Next Article in Journal
Metabolomic Profiling of Aqueous Humor from Pathological Myopia Patients with Choroidal Neovascularization
Next Article in Special Issue
Serum Metabolomic and Lipidomic Profiling Reveals the Signature for Postoperative Obesity among Adult-Onset Craniopharyngioma
Previous Article in Journal
Pharmacokinetic and Pharmacodynamic Drug–Drug Interactions: Research Methods and Applications
Previous Article in Special Issue
Exercise Metabolome: Insights for Health and Performance
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
Journals
Metabolites
Volume 13
Issue 8
10.3390/metabo13080899
metabolites-logo
Submit to this Journal Review for this Journal Propose a Special Issue
► Article Menu

Article Menu

share Share announcement Help format_quote Cite question_answer Discuss in SciProfiles thumb_up
...
Endorse textsms
...
Comment
first_page
settings
Order Article Reprints
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Untargeted Metabolomics and Body Mass in Adolescents: A Cross-Sectional and Longitudinal Analysis

by
Amarnath Singh
1,
Garrett Kinnebrew
2,
Ping-Ching Hsu
3,
Daniel Y. Weng
1,
Min-Ae Song
4,
Sarah A. Reisinger
5,
Joseph P. McElroy
6,
Brittney Keller-Hamilton
5,7,
Amy K. Ferketich
4,
Jo L. Freudenheim
8 and
Peter G. Shields
1,*
1
Comprehensive Cancer Center, The Ohio State University, Columbus, OH 43210-1240, USA
2
Department of Biomedical Informatics, Biomedical Informatics Shared Resources (BISR), The Ohio State University, Columbus, OH 43210-1240, USA
3
Department of Environmental Health Sciences, Fay W. Boozman College of Public Health, University of Arkansas for Medical Sciences, Little Rock, AR 72205, USA
4
College of Public Health, The Ohio State University, Columbus, OH 43210-1240, USA
5
Center for Tobacco Research, Comprehensive Cancer Center, The Ohio State University, Columbus, OH 43210-1240, USA
6
Center for Biostatistics, Department of Biomedical Informatics, The Ohio State University, Columbus, OH 43210-1240, USA
7
Department of Internal Medicine, College of Medicine, The Ohio State University, Columbus, OH 43210-1240, USA
8
Department of Epidemiology and Environmental Health, University at Buffalo, Buffalo, NY 14214, USA
*
Author to whom correspondence should be addressed.
Metabolites 2023, 13(8), 899; https://doi.org/10.3390/metabo13080899
Submission received: 7 June 2023 / Revised: 25 July 2023 / Accepted: 27 July 2023 / Published: 30 July 2023
(This article belongs to the Special Issue Metabolomics in the Interventions of Metabolic Diseases and Physical Activity)
Browse Figures
Versions Notes

Abstract

:
Obesity in children and adolescents has increased globally. Increased body mass index (BMI) during adolescence carries significant long-term adverse health outcomes, including chronic diseases such as cardiovascular disease, stroke, diabetes, and cancer. Little is known about the metabolic consequences of changes in BMI in adolescents outside of typical clinical parameters. Here, we used untargeted metabolomics to assess changing BMI in male adolescents. Untargeted metabolomic profiling was performed on urine samples from 360 adolescents using UPLC–QTOF-MS. The study includes a baseline of 235 subjects in a discovery set and 125 subjects in a validation set. Of them, a follow-up of 81 subjects (1 year later) as a replication set was studied. Linear regression analysis models were used to estimate the associations of metabolic features with BMI z-score in the discovery and validation sets, after adjusting for age, race, and total energy intake (kcal) at false-discovery-rate correction (FDR) ≤ 0.1. We identified 221 and 16 significant metabolic features in the discovery and in the validation set, respectively. The metabolites associated with BMI z-score in validation sets are glycylproline, citrulline, 4-vinylsyringol, 3′-sialyllactose, estrone sulfate, carnosine, formiminoglutamic acid, 4-hydroxyproline, hydroxyprolyl-asparagine, 2-hexenoylcarnitine, L-glutamine, inosine, N-(2-Hydroxyphenyl) acetamide glucuronide, and galactosylhydroxylysine. Of those 16 features, 9 significant metabolic features were associated with a positive change in BMI in the replication set 1 year later. Histidine and arginine metabolism were the most affected metabolic pathways. Our findings suggest that obesity and its metabolic outcomes in the urine metabolome of children are linked to altered amino acids, lipid, and carbohydrate metabolism. These identified metabolites may serve as biomarkers and aid in the investigation of obesity’s underlying pathological mechanisms. Whether these features are associated with the development of obesity, or a consequence of changing BMI, requires further study.

Graphical Abstract

1. Introduction

More than 379 million children and adolescents worldwide are overweight or obese [1]. In the United States from 2017 to 2020, the prevalence of obesity was 19.7% in the age group of 2–19 years and affected about 14.7 million children and adolescents [2]. Up to 80% of obese children become obese adults, indicating that childhood obesity typically extends into adulthood [3]. Obesity is a common, serious, and multifactorial disease associated with genetics [4], environment [5], physiology [6], increased consumption of hypercaloric foods [7], and sedentary lifestyles [8]. Increased body mass index (BMI) and obesity during adolescence carry significant long-term adverse health outcomes, including the development of chronic diseases and mortality, such as cardiovascular disease, stroke, insulin resistance diabetes, and cancer [9,10,11,12,13]. Therefore, efficient interventions and predictive biomarkers for the development of obesity are needed. As children grow, the assessment of body mass index (BMI) requires adjustment for age and gender [14]. While a study of BMI typically categorizes adolescents as normal, overweight, or obese based on BMI percentile [15,16], a better measure of body mass index is the BMI z-score as a continuous variable because the BMI z-score is age and gender-specific [17]. In this study, we used the BMI z-score to investigate the association between the urinary metabolome and obesity.
Untargeted metabolomics profiles low-molecular-weight metabolites (<1500 Da) in biospecimens to elucidate cell physiology and disease mechanisms, and to identify biomarkers of disease risk [18,19]. This approach helps identify the molecular mechanisms of complex diseases, as well as for disease monitoring and risk assessment. In the past, targeted and untargeted metabolomics studies have been used to study the metabolome signature of obesity in response to dietary intake [20], the effects of specific dietary patterns or weight loss [21,22] or gain interventions, and body fat (%) [23].
In our comprehensive literature review, we identified 41 studies on adolescent metabolomics and obesity [10]. Out of the total studies, 33 specifically focused on blood (plasma, serum), 3 studies examined umbilical cord blood, and 1 study used saliva samples, while 6 studies utilized urine samples [10,24,25,26,27,28,29]. Urine contains diverse metabolites reflecting the overall metabolic status of an individual, offering a comprehensive view of obesity-related changes including the excretion of metabolites derived from various biological processes [30]. Unlike plasma and serum, urine is less affected by factors such as diet, medication, or circadian rhythms [31]. Metabolomics studies in adolescent populations are still limited. To gain a deeper understanding of obesity’s development, it is essential to validate previous research and examine metabolic changes in obese children who have not yet manifested disease symptoms.
Here, we examined the association of adolescent urinary metabolic signature with BMI z-scores to investigate the mechanisms of progression of childhood obesity at a metabolite level, which may aid in identifying adverse effects of obesity. This is the largest study of untargeted metabolomics in adolescents and is strengthened by longitudinal assessment.

2. Materials and Methods

2.1. Study Recruitment and Design

A total of 1220 male youths were enrolled in the Buckeye Teen Health Study (BTH) from January 2015 to June 2016, a longitudinal cohort study focusing on lifestyle and tobacco use behaviors [32,33,34,35,36,37,38]. Out of this, a total of 360 subjects additionally provided urine samples as a “biomarker cohort” based on age, county, and date of sample collection (Figure 1). Eligible male youths were aged 11 to 16 years and lived in either urban Franklin County, Ohio, or 1 of 9 Appalachian Ohio counties. One year follow-up metabolomics study, referred to as a replication set, was conducted with eighty-one participants from the biomarker cohort (Figure 1). All participants in this study were informed in writing regarding collecting their samples for research aims and given the right to refuse such uses. Male youths provided consent, and their parents provided permission for them to enroll in the study. Exclusion criteria included any hearing or vision impairments or the inability to read and speak English. The baseline and one-year follow-up sessions were completed in person at participants’ homes or a mutually agreed-upon public location. The anthropometric measurements and data collection was conducted by a trained interviewer who resided in the same region as the participants. Data collected include demographic, family, and socioeconomic characteristics, as well as anthropometric measurements, such as height, body weight, and dietary records. BMI (kg/m2) or BMI percentile, and BMI z-score (https://zscore.research.chop.edu/calcbmi.php, accessed on 22 May 2022) were calculated based on the measured height and weight data. The questionnaire included both interviewer-administered and audio-administered items, depending on the level of sensitivity of the item to provide the boy’s privacy when their parents were around. Participants listened to these questions on a headset and responded on the computer without the assistance of the interviewer. The study protocol was approved by the Ohio State University Clinical Scientific Research Committee and Institutional Review Board (2014C0030).

