based research

BACKGROUND

Metabolomics aims to identify the changes in endogenous metabolites of biological systems in response to intrinsic and extrinsic factors. This is accomplished through untargeted, semi-targeted and targeted based approaches. Untargeted and semi-targeted methods are typically applied in hypothesis-generating investigations (aimed at measuring as many metabolites as possible), while targeted approaches analyze a relatively smaller subset of biochemically important and relevant metabolites. Regardless of approach, it is well recognized amongst the metabolomics community that gas chromatography-mass spectrometry (GC-MS) is one of the most efficient, reproducible and well used analytical platforms for metabolomics research. This is due to the robust, reproducible and selective nature of the technique, as well as the large number of well-established libraries of both commercial and 'in house' metabolite databases available.

AIM OF REVIEW

This review provides an overview of developments in GC-MS based metabolomics applications, with a focus on sample preparation and preservation techniques. A number of chemical derivatization (in-time, in-liner, offline and microwave assisted) techniques are also discussed. Electron impact ionization and a summary of alternate mass analyzers are highlighted, along with a number of recently reported new GC columns suited for metabolomics. Lastly, multidimensional GC-MS and its application in environmental and biomedical research is presented, along with the importance of bioinformatics.

KEY SCIENTIFIC CONCEPTS OF REVIEW

The purpose of this review is to both highlight and provide an update on GC-MS analytical techniques that are common in metabolomics studies. Specific emphasis is given to the key steps within the GC-MS workflow that those new to this field need to be aware of and the common pitfalls that should be looked out for when starting in this area.

Metabolomic Markers of Phthalate Exposure in Plasma and Urine of Pregnant Women

Phthalates are known endocrine disruptors and found in almost all people with several associated adverse health outcomes reported in humans and animal models. Limited data are available on the relationship between exposure to endocrine disrupting chemicals and the human metabolome. We examined the relationship of metabolomic profiles in plasma and urine of 115 pregnant women with eleven urine phthalate metabolites measured at 26 weeks of gestation to identify potential biomarkers and relevant pathways. Targeted metabolomics was performed by selected reaction monitoring liquid chromatography and triple quadrupole mass spectrometry to measure 415 metabolites in plasma and 151 metabolites in urine samples. We have chosen metabolites with the best defined peaks for more detailed analysis (138 in plasma and 40 in urine). Relationship between urine phthalate metabolites and concurrent metabolomic markers in plasma and urine suggested potential involvement of diverse pathways including lipid, steroid, and nucleic acid metabolism and enhanced inflammatory response. Most of the correlations were positive for both urine and plasma, and further confirmed by regression and PCA analysis. However, after the FDR adjustment for multiple comparisons, only 9 urine associations remained statistically significant (q-values 0.0001–0.0451), including Nicotinamide mononucleotide, Cysteine T2, Cystine, and L-Aspartic acid. Additionally, we found negative associations of maternal pre-pregnancy body mass index (BMI) with more than 20 metabolomic markers related to lipid and amino-acid metabolism and inflammation pathways in plasma (p = 0.01–0.0004), while Mevalonic acid was positively associated (p = 0.009). Nicotinic acid, the only significant metabolite in urine, had a positive association with maternal BMI (p = 0.002). In summary, when evaluated in the context of metabolic pathways, the findings suggest enhanced lipid biogenesis, inflammation and altered nucleic acid metabolism in association with higher phthalate levels. These results provide new insights into the relationship between phthalates, common in most human populations, and metabolomics, a novel approach to exposure and health biomonitoring.

Characterization of missing values in untargeted MS-based metabolomics data and evaluation of missing data handling strategies

BACKGROUND

Untargeted mass spectrometry (MS)-based metabolomics data often contain missing values that reduce statistical power and can introduce bias in biomedical studies. However, a systematic assessment of the various sources of missing values and strategies to handle these data has received little attention. Missing data can occur systematically, e.g. from run day-dependent effects due to limits of detection (LOD); or it can be random as, for instance, a consequence of sample preparation.

METHODS

We investigated patterns of missing data in an MS-based metabolomics experiment of serum samples from the German KORA F4 cohort (n = 1750). We then evaluated 31 imputation methods in a simulation framework and biologically validated the results by applying all imputation approaches to real metabolomics data. We examined the ability of each method to reconstruct biochemical pathways from data-driven correlation networks, and the ability of the method to increase statistical power while preserving the strength of established metabolic quantitative trait loci.

RESULTS

Run day-dependent LOD-based missing data accounts for most missing values in the metabolomics dataset. Although multiple imputation by chained equations performed well in many scenarios, it is computationally and statistically challenging. K-nearest neighbors (KNN) imputation on observations with variable pre-selection showed robust performance across all evaluation schemes and is computationally more tractable.