2.2. Urine Sample Collection

Each study subject submitted a random single-void urine sample for analysis in a sterile container. Urine samples collected were shipped using an ice pack after freezing to OSU. Samples were stored at −80 °C for long-term storage until further analysis.

2.3. Reagents and Chemicals

All reagents and solvents were of HPLC grade. Formic acid, acetonitrile (ACN), 4-nitrobenzoic acid (4-NBA), and 13 C-labelled phenylalanine were purchased from Sigma-Aldrich (St. Louis, MO, USA) and water was purchased from Fisher Optima grade (Fisher Scientific, Waltham, MA, USA).

2.4. Urine Sample Preparation for Metabolomics

Urine was thawed at room temperature for 15 min and centrifuged for 10 min at 15,000 rpm at 4 °C. For positive mode sample analysis, the supernatant of the urine (20 µL) sample was diluted with water (180 µL) containing formic acid (0.1%) and 13 C-labelled phenylalanine (1.0 µM) as internal standard. For negative mode sample analysis, the supernatant of the urine (20 µL) sample was diluted with water (180 µL) containing formic acid (0.1%) and 4-NBA (2.0 µM) as internal standard. The final solution was vortexed for 30 s, transferred into HPLC vials, and placed into an autosampler tray for analysis.

2.5. UPLC–QTOF-MS Analysis

The metabolite separations of the urine sample were obtained on the Agilent 1290 Infinity Quaternary LC System (Agilent Technologies, Santa Clara, United States) using the ACQUITY UPLC HSS T3 column (2.1 × 100 mm, 1.8 μm). The injection volume was 1 μL, followed by a standard needle wash. The mobile phase (A) consisted of 100% water (H2O) with 0.1% formic acid and a mobile phase (B) consisted of 100% ACN with 0.1% formic acid with the following gradient elution. The flow rate was set as 0.5 mL/min and the gradient consisted of 100% A; 0–1.5 min, 0% B; 1.5–7.5 min, 50% B; 7.5–8.5 min, 95% B; 8.5–10 min, 95% B; 10–10.1 min, 0% B; and 10.10–12 min, 0% B. The column eluent (1 µL) was introduced directly into the mass spectrometer by electrospray. The autosampler tray temperature was set to 4 °C and the column temperature was 40 °C. The metabolic profiling analysis of the urine sample was conducted on an Agilent 6550 iFunnel Q-TOF LC/MS (Agilent Technologies, USA) with Dual Agilent Jet Stream Electrospray Ionization (Agilent Technologies), and its parameter was set as follows: for positive mode, dry gas temperature, 150 °C; dry gas flow, 18 L min−1; nebulizer pressure, 30 psig; sheath gas temperature, 300 °C; and sheath gas flow, 12 L min−1, and for negative mode, the instrument parameter settings are dry gas temperature, 200 °C; dry gas flow, 18 L min−1; nebulizer pressure, 35 psig; sheath gas temperature, 320 °C; and sheath gas flow, 12 L min−1. Mass spectrometry was performed in both positive-ion (ESI+) or negative-ion (ESI−) electrospray ionization mode with a capillary voltage of 4000 V and a sampling cone voltage of 65 V in both negative mode and positive mode. The scan range was adjusted to centroid mode using a scan rate of 3.00 spectra/second, and a mass range of 50–1700 m/z.
A quality control (QC) sample was prepared by mixing an equal volume of a pooled urine sample of subjects containing 4-NBA (2.0 µM) and 13 C-labelled phenylalanine (1.0 µM) as internal standard, which was then aliquoted into small vials. The pooled QC samples were used to condition the analytical platform at the beginning of the run, placed as every tenth sample, and analyzed periodically after 10 sample runs to check the performance of the analytical system in terms of retention times, accurate mass measurements, and signal intensities.

2.6. Data Analysis

After completing the metabolomics run on the instrument, we processed the raw data from both the discovery and validation set together. The raw data (.d file) from the ultra-performance liquid chromatography-quadrupole-time-of-flight mass spectrometry (UPLC-QTOF-MS) instrument were converted to mzML files using the MS convert program from ProteoWizard (https://proteowizard.sourceforge.io/, accessed on 22 May 2022). In the resulting mzML files, intensity and m/z values were stored as 32-bit floating points with zlib compression, and the “vendor peak picking” option was selected to convert data to centroid mode. The mzML files were imported into the R statistical computing environment using the R package XCMS [39]. XCMS was used for data processing including retention time alignment, peak detection, peak grouping, and peak filling to obtain a sample-feature matrix. Retention time alignment was performed using the ObiWarp algorithm. Each sample was aligned against the pooled QC sample in both EIC+ and EIC− modes. Peak detection was performed using the CentWave algorithm (parameters: peak width = 2–30 s, signal-to-noise ratio = 3, mzdiff = −0.005, integration method = 1, prefilter = 3 peaks, 10 intensity). Peak grouping was performed using the peak density method (parameters: minFraction = 0, minSamples = 2, bw = 5, binSize = 0.015). Integration of signal from peaks missing after grouping peaks across samples was performed with the fillChromPeaks function and default parameters (expandMz = 0, expandRt = 0, ppm = 0, fixedMz = 0, fixedRt = 0). The resulting data contained a set of features corresponding to peaks which were found in multiple samples each with an approximate m/z and retention time. The matrix of intensity values by feature and sample was used for further analysis. During the profiling, the analytical robustness of UPLC-QTOF-MS QC samples was repeatedly analyzed. Metabolite features that were missing in more than one pooled QC sample or showing a large coefficient of variation (CV) in the pooled QC samples greater than 30% were discarded as unreliable before conducting the statistical analysis. Intensity values were normalized by the mass spectrometry total usable signal (MSTUS). Briefly, an MSTUS value was calculated by summing the intensity of all features, excluding those features which were either found in less than half of the samples, in the bottom 20th percentile of peak intensities, in more than 80% of samples, or were in the 90th percentile in at least one but less than 10% of samples. Feature intensities for each sample were then multiplied by a scaling factor calculated as the median MSTUS value over the sample MSTUS value. Missing feature intensities after peak filling were then imputed as half of the minimum value by feature [40].

2.7. Statistical Analysis

Before metabolomics analyses, patients were randomized into a modeling set for the discovery (n = 235), validation (n = 125), and one-year follow-up as replication sets (n = 81) (Figure 1).
JMP Pro 15 (100 SAS Campus Drive Cary, NC, USA) was used for statistical analysis. Data of normal distribution are expressed as mean ± SD. Measurement data of non-normal distribution are expressed as median (interquartile range). Before statistical analyses, urinary metabolomics peak intensity data were log10-transformed. Preliminarily annotated metabolites present in both cohorts were analyzed. The Partek Genomics suite was used for the principal component analysis [41]. Linear regression analysis models were used to estimate the associations of metabolic features with a BMI z-score in the discovery and validation set, after adjusting for age, race, and total energy intake (kcal). In the present study, smoking status was less than 4.2%, and smoking status and socioeconomic status were initially included in the models but were removed due to the lack of significance of the effect (Supplement Tables S4 and S5).
After pooled QC filtering of CV < 30% and performing putative annotation, a total of 1532 metabolite features were obtained in both positive and negative modes. These features were then subjected to linear regression analysis in the discovery set. Subsequently, 221 features that exhibited significant association with BMI Z-score in the discovery set were selected for further investigation of their association with BMI Z-score in the validation set.
Correction for multiple comparisons was performed using a false discovery rate (FDR) [42] corrected alpha of ≤0.1. Graph Pad Prism version 8.4 was used to produce the volcano plots. To identify pathways that may have been perturbated in adolescent children, metabolic pathway analysis was performed using Metaboanalyst 5.0 [43,44] (http://www.metaboanalyst.ca/). Metabolites that showed an association with BMI Z-score in the discovery set and validation set were queried against the human metabolome database (HMDB, www.hmdb.ca, accessed on 22 May 2022). The list of putative metabolites generated by HMDB was imported into Metaboanalyst and mapped based on the Kyoto Encyclopedia of Genes and Genomes (KEGG). Pathway enrichment ratio values > 1.0 and a p-value < 0.05 was considered to define a perturbed pathway. The top 25 pathways were reported as the most perturbated pathways.

3. Results

3.1. Characteristics of Study Subjects

The baseline characteristics of the discovery, validation set, and replication set study participants are shown in Table 1. In total, 360 study participants were recruited into the BTH study for urine metabolomics, including 235 in the discovery set and 125 in the validation set. The average age (mean ± SD) of the study participant was 14.80 ± 1.42 years, 14.84 ± 1.31 years, and 16.08 ± 1.20 years in the discovery set, validation set, and replication set study participants, respectively. The age and height of the study participants in the discovery and validation sets were similar, but the weight of the study participants in the discovery (158.50 lb.) and validation set (149.85 lb.) slightly differed. As expected, we observed an increase in weight (170.64 lb.) in the replication set study participants. The average BMI (Kg/m2) of the study participants in the discovery set, validation set, and replication set study were 24.64 ± 7.13, 23.28 ± 6.01, and 25.16 ± 6.91, respectively. At baseline, 31.49 and 22.40% of the study participants were obese (Table 1).

3.2. Untargeted Metabolic Profiling of Urine by UPLC-QTOF-MS

In the UPLC-QTOF-MS dataset, a total of 4596 features were detected in the positive mode (2537) and negative mode (2059) after pooled QC filtering of CV < 30%. Further, after annotation using the Human Metabolome Database (HMDB) (https://hmdb.ca/spectra/ms/search), we obtained that 1532 metabolite features in both positive mode and negative mode were used for statistical analysis. The overall composition of the distributions of these metabolites via principal component analysis found complex underlying correlations; in particular, the first principal component explained 17.7% of the total variation in the levels of these 1532 metabolites (Supplementary Figures S1 and S2A,B).

3.3. Association of Metabolites with BMI Z-Score in Discovery and Validation Set, and Changes in BMI at the 1-Year Follow-Up

In the discovery set, a total of 221 features were significantly associated with BMI z-score after adjustment for covariables (age, race, and total energy intake) and multiple comparisons (fdr ≤ 0.1). The estimates of the association of metabolites with BMI z-score and their corresponding p-value after multiple test corrections are displayed in Figure 2A, Supplement Table S1, and the top 30 are displayed in Table 2. Out of 221 significant features associated with BMI z-score, 32 metabolites are negatively associated with BMI z-score, and 189 metabolites were positively associated with BMI z-score (Supplement Table S1).
These 221 putatively identified metabolites in discovery sets refer to 11 chemical taxonomy super-class and 28 sub-class families (Supplemental Table S1). Out of these 11 super-class taxonomies, the most perturbated super-class of compounds belonged to organic acids and their derivatives (amino acids, peptides, and analogues), lipids and lipid-like molecules, organic oxygen compounds carbohydrates, and carbohydrate conjugates), organoheterocyclic compounds (pyridines, indol, imidazoles, indolyl carboxylic acids, and derivatives), nucleosides, nucleotides, and analogues (Figure 3, Supplement Table S1).
In the validation set analysis, we assessed 221 significant metabolites from the discovery set, and after controlling for covariates and multiple comparisons, we found that 16 of the 221 metabolites remained significantly associated with BMI z-score in the validation set (FDR = 0.1) (Figure 2B, Table 3, Supplement Table S1). Among 16 significant metabolites associated with a BMI Z-score in the validation set, 8 metabolites were negatively associated and 8 were positively associated (Figure 2B, Table 3). The positive change in BMI in the replication set (1 year of follow-up) indicated that among the 16 replicated in the validation set, 9 were also found to be correlated with the positive change in BMI (Table 4).

3.4. Differential Mapping of Metabolites in Pathway Analysis

To explore the metabolic pathways that potentially contribute to obesity progression in adolescent children, we carried out a metabolomic pathway enrichment analysis using MetPA (www.metaboanalyst.ca) with 221 significant features from the discovery set and 16 significant features from the validation set. p-values < 0.05 and enrichment ratios > 0.1 were considered to indicate highly influential pathways. For the discovery set, pathway enrichment results showed that 25 metabolic pathways were identified in KEGG. The highest enrichment ratios were obtained for the histidine metabolism pathways, beta-Alanine metabolism, aminoacyl-tRNA biosynthesis, arginine biosynthesis, arginine and proline metabolism, and pyrimidine metabolism are the most enriched metabolic pathways, as shown in Figure 4A, and detailed pathway enrichment results are displayed in Supplement Table S2. Similarly, in the validation set, arginine biosynthesis, histidine metabolism, D-glutamine and D-glutamate metabolism, and nitrogen metabolism were the most significantly enriched metabolic pathways (Figure 4B and Supplement Table S3). In both instances of pathway enrichment analysis, arginine biosynthesis and histidine metabolism were the most enriched metabolic pathways.