CONCLUSION

Missing data in untargeted MS-based metabolomics data occur for various reasons. Based on our results, we recommend that KNN-based imputation is performed on observations with variable pre-selection since it showed robust results in all evaluation schemes.

Glucose challenge metabolomics implicates medium-chain acylcarnitines in insulin resistance

Insulin resistance (IR) predisposes to type 2 diabetes and cardiovascular disease but its causes are incompletely understood. Metabolic challenges like the oral glucose tolerance test (OGTT) can reveal pathogenic mechanisms. We aimed to discover associations of IR with metabolite trajectories during OGTT. In 470 non-diabetic men (age 70.6 ± 0.6 years), plasma samples obtained at 0, 30 and 120 minutes during an OGTT were analyzed by untargeted liquid chromatography-mass spectrometry metabolomics. IR was assessed with the hyperinsulinemic-euglycemic clamp method. We applied age-adjusted linear regression to identify metabolites whose concentration change was related to IR. Nine trajectories, including monounsaturated fatty acids, lysophosphatidylethanolamines and a bile acid, were significantly associated with IR, with the strongest associations observed for medium-chain acylcarnitines C10 and C12, and no associations with L-carnitine or C2-, C8-, C14- or C16-carnitine. Concentrations of C10- and C12-carnitine decreased during OGTT with a blunted decline in participants with worse insulin resistance. Associations persisted after adjustment for obesity, fasting insulin and fasting glucose. In mouse 3T3-L1 adipocytes exposed to different acylcarnitines, we observed blunted insulin-stimulated glucose uptake after treatment with C10- or C12-carnitine. In conclusion, our results identify medium-chain acylcarnitines as possible contributors to IR.

Machine learning approaches to the social determinants of health in the health and retirement study

BACKGROUND

Social and economic factors are important predictors of health and of recognized importance for health systems. However, machine learning, used elsewhere in the biomedical literature, has not been extensively applied to study relationships between society and health. We investigate how machine learning may add to our understanding of social determinants of health using data from the Health and Retirement Study.

METHODS

A linear regression of age and gender, and a parsimonious theory-based regression additionally incorporating income, wealth, and education, were used to predict systolic blood pressure, body mass index, waist circumference, and telomere length. Prediction, fit, and interpretability were compared across four machine learning methods: linear regression, penalized regressions, random forests, and neural networks.

RESULTS

All models had poor out-of-sample prediction. Most machine learning models performed similarly to the simpler models. However, neural networks greatly outperformed the three other methods. Neural networks also had good fit to the data (R2 between 0.4-0.6, versus <0.3 for all others). Across machine learning models, nine variables were frequently selected or highly weighted as predictors: dental visits, current smoking, self-rated health, serial-seven subtractions, probability of receiving an inheritance, probability of leaving an inheritance of at least $10,000, number of children ever born, African-American race, and gender.

DISCUSSION

Some of the machine learning methods do not improve prediction or fit beyond simpler models, however, neural networks performed well. The predictors identified across models suggest underlying social factors that are important predictors of biological indicators of chronic disease, and that the non-linear and interactive relationships between variables fundamental to the neural network approach may be important to consider.

A Serum Small Molecule Biosignature of Radiation Exposure from Total Body Irradiated Patients

The potential for radiological accidents and nuclear terrorism has increased the need for the development of new rapid biodosimetry methods. In addition, in a clinical setting the issue of an individual's radiosensitivity should be taken into consideration during radiotherapy. We utilized metabolomics and lipidomics to investigate changes of metabolites in serum samples following exposure to total body ionizing radiation in humans. Serum was collected prior to irradiation, at 3-8 h after a single dose of 1.25-2 Gy, and at 24 h with a total delivered dose of 2-3.75 Gy. Metabolomics revealed perturbations in glycerophosphocholine, phenylalanine, ubiquinone Q2, and oxalic acid. Alterations were observed in circulating levels of lipids from monoacylglycerol, triacylglycerol, phosphatidylcholine, and phosphatidylglycerol lipid classes. Polyunsaturated fatty acids were some of the most dysregulated lipids, with increased levels linked to proinflammatory processes. A targeted metabolomics approach for eicosanoids was also employed. The results showed a rapid response for proinflammatory eicosanoids, with a dampening of the signal at the later time point. Sex differences were observed in the markers from the untargeted approach but not the targeted method. The ability to identify and quantify small molecules in blood can therefore be utilized to monitor radiation exposure in human populations.