4. Discussion

The purpose of the study was to identify the metabolomic signature of obesity in adolescents. Prior studies in the field of metabolomics have predominantly focused on characterizing metabolites linked to childhood obesity or BMI in contrast to non-obese subjects [27,45,46,47,48]. Obesity mainly classifies adolescents as normal, overweight, or obese based on their BMI percentile [15,16]. As children grow up, the measurement of body mass index (BMI) requires adjustment for age and gender [14]. However, the BMI z-score as a continuous variable, which is age and gender-specific in children, is considered a more suitable metric for determining body mass index (BMI) than the conventional BMI measurement and provides a better estimation of an individual’s adiposity status [17].
The present cross-sectional study demonstrated the perturbation of the urinary metabolites associated with BMI z-score in male adolescents and replicated results in the longitudinal analysis for changes in BMI. Additionally, metabolomic signatures were identified with the potential to examine metabolic health in adolescents. To our knowledge, this is the first longitudinal study to relate positive change in BMI to the urine metabolome in adolescents. In this study, 1532 putatively annotated features were used to study the association analysis with BMI Z-score in adolescents. Following the adjustment of age, race, and total energy intake, 221 metabolites were identified as exhibiting a robust association with BMI z-score within the discovery set. Out of 221 lists of significant features from the discovery set, 16 metabolites were found to be significantly associated with the BMI z-score in the validation dataset in adolescents. Obesity is a serious and growing health problem that affects people of all ages all over the world, including children. In the past, research on metabolomic biomarkers in obesity has largely focused on targeted metabolomics and a specific group of compounds rather than evaluating a large number of metabolites at the same time [49,50,51,52]. However, developing evidence indicates that not only absolute metabolite levels of specific compounds are important, but their relationship with other metabolites (profiles) and pathways play an important role in the biology of metabolism [53]. Therefore, it is crucial to simultaneously investigate a larger number of metabolites to arrive at a more accurate etiologic picture. Our findings highlight the presence of changes in the urine metabolome associated with positive changes in BMI in adolescence. We observed that the six urinary metabolites (3′-sialyllactose, formiminoglutamic acid, 4-hydroxyproline, citrulline, inosine) showed a positive association with positive change in BMI and three metabolites (glycylproline, 4-vinylsyringol) showed a negative association with positive change in BMI. Accumulated evidence showed that the perturbated metabolites are amino acids and carbohydrates.
In several adolescent studies, the relationship between the metabolome and obesity has been thoroughly described [1,17,21,27]. The present study's finding is consistent with previous findings and most of the metabolites strongly influencing metabolic BMI are documented in the present study. In the study by Cirulli et al. (2019), which included 1969 individuals from the TwinsUK cohort, 49 metabolites showed the strongest associations with BMI out of a total of 650 metabolites, and a majority of the 49 metabolites were identified as significant predictors of BMI. These included glutamate, asparagine, leucine, N2,N2-dimethylguanosine, and kynurenate, among others [17]. Thus, only 7.54% of metabolites were replicated across multiple studies [17], a finding similar to that of the present study, where 7.27% of metabolites were replicated in validation sets. Similarly, Sohn et al. (2022) investigated metabolomic signatures associated with weight control interventions in children with obesity using untargeted metabolomics in plasma samples and observed 12 metabolites were significant at both time points including asparagine, glutamine, O-acetylcarnitine, and most perturbated metabolic pathways was D-glutamine and glutamate metabolism and arginine biosynthesis [1]. The present study results also find D-glutamine and glutamate metabolism (L-Glutamine) and arginine biosynthesis (L-Glutamine) were the most perturbated metabolic pathways. Another study by Cho et al. (2017) included non-obese (n = 91) and obese (n = 93) adolescents from both sexes and also conducted untargeted and targeted urinary metabolomics. Inflammation-related metabolites were identified with strong predictive power to distinguish obese and non-obese groups, and acylcarnitines (hexanoylcarnitine), amino acid (glutamine, asparagine), amines (carnosine), glycerophospholipids, and sphingolipids were significantly high in obese adolescents [27]. In our study, we found 2-hexanoylcarnitine, L-glutamine, carnosine, and hydroxyprolyl-asparagine are associated with BMI Z-score. Obesity is associated with increased acylcarnitine levels in blood plasma sample of Hispanic children, e.g., propionyl-, butyryl-, hexanoyl-, stearoyl-, and oleoylcarnitine [46], and in adolescence, e.g., propionyl-, 2-methylbutyryl-, isovaleryl-, and isobutyrylcarnitine [54]. In this study, 2-hexenoylcarnitine was associated with the BMI z-score. Acylcarnitines play an important role in many cellular energy metabolism pathways, e.g., transporting acyl groups (organic acids and fatty acids) from the cytoplasm to the mitochondria for the production of energy [55,56]. Papandreou et al. (2021) reported acylcarnitine metabolites (hexanoylcarnitine, hexadecenoylcarnitine) were associated with body fat% [57], indicating that higher body fat correlates with upregulated beta-oxidation of fatty acids. Acylcarnitines are important biomarkers in metabolic studies such as metabolic disorders, cardiovascular diseases, diabetes [56], kidney cancer [58], and hepatocellular carcinoma [59]. Observed elevated levels of carnitine metabolites in adolescents require particular attention and may be a target for obesity management to prevent the development of disease complications.
As previously reported by Brachem et al. (2020), the urinary level of glucuronide of C10H18O2 (12) was positively associated with BMI and body fat in adolescents [23]. In addition, in another study, Tchernof et al. (1997) reported the plasma level of androstane-3α,17β-Diol glucuronide concentration was significantly higher in overweight men [60]. The liver is the primary site of altered glucuronidation, with the help of the UDP-glucuronyl transferase enzyme in conjugation with glucuronate, resulting in the removal of toxic substances, drugs, or other xenobiotics [61,62]. Glucuronides in adipose tissue have been shown to demonstrate higher activity in obese individuals [60,63]. In the current study, N-(2-hydroxyphenyl) acetamide glucuronide was found to be positively associated with BMI Z-score and positive change in BMI in adolescents. The current study observation is consistent with previous studies in that the glucouronide product is associated with obesity [64]. The mechanism of glucouronide products and their relationship with obesity is currently unknown.
Increased adiposity is also associated with changes in amino acid metabolism [47]. In the etiology of obesity and diabetes mellitus, amino acids (AAs) are emerging as a new class of potent molecules. Amino acids (AA) are the building blocks of proteins and play essential roles in gene expression, cell signaling, inflammatory responses, metabolism, oxidative stress, and detoxification [65]. Yamakado et al. reported that the change in the amino acid profile was closely related to the development of metabolic complications such as insulin resistance, diabetes, and visceral fat accumulation [66]. We observed a negative association of BMI z-score with amino acids and their derivatives such as glycylproline, L-glutamine, hydroxyprolyl-asparagine, and galactosylhydroxylysine. In contrast, formiminoglutamic acid, 4-hydroxyproline, and citrulline were positively associated with BMI Z-score. These findings are consistent with those of previous research, where obesity was associated with higher levels of amino acids such as lysine, tryptophan, cystine, and glutamate, but lower levels of asparagine, glutamine, glycine, and serine in Japanese adults [49]. In comparison to children of normal weight, children with obesity had reported higher serum levels of phenylalanine, proline, histidine, and lysine and lower serum levels of glutamine [67]. Citrulline is associated with the urea cycle. A previous study reported a decrease in citrulline in obese adolescents [27,46,67]. While the present study results are inconsistent, we found citrulline is positively associated with BMI z-score and a positive change in BMI in adolescents. Moreover, branched-chain amino acids (BCAAs) and aromatic amino acids have previously been proposed as biomarkers of metabolic syndrome [50]. BCAAs promote protein synthesis and turnover, signaling pathways, and metabolism of glucose. Oxidation of BCAAs may increase fatty acid oxidation and play a role in obesity. In the present study, we also observed a significant positive association of 4-hydroxyproline with BMI z-score and a positive change in BMI in a follow-up study in adolescents. Increasing evidence has been reported that 4-hydroxyproline may play a significant role in protecting mammalian cells from oxidative stress and injury [68]. Additionally, 4-hydroxyproline enhances human nutrition and health, including metabolic, immune, and cardiovascular health [69]. 4-hydroxyproline and carnosine from dietary sources are beneficial for preventing and treating obesity, cardiovascular dysfunction, and aging-related disorders, as well as inhibiting tumorigenesis in children and adults [69]. Our results are consistent with Cho et al. (2017), where carnosine a dipeptide synthesized in the body from β-alanine and L-histidine was found to decrease in obese adolescents [27]. In the current study, we observed galactosylhydroxylysine was negatively associated with BMI Z-score. A study conducted urine metabolic profiling in normal-weight young men and obese men with hyperlipidemia. It was observed that glucosylgalactosyl hydroxylysine along with eight other metabolites has a significant impact on the development and manifestation of obesity-related disorders [70].
Glycylproline, a dipeptide composed of glycine and proline, is considered a building block for proteins. Glycine is also required for multiple metabolic pathways. Several studies reported the level of glycine was found to decrease in children with obesity in plasma samples [46,52]. Similarly, Wahl et al. (2012) reported the level of proline decreased in children with obesity [47]. However, Short et al. investigated plasma profiles of amino acid-related metabolites among Indian American adolescents and observed higher levels of proline in obese children as compared to normal-weight children [67]. In the present study, glycine-proline is negatively associated with BMI z-score and a positive change in BMI after a 1-year follow-up. In vitro and in vivo studies reported cyclic glycine-proline (cGP) mediates the homeostasis of insulin-like growth factor (IGF)-1 function and the cGP/IGF-1 ratio, which determines IGF-1 bioactivity. Plasma IGF-1 is predominantly inactive and slightly related to obesity and hypertension in humans [71]. Glycine and proline are non-essential amino acids (NEAAs) that play a crucial role in nutrition serving and are closely related to the development of tumors. Glycine deficiency reduces the synthesis of glutathione (GSH) and enhances reactive oxygen species (ROS) production. A decline in GSH levels and accumulation of ROS promote lipid peroxidation ultimately leads to tumor suppression through ferroptosis-mediated mechanisms [72]. Formiminoglutamic acid is an intermediate metabolite in the degradative conversion of histidine to glutamic acid [73]. Formiminoglutamic acid is not directly related to obesity, but increased levels of urinary formiminoglutamic acid are associated with a deficiency of folic acid and vitamin B12, which may lead to liver disease [74], CVD [75], and heart disease [76]. In this study, formiminoglutamic acid is positively associated with BMI z-score and a positive change in BMI. The pathway analysis report suggests which metabolic pathways might be altered in individuals with different BMI z-scores and positive changes in BMI. The most perturbated metabolic pathway in the current study is arginine biosynthesis. Arginine is synthesized from citrulline and L-glutamine. Recent findings from human and animal-based research suggested that arginine plays a crucial role in modulating the metabolism of energy substrates [77]. Arginine metabolism stimulates the expression of specific genes in adipose tissue, potentially increasing fatty acid and glucose oxidation [78]. Previous investigations have exhibited that the intake of L-arginine can improve endothelial function, insulin secretion and sensitivity, and inflammation, all of which are interconnected with obesity-induced ailments such as type 2 diabetes mellitus and cardiovascular diseases [79,80,81].
Our study has some limitations. We evaluated a sample of individuals mainly consisting of only male adolescents with normal/overweight/obesity, which could limit the generalizability of our results to other populations and girls. Additionally, we provide no information about the association of obesity and metabolomic features in adults, when BMI is generally more stable and not impacted by changes in hormones (e.g., growth hormones). Second, in our study, we used BMI z-score as a criterion for obesity, rather than body-fat and waist circumference, which can be a more accurate parameter of adiposity. Third, since metabolites were only measured at a one-time point (baseline) and one-year follow-up with a limited sample size (n = 81), we could not provide any data regarding the stability of the metabolome associated with BMI z-score and positive change in BMI over longer times. Fourth, this study did not positively identify the metabolites and apply a targeted approach so that the putative identifications may not be correct. Fifth, we used random urine samples for metabolomics analysis and variability in the metabolite content may be a limitation.
The present study has notable strengths. This current study focuses on adolescent males, capturing what features remain associated with obesity over time. It incorporates both cross-sectional and longitudinal analysis. Replication was enhanced by using a discovery and validation set, which was then applied to longitudinal analysis. This increases the confidence of true findings. Reasons why the sampling at two time points is not concordant could be due to differences in lifestyle at the different time points, changing metabolomic profiles as the participants age, and/or false negatives in spite of multiple comparison corrections. Another strength is the use of the z-score, which is better reflective of future health status as it accounts for age and gender. Additionally, comprehensive untargeted metabolome profiling was performed on UPLC-Q-TOF-MS metabolomic platforms, which provides maximum coverage as compared to targeted metabolomics and another approach to qualitatively analyze a wide range of metabolites. It also helps in the identification of new metabolites associated with obesity and obesity-related complications.

5. Conclusions

The present study assessed the metabolomic profile using an untargeted metabolomics approach and most of the significant metabolic features observed in the current study are consistent with the adolescent study. The current study indicates histidine metabolism as a key mechanism related to obesity. The novel whole-metabolome approach evidenced several biomarkers related to obesity, most of which are carboxylic acid derivatives (glycylproline, 4-hydroxyproline L-glutamine, citrulline, galactosylhydroxylysine), carnitines (2-hexenoylcarnitine), lipids (N-(2-hydroxyphenyl) acetamide glucuronide), carbohydrates (3′-sialyllactose), while several others have been evidenced in the discovery set. These metabolites may directly or indirectly contribute to the development of obesity, particularly its effects, which can be studied in more detail in mechanistic studies. Obesity is a multifaceted and complex health concern that arises from a combination of genetic, behavioral, social, and environmental factors. While biological factors certainly play a role in an individual’s susceptibility to obesity, the environment in which we live greatly influences our food choices, physical activity levels, and overall lifestyle. Even though the use of metabolomics in childhood obesity research is still in its early stages, the identified metabolites have provided additional insight into the pathogenesis of some obesity-related diseases. Furthermore, future research should attempt to replicate our findings in a different adolescent population and extend the analysis to a longitudinal design to better understand the possible correlation of body composition with urine metabolome.

Supplementary Materials

The following supporting information can be downloaded at: https://www.mdpi.com/article/10.3390/metabo13080899/s1, Figure S1: Principal component analysis (PCA) of 1532 annotated urinary metabolomics data with three groups (Discovery set (n = 235), Validation set (n = 125), and replication set (n = 81); Figure S2A: Principal Components analysis with outliers; Figure S2B: T2 Hotelling for I2 Principal Components; Table S1: Supplement Table S1. To study the association of metabolites with BMI z-score in discovery set and validation set; Supplement Table S2. Top enriched metabolic pathways using list of significant features in discovery set; Supplement Table S3. Top enriched metabolic pathways using list of significant features in the validation set; Supplement Table S4. Association of metabolic feature with BMI z-score in discovery set after adjusting for covariate (age, race, total energy intake); Supplement Table S5. Association of metabolic feature with BMI z-score in discovery set after adjusting for covariate (age, race, total energy intake, smoking, socioeconomic status)

Author Contributions

A.S. and P.G.S. made significant contributions to the conception or design of this study. A.S. and G.K. analyzed data. A.S. and P.G.S. interpreted the findings. A.S. wrote the first draft. P.G.S., K.G., P.-C.H., M.-A.S., D.Y.W., S.A.R., J.P.M., J.L.F., B.K.-H. and A.K.F. made significant contributions to the evaluation of the data and findings, the review, and revision of the manuscript. J.P.M. oversaw the statistical data analysis. All authors have read and agreed to the published version of the manuscript.

Funding

Research reported in this publication was supported by The Ohio State University Comprehensive Cancer Center and the National Institutes of Health under grant number P30CA016058 and National Cancer Institute and Food and Drug Administration’s Center for Tobacco Products P50CA180908. The content is solely the responsibility of the authors and does not necessarily represent the official views of the NIH, NCI, or the FDA.

Institutional Review Board Statement

The study is approved by The Ohio State University’s Institutional Review Board.

Informed Consent Statement

Informed consent was obtained from all participants involved in the study.

Data Availability Statement

Data are contained within the article. The raw data supporting the conclusions of this article will be made available by the authors, upon reasonable request.

Acknowledgments

We thank the Recruitment, Intervention and Survey Shared Resource at The Ohio State University Comprehensive Cancer Center for data management, the Nutrient & Phytochemical Analytics Shared Resource (NPASR) for laboratory (metabolomics) analysis, and I also acknowledge the support of the Bioinformatics Shared Resource and the Biostatistics Shared Resource at The Ohio State University (Columbus, OH). We also thank the study participants and staff for recruiting participants as well as data and sample collection.

Conflicts of Interest

The authors declare no conflict of interest.

Abbreviations

BMIBody mass index
UPLC–QTOF-MSUltra-high performance liquid chromatography-quadrupole time-of-flight mass spectrometry
FDRFalse-discovery-rate
BTHBuckeye Teen Health Study
OSUOhio State University
HPLCHigh-performance liquid chromatography
ACNAcetonitrile
4-NBA4-nitrobenzoic acid
ESIElectrospray ionization
QCQuality control
CVCoefficient of variation
MSTUSMass spectrometry total usable signal
HMDBHuman metabolome database
KEGGKyoto Encyclopedia of Genes and Genomes
CVCoefficient of variation
MetPAMetabolomic pathway enrichment analysis
AAsAmino acids
BCAAsBranched-chain amino acids
IGF-1Insulin-like growth factor-1
cGPcyclic Glycine-Proline
NEAAsNon-essential amino acids
ROSReactive oxygen species
GSHGlutathione
CVDCardiovascular diseases

References

  1. Szczerbinski, L.; Wojciechowska, G.; Olichwier, A.; Taylor, M.A.; Puchta, U.; Konopka, P.; Paszko, A.; Citko, A.; Goscik, J.; Fiehn, O.; et al. Untargeted Metabolomics Analysis of the Serum Metabolic Signature of Childhood Obesity. Nutrients 2022, 14, 214. [Google Scholar] [CrossRef] [PubMed]
  2. Stierman, B.; Afful, J.; Carroll, M.D.; Chen, T.-C.; Davy, O.; Fink, S.; Fryar, C.D.; Gu, Q.; Hales, C.M.; Hughes, J.P. National Health and Nutrition Examination Survey 2017–March 2020 Prepandemic Data Files Development of Files and Prevalence Estimates for Selected Health Outcomes. Natl. Health Stat. Rep. 2021, 158. [Google Scholar]
  3. Simmonds, M.; Llewellyn, A.; Owen, C.G.; Woolacott, N. Predicting adult obesity from childhood obesity: A systematic review and meta-analysis. Obes. Rev. 2016, 17, 95–107. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  4. Loos, R.J.; Yeo, G.S. The bigger picture of FTO—The first GWAS-identified obesity gene. Nat. Rev. Endocrinol. 2014, 10, 51–61. [Google Scholar] [CrossRef]
  5. Kahan, L.G.; Mehrzad, R. Chapter 10-Environmental factors related to the obesity epidemic. In Obesity; Mehrzad, R., Ed.; Elsevier: Amsterdam, The Netherlands, 2020; pp. 117–139. [Google Scholar] [CrossRef]
  6. Pérez, L.M.; García, K.; Herrera, R. Psychological, behavioral and familial factors in obese Cuban children and adolescents. MEDICC Rev. 2013, 15, 24–28. [Google Scholar]
  7. Karatsoreos, I.N.; Thaler, J.P.; Borgland, S.L.; Champagne, F.A.; Hurd, Y.L.; Hill, M.N. Food for thought: Hormonal, experiential, and neural influences on feeding and obesity. J. Neurosci. 2013, 33, 17610–17616. [Google Scholar] [CrossRef] [Green Version]
  8. Romieu, I.; Dossus, L.; Barquera, S.; Blottière, H.M.; Franks, P.W.; Gunter, M.; Hwalla, N.; Hursting, S.D.; Leitzmann, M.; Margetts, B. Energy balance and obesity: What are the main drivers? Cancer Causes Control 2017, 28, 247–258. [Google Scholar] [CrossRef] [Green Version]
  9. Bardugo, A.; Fishman, B.; Libruder, C.; Tanne, D.; Ram, A.; Hershkovitz, Y.; Zucker, I.; Furer, A.; Gilon, R.; Chodick, G.; et al. Body Mass Index in 1.9 Million Adolescents and Stroke in Young Adulthood. Stroke 2021, 52, 2043–2052. [Google Scholar] [CrossRef]
  10. Handakas, E.; Lau, C.H.; Alfano, R.; Chatzi, V.L.; Plusquin, M.; Vineis, P.; Robinson, O. A systematic review of metabolomic studies of childhood obesity: State of the evidence for metabolic determinants and consequences. Obes. Rev. 2022, 23, e13384. [Google Scholar] [CrossRef]
  11. Hruby, A.; Hu, F.B. The Epidemiology of Obesity: A Big Picture. Pharmacoeconomics 2015, 33, 673–689. [Google Scholar] [CrossRef]
  12. Steinberger, J.; Daniels, S.R. Obesity, Insulin Resistance, Diabetes, and Cardiovascular Risk in Children. Circulation 2003, 107, 1448–1453. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  13. Daneshzad, E.; Rostami, S.; Aghamahdi, F.; Mahdavi-Gorabi, A.; Qorbani, M. Association of cardiometabolic risk factors with insulin resistance in overweight and obese children. BMC Endocr. Disord. 2022, 22, 320. [Google Scholar] [CrossRef]
  14. Vanderwall, C.; Randall Clark, R.; Eickhoff, J.; Carrel, A.L. BMI is a poor predictor of adiposity in young overweight and obese children. BMC Pediatr. 2017, 17, 135. [Google Scholar] [CrossRef] [Green Version]
  15. Vehrs, P.R.; Fellingham, G.W.; McAferty, A.; Kelsey, L. Trends in BMI Percentile and Body Fat Percentage in Children 12 to 17 Years of Age. Children 2022, 9, 744. [Google Scholar] [CrossRef]
  16. Grossman, D.C.; Bibbins-Domingo, K.; Curry, S.J.; Barry, M.J.; Davidson, K.W.; Doubeni, C.A.; Epling, J.W.; Kemper, A.R.; Krist, A.H.; Kurth, A.E. Screening for obesity in children and adolescents: US Preventive Services Task Force recommendation statement. Jama 2017, 317, 2417–2426. [Google Scholar]
  17. Cirulli, E.T.; Guo, L.; Swisher, C.L.; Shah, N.; Huang, L.; Napier, L.A.; Kirkness, E.F.; Spector, T.D.; Caskey, C.T.; Thorens, B. Profound perturbation of the metabolome in obesity is associated with health risk. Cell Metab. 2019, 29, 488–500.e482. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  18. Jacob, M.; Lopata, A.L.; Dasouki, M.; Abdel Rahman, A.M. Metabolomics toward personalized medicine. Mass. Spectrom. Rev. 2019, 38, 221–238. [Google Scholar] [CrossRef] [PubMed]
  19. Wishart, D.S. Emerging applications of metabolomics in drug discovery and precision medicine. Nat. Rev. Drug Discov. 2016, 15, 473–484. [Google Scholar] [CrossRef]
  20. Yu, E.Y.-W.; Ren, Z.; Mehrkanoon, S.; Stehouwer, C.D.A.; van Greevenbroek, M.M.J.; Eussen, S.J.P.M.; Zeegers, M.P.; Wesselius, A. Plasma metabolomic profiling of dietary patterns associated with glucose metabolism status: The Maastricht Study. BMC Med. 2022, 20, 450. [Google Scholar] [CrossRef]
  21. Sohn, M.J.; Chae, W.; Ko, J.S.; Cho, J.Y.; Kim, J.E.; Choi, J.Y.; Jang, H.B.; Lee, H.J.; Park, S.I.; Park, K.H.; et al. Metabolomic Signatures for the Effects of Weight Loss Interventions on Severe Obesity in Children and Adolescents. Metabolites 2021, 12, 27. [Google Scholar] [CrossRef]
  22. Rigamonti, A.E.; Frigerio, G.; Caroli, D.; De Col, A.; Cella, S.G.; Sartorio, A.; Fustinoni, S. A Metabolomics-Based Investigation of the Effects of a Short-Term Body Weight Reduction Program in a Cohort of Adolescents with Obesity: A Prospective Interventional Clinical Study. Nutrients 2023, 15, 529. [Google Scholar] [CrossRef]
  23. Brachem, C.; Langenau, J.; Weinhold, L.; Schmid, M.; Nöthlings, U.; Oluwagbemigun, K. Associations of BMI and Body Fat with Urine Metabolome in Adolescents Are Sex-Specific: A Cross-Sectional Study. Metabolites 2020, 10, 330. [Google Scholar] [CrossRef] [PubMed]
  24. Concepcion, J.; Chen, K.; Saito, R.; Gangoiti, J.; Mendez, E.; Nikita, M.E.; Barshop, B.A.; Natarajan, L.; Sharma, K.; Kim, J.J. Identification of pathognomonic purine synthesis biomarkers by metabolomic profiling of adolescents with obesity and type 2 diabetes. PLoS ONE 2020, 15, e0234970. [Google Scholar] [CrossRef] [PubMed]
  25. Lau, C.-H.E.; Siskos, A.P.; Maitre, L.; Robinson, O.; Athersuch, T.J.; Want, E.J.; Urquiza, J.; Casas, M.; Vafeiadi, M.; Roumeliotaki, T. Determinants of the urinary and serum metabolome in children from six European populations. BMC Med. 2018, 16, 1–19. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  26. Troisi, J.; Pierri, L.; Landolfi, A.; Marciano, F.; Bisogno, A.; Belmonte, F.; Palladino, C.; Guercio Nuzio, S.; Campiglia, P.; Vajro, P. Urinary Metabolomics in Pediatric Obesity and NAFLD Identifies Metabolic Pathways/Metabolites Related to Dietary Habits and Gut-Liver Axis Perturbations. Nutrients 2017, 9, 485. [Google Scholar] [CrossRef] [Green Version]
  27. Cho, K.; Moon, J.S.; Kang, J.H.; Jang, H.B.; Lee, H.J.; Park, S.I.; Yu, K.S.; Cho, J.Y. Combined untargeted and targeted metabolomic profiling reveals urinary biomarkers for discriminating obese from normal-weight adolescents. Pediatr. Obes. 2017, 12, 93–101. [Google Scholar] [CrossRef]
  28. Pathmasiri, W.; Pratt, K.; Collier, D.; Lutes, L.; McRitchie, S.; Sumner, S. Integrating metabolomic signatures and psychosocial parameters in responsivity to an immersion treatment model for adolescent obesity. Metabolomics 2012, 8, 1037–1051. [Google Scholar] [CrossRef]
  29. Leal-Witt, M.J.; Ramon-Krauel, M.; Samino, S.; Llobet, M.; Cuadras, D.; Jimenez-Chillaron, J.C.; Yanes, O.; Lerin, C. Untargeted metabolomics identifies a plasma sphingolipid-related signature associated with lifestyle intervention in prepubertal children with obesity. Int. J. Obes. 2018, 42, 72–78. [Google Scholar] [CrossRef]
  30. Villarreal-Pérez, J.Z.; Villarreal-Martínez, J.Z.; Lavalle-González, F.J.; Torres-Sepúlveda, M.d.R.; Ruiz-Herrera, C.; Cerda-Flores, R.M.; Castillo-García, E.R.; Rodríguez-Sánchez, I.P.; Martínez de Villarreal, L.E. Plasma and urine metabolic profiles are reflective of altered beta-oxidation in non-diabetic obese subjects and patients with type 2 diabetes mellitus. Diabetol. Metab. Syndr. 2014, 6, 129. [Google Scholar] [CrossRef] [Green Version]
  31. Robinson, O.; Chadeau Hyam, M.; Karaman, I.; Climaco Pinto, R.; Ala-Korpela, M.; Handakas, E.; Fiorito, G.; Gao, H.; Heard, A.; Jarvelin, M.R.; et al. Determinants of accelerated metabolomic and epigenetic aging in a UK cohort. Aging Cell 2020, 19, e13149. [Google Scholar] [CrossRef]
  32. Friedman, K.L.; Roberts, M.E.; Keller-Hamilton, B.; Yates, K.A.; Paskett, E.D.; Berman, M.L.; Slater, M.D.; Lu, B.; Ferketich, A.K. Attitudes toward Tobacco, Alcohol, and Non-Alcoholic Beverage Advertisement Themes among Adolescent Boys. Subst. Use Misuse 2018, 53, 1706–1714. [Google Scholar] [CrossRef] [PubMed]
  33. Keller-Hamilton, B.; Muff, J.; Blue, T.; Lu, B.; Slater, M.D.; Roberts, M.E.; Ferketich, A.K. Tobacco and Alcohol on Television: A Content Analysis of Male Adolescents' Favorite Shows. Prev. Chronic Dis. 2018, 15, E134. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  34. Keller-Hamilton, B.; Roberts, M.E.; Slater, M.D.; Ferketich, A.K. Memorability of Cigarette Advertisements Making "Natural" Claims Among Adolescents. Tob. Regul. Sci. 2019, 5, 326–331. [Google Scholar] [CrossRef]
  35. Keller-Hamilton, B.; Lu, B.; Roberts, M.E.; Berman, M.L.; Root, E.D.; Ferketich, A.K. Electronic cigarette use and risk of cigarette and smokeless tobacco initiation among adolescent boys: A propensity score matched analysis. Addict. Behav. 2021, 114, 106770. [Google Scholar] [CrossRef] [PubMed]
  36. Burgoon, M.L.; Albani, T.; Keller-Hamilton, B.; Lu, B.; Roberts, M.E.; Craigmile, P.F.; Browning, C.; Xi, W.; Ferketich, A.K. Exposures to the tobacco retail environment among adolescent boys in urban and rural environments. Am. J. Drug Alcohol. Abus. 2019, 45, 217–226. [Google Scholar] [CrossRef]
  37. Nixon, D.E.; Ferketich, A.K.; Slater, M.D.; Mays, D.; Keller-Hamilton, B. Prospective associations between attitudes toward alcohol advertisements and alcohol use behaviors among adolescent boys. Addict. Behav. Rep. 2022, 15, 100428. [Google Scholar] [CrossRef]
  38. Roberts, M.E.; Keller-Hamilton, B.; Ferketich, A.K. Testing if attitudes mediate the association between advertising exposure and adolescent tobacco use. Addict. Behav. 2022, 134, 107415. [Google Scholar] [CrossRef]
  39. Mahieu, N.G.; Genenbacher, J.L.; Patti, G.J. A roadmap for the XCMS family of software solutions in metabolomics. Curr. Opin. Chem. Biol. 2016, 30, 87–93. [Google Scholar] [CrossRef] [Green Version]
  40. Faquih, T.; van Smeden, M.; Luo, J.; le Cessie, S.; Kastenmüller, G.; Krumsiek, J.; Noordam, R.; van Heemst, D.; Rosendaal, F.R.; van Hylckama Vlieg, A.; et al. A Workflow for Missing Values Imputation of Untargeted Metabolomics Data. Metabolites 2020, 10, 486. [Google Scholar] [CrossRef]
  41. Abdi, H.; Williams, L.J. Principal component analysis. WIREs Comput. Stat. 2010, 2, 433–459. [Google Scholar] [CrossRef]
  42. Benjamini, Y.; Hochberg, Y. Controlling the False Discovery Rate: A Practical and Powerful Approach to Multiple Testing. J. R. Stat. Society. Ser. B (Methodol.) 1995, 57, 289–300. [Google Scholar] [CrossRef]
  43. Xia, J.; Wishart, D.S. Metabolomic Data Processing, Analysis, and Interpretation Using MetaboAnalyst. Curr. Protoc. Bioinform. 2011, Chapter 14. 14.10.11–14.10.48. [Google Scholar] [CrossRef] [PubMed]
  44. Xia, J.; Sinelnikov, I.V.; Han, B.; Wishart, D.S. MetaboAnalyst 3.0—Making Metabolomics More Meaningful. Nucleic Acids Research 2015, 43, W251–W257. [Google Scholar] [CrossRef] [Green Version]
  45. Sorrow, P.; Maguire, R.; Murphy, S.K.; Belcher, S.M.; Hoyo, C. Elevated metabolites of acetaminophen in cord blood of children with obesity. Pediatr. Obes. 2019, 14, e12465. [Google Scholar] [CrossRef] [Green Version]
  46. Butte, N.F.; Liu, Y.; Zakeri, I.F.; Mohney, R.P.; Mehta, N.; Voruganti, V.S.; Göring, H.; Cole, S.A.; Comuzzie, A.G. Global metabolomic profiling targeting childhood obesity in the Hispanic population. Am. J. Clin. Nutr. 2015, 102, 256–267. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  47. Wahl, S.; Yu, Z.; Kleber, M.; Singmann, P.; Holzapfel, C.; He, Y.; Mittelstrass, K.; Polonikov, A.; Prehn, C.; Römisch-Margl, W.; et al. Childhood Obesity Is Associated with Changes in the Serum Metabolite Profile. Obes. Facts 2012, 5, 660–670. [Google Scholar] [CrossRef]
  48. Gawlik, A.; Shmoish, M.; Hartmann, M.F.; Malecka-Tendera, E.; Wudy, S.A.; Hochberg, Z. Steroid Metabolomic Disease Signature of Nonsyndromic Childhood Obesity. J. Clin. Endocrinol. Metab. 2016, 101, 4329–4337. [Google Scholar] [CrossRef] [Green Version]
  49. Takashina, C.; Tsujino, I.; Watanabe, T.; Sakaue, S.; Ikeda, D.; Yamada, A.; Sato, T.; Ohira, H.; Otsuka, Y.; Oyama-Manabe, N.; et al. Associations among the plasma amino acid profile, obesity, and glucose metabolism in Japanese adults with normal glucose tolerance. Nutr Metab (Lond) 2016, 13, 5. [Google Scholar] [CrossRef]
  50. Gogna, N.; Krishna, M.; Oommen, A.M.; Dorai, K. Investigating correlations in the altered metabolic profiles of obese and diabetic subjects in a South Indian Asian population using an NMR-based metabolomic approach. Mol. BioSystems 2015, 11, 595–606. [Google Scholar] [CrossRef]
  51. Kim, J.Y.; Park, J.Y.; Kim, O.Y.; Ham, B.M.; Kim, H.-J.; Kwon, D.Y.; Jang, Y.; Lee, J.H. Metabolic profiling of plasma in overweight/obese and lean men using ultra performance liquid chromatography and Q-TOF mass spectrometry (UPLC− Q-TOF MS). J. Proteome Res. 2010, 9, 4368–4375. [Google Scholar] [CrossRef]
  52. Mihalik, S.J.; Michaliszyn, S.F.; de las Heras, J.; Bacha, F.; Lee, S.; Chace, D.H.; DeJesus, V.R.; Vockley, J.; Arslanian, S.A. Metabolomic Profiling of Fatty Acid and Amino Acid Metabolism in Youth With Obesity and Type 2 Diabetes: Evidence for enhanced mitochondrial oxidation. Diabetes Care 2012, 35, 605–611. [Google Scholar] [CrossRef] [Green Version]
  53. Newgard, C.B.; An, J.; Bain, J.R.; Muehlbauer, M.J.; Stevens, R.D.; Lien, L.F.; Haqq, A.M.; Shah, S.H.; Arlotto, M.; Slentz, C.A.; et al. A branched-chain amino acid-related metabolic signature that differentiates obese and lean humans and contributes to insulin resistance. Cell Metab. 2009, 9, 311–326. [Google Scholar] [CrossRef] [Green Version]
  54. Perng, W.; Rifas-Shiman, S.L.; Hivert, M.F.; Chavarro, J.E.; Oken, E. Branched Chain Amino Acids, Androgen Hormones, and Metabolic Risk Across Early Adolescence: A Prospective Study in Project Viva. Obes. (Silver Spring) 2018, 26, 916–926. [Google Scholar] [CrossRef]
  55. Kang, M.; Yoo, H.J.; Kim, M.; Kim, M.; Lee, J.H. Metabolomics identifies increases in the acylcarnitine profiles in the plasma of overweight subjects in response to mild weight loss: A randomized, controlled design study. Lipids Health Dis. 2018, 17, 237. [Google Scholar] [CrossRef] [Green Version]
  56. Dambrova, M.; Makrecka-Kuka, M.; Kuka, J.; Vilskersts, R.; Nordberg, D.; Attwood, M.M.; Smesny, S.; Sen, Z.D.; Guo, A.C.; Oler, E.; et al. Acylcarnitines: Nomenclature, Biomarkers, Therapeutic Potential, Drug Targets, and Clinical Trials. Pharmacol. Rev. 2022, 74, 506–551. [Google Scholar] [CrossRef]
  57. Papandreou, C.; García-Gavilán, J.; Camacho-Barcia, L.; Hansen, T.T.; Sjödin, A.; Harrold, J.A.; Halford, J.C.G.; Bulló, M. Circulating Metabolites Associated with Body Fat and Lean Mass in Adults with Overweight/Obesity. Metabolites 2021, 11, 317. [Google Scholar] [CrossRef] [PubMed]
  58. Ganti, S.; Taylor, S.L.; Kim, K.; Hoppel, C.L.; Guo, L.; Yang, J.; Evans, C.; Weiss, R.H. Urinary acylcarnitines are altered in human kidney cancer. Int. J. Cancer 2012, 130, 2791–2800. [Google Scholar] [CrossRef]
  59. Li, S.; Gao, D.; Jiang, Y. Function, Detection and Alteration of Acylcarnitine Metabolism in Hepatocellular Carcinoma. Metabolites 2019, 9, 36. [Google Scholar] [CrossRef] [Green Version]
  60. Tchernof, A.; Labrie, F.; Bélanger, A.; Prud’homme, D.; Bouchard, C.; Tremblay, A.; Nadeau, A.; Després, J.-P. Androstane-3α,17β-Diol Glucuronide as a Steroid Correlate of Visceral Obesity in Men*. J. Clin. Endocrinol. Metab. 1997, 82, 1528–1534. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  61. Dalesio, N.M.; Lee, C.K.K.; Hendrix, C.W.; Kerns, N.; Hsu, A.; Clarke, W.; Collaco, J.M.; McGrath-Morrow, S.; Yaster, M.; Brown, R.H.; et al. Effects of Obstructive Sleep Apnea and Obesity on Morphine Pharmacokinetics in Children. Anesth. Analg. 2020, 131, 876–884. [Google Scholar] [CrossRef] [PubMed]
  62. Schiffer, L.; Barnard, L.; Baranowski, E.S.; Gilligan, L.C.; Taylor, A.E.; Arlt, W.; Shackleton, C.H.L.; Storbeck, K.-H. Human steroid biosynthesis, metabolism and excretion are differentially reflected by serum and urine steroid metabolomes: A comprehensive review. J. Steroid Biochem. Mol. Biol. 2019, 194, 105439. [Google Scholar] [CrossRef] [PubMed]
  63. Moore, S.C.; Matthews, C.E.; Sampson, J.N.; Stolzenberg-Solomon, R.Z.; Zheng, W.; Cai, Q.; Tan, Y.T.; Chow, W.-H.; Ji, B.-T.; Liu, D.K. Human metabolic correlates of body mass index. Metabolomics 2014, 10, 259–269. [Google Scholar] [CrossRef]
  64. Vandenput, L.; Mellström, D.; Lorentzon, M.; Swanson, C.; Karlsson, M.K.; Brandberg, J.; Lönn, L.; Orwoll, E.; Smith, U.; Labrie, F.; et al. Androgens and Glucuronidated Androgen Metabolites Are Associated with Metabolic Risk Factors in Men. J. Clin. Endocrinol. Metab. 2007, 92, 4130–4137. [Google Scholar] [CrossRef] [Green Version]
  65. Kelly, B.; Pearce, E.L. Amino assets: How amino acids support immunity. Cell Metab. 2020, 32, 154–175. [Google Scholar] [CrossRef]
  66. Yamakado, M.; Tanaka, T.; Nagao, K.; Ishizaka, Y.; Mitushima, T.; Tani, M.; Toda, A.; Toda, E.; Okada, M.; Miyano, H.; et al. Plasma amino acid profile is associated with visceral fat accumulation in obese Japanese subjects. Clin. Obes. 2012, 2, 29–40. [Google Scholar] [CrossRef] [PubMed]
  67. Short, K.R.; Chadwick, J.Q.; Teague, A.M.; Tullier, M.A.; Wolbert, L.; Coleman, C.; Copeland, K.C. Effect of obesity and exercise training on plasma amino acids and amino metabolites in American Indian adolescents. J. Clin. Endocrinol. Metab. 2019, 104, 3249–3261. [Google Scholar] [CrossRef] [PubMed]
  68. Wu, Z.; Hou, Y.; Dai, Z.; Hu, C.-A.A.; Wu, G. Metabolism, nutrition, and redox signaling of hydroxyproline. Antioxid. Redox Signal. 2019, 30, 674–682. [Google Scholar] [CrossRef]
  69. Wu, G. Important roles of dietary taurine, creatine, carnosine, anserine and 4-hydroxyproline in human nutrition and health. Amino Acids 2020, 52, 329–360. [Google Scholar] [CrossRef] [Green Version]
  70. Wang, C.; Feng, R.; Sun, D.; Li, Y.; Bi, X.; Sun, C. Metabolic profiling of urine in young obese men using ultra performance liquid chromatography and Q-TOF mass spectrometry (UPLC/Q-TOF MS). J. Chromatogr. B Anal. Technol. Biomed. Life Sci. 2011, 879, 2871–2876. [Google Scholar] [CrossRef]
  71. Singh-Mallah, G.; Singh, K.; McMahon, C.D.; Harris, P.; Brimble, M.A.; Thorstensen, E.; Guan, J. Maternally Administered Cyclic Glycine-Proline Increases Insulin-Like Growth Factor-1 Bioavailability and Novelty Recognition in Developing Offspring. Endocrinology 2016, 157, 3130–3139. [Google Scholar] [CrossRef] [Green Version]
  72. Qu, C.; Peng, Y.; Liu, S. Ferroptosis Biology and Implication in Cancers. Front. Mol. Biosci. 2022, 9, 892957. [Google Scholar] [CrossRef]
  73. Haam, J.H.; Kim, Y.S.; Cho, D.Y.; Chun, H.; Choi, S.W.; Lee, Y.K.; Lim, S.W.; Koo, H.S.; Kim, M.J. Elevated levels of urine isocitrate, hydroxymethylglutarate, and formiminoglutamate are associated with arterial stiffness in Korean adults. Sci. Rep. 2021, 11, 10180. [Google Scholar] [CrossRef]
  74. Rose, D.P. Value of Detection of Formiminoglutamic Acid in Urine for the Diagnosis of Malabsorption States. Br. Med. J. 1965, 1, 1031–1034. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  75. Verhaar, M.C.; Stroes, E.; Rabelink, T.J. Folates and Cardiovascular Disease. Arterioscler. Thromb. Vasc. Biol. 2002, 22, 6–13. [Google Scholar] [CrossRef] [Green Version]
  76. Luhby, A.L.; Cooperman, J.M.; Teller, D.N. Urinary excretion of formiminoglutamic acid: Application in diagnosis of clinical folic acid deficiency. Am. J. Clin. Nutr. 1959, 7, 397–406. [Google Scholar] [CrossRef] [PubMed]
  77. McKnight, J.R.; Satterfield, M.C.; Jobgen, W.S.; Smith, S.B.; Spencer, T.E.; Meininger, C.J.; McNeal, C.J.; Wu, G. Beneficial effects of L-arginine on reducing obesity: Potential mechanisms and important implications for human health. Amino Acids 2010, 39, 349–357. [Google Scholar] [CrossRef]
  78. Fu, W.J.; Haynes, T.E.; Kohli, R.; Hu, J.; Shi, W.; Spencer, T.E.; Carroll, R.J.; Meininger, C.J.; Wu, G. Dietary L-arginine supplementation reduces fat mass in Zucker diabetic fatty rats. J. Nutr. 2005, 135, 714–721. [Google Scholar] [CrossRef] [Green Version]
  79. Lucotti, P.; Setola, E.; Monti, L.D.; Galluccio, E.; Costa, S.; Sandoli, E.P.; Fermo, I.; Rabaiotti, G.; Gatti, R.; Piatti, P. Beneficial effects of a long-term oral L-arginine treatment added to a hypocaloric diet and exercise training program in obese, insulin-resistant type 2 diabetic patients. Am. J. Physiol. Endocrinol. Metab. 2006, 291, E906–E912. [Google Scholar] [CrossRef] [PubMed]
  80. Pahlavani, N.; Entezari, M.; Nasiri, M.; Miri, A.; Rezaie