Exploration of potential biomarkers and related biological pathways for PCB exposure in maternal and cord serum: A pilot birth cohort study in Chiba, Japan

Polychlorinated biphenyls (PCBs) have been associated with adverse human reproductive and fetal developmental measures or outcomes because of their endocrine-disrupting effects; however, the biological mechanisms of adverse effects of PCB exposure in humans are not currently well established. In this study, we aimed to identify the biological pathways and potential biomarkers of PCB exposure in maternal and umbilical cord serum using a hydrophilic interaction chromatography-tandem mass spectrometry (HILIC-MS/MS) metabolomics platform. The median concentration of total PCBs in maternal (n=93) and cord serum (n=93) were 350 and 70pgg^-1 wet wt, respectively. PCB levels in maternal and fetal serum from the Chiba Study of Mother and Children's Health (C-MACH) cohort are comparable to those of earlier cohort studies conducted in Japan, the USA, and European countries. We used the random forest model with the metabolome profile to predict exposure levels of PCB (first quartile [Q1] and fourth quartile [Q4]) for pregnant women and fetuses. In the prediction model for classification of Q1 versus Q4 (area-under-curve [AUC]: pregnant women=0.812 and fetuses=0.919), citraconic acid level in maternal serum and ethanolamine, p-hydroxybenzoate, and purine levels in cord serum had >0.70 AUC values. These candidate biomarkers and metabolite included in composited models were related to glutathione and amino acid metabolism in maternal serum and the amino acid metabolism and ubiquinone and other terpenoid-quinone biosynthesis in cord serum (FDR <0.10), indicating disruption of metabolic pathways by PCB exposure in pregnant women and fetuses. These results showed that metabolome analysis might be useful to explore potential biomarkers and related biological pathways for PCB exposure. Thus, more detailed studies are needed to verify sensitivity of the biomarkers and clarify the biochemical changes resulting from PCB exposure.

Frontiers of high-throughput metabolomics

Large scale metabolomics studies are increasingly used to investigate genetically different individuals and time-dependent responses to environmental stimuli. New mass spectrometric approaches with at least an order of magnitude more rapid analysis of small molecules within the cell's metabolome are now paving the way towards true high-throughput metabolomics, opening new opportunities in systems biology, functional genomics, drug discovery, and personalized medicine. Here we discuss the impact and advantages of the progress made in profiling large cohorts and dynamic systems with high temporal resolution and automated sampling. In both areas, high-throughput metabolomics is gaining traction because it can generate hypotheses on molecular mechanisms and metabolic regulation. We conclude with the current status of the less mature single cell analyses where high-throughput analytics will be indispensable to resolve metabolic heterogeneity in populations and compartmentalization of metabolites.

Identifying biomarkers for asthma diagnosis using targeted metabolomics approaches

BACKGROUND

The diagnosis of asthma in children is challenging and relies on a combination of clinical factors and biomarkers including methacholine challenge, lung function, bronchodilator responsiveness, and presence of airway inflammation. No single test is diagnostic. We sought to identify a pattern of inflammatory biomarkers that was unique to asthma using a targeted metabolomics approach combined with data science methods.

METHODS

We conducted a nested case-control study of 100 children living in a peri-urban community in Lima, Peru. We defined cases as children with current asthma, and controls as children with no prior history of asthma and normal lung function. We further categorized enrollment following a factorial design to enroll equal numbers of children as either overweight or not. We obtained a fasting venous blood sample to characterize a comprehensive panel of targeted markers using a metabolomics approach based on high performance liquid chromatography-mass spectrometry.

RESULTS

A statistical comparison of targeted metabolites between children with asthma (n = 50) and healthy controls (n = 49) revealed distinct patterns in relative concentrations of several metabolites: children with asthma had approximately 40-50% lower relative concentrations of ascorbic acid, 2-isopropylmalic acid, shikimate-3-phosphate, and 6-phospho-d-gluconate when compared to children without asthma, and 70% lower relative concentrations of reduced glutathione (all p < 0.001 after Bonferroni correction). Moreover, a combination of 2-isopropylmalic acid and betaine strongly discriminated between children with asthma (2-isopropylmalic acid ≤ 13 077 normalized counts/second) and controls (2-isopropylmalic acid > 13 077 normalized counts/second and betaine ≤ 16 47 121 normalized counts/second).

CONCLUSIONS

By using a metabolomics approach applied to serum, we were able to discriminate between children with and without asthma by revealing different metabolic patterns. These results suggest that serum metabolomics may represent a diagnostic tool for asthma and may be helpful for distinguishing asthma phenotypes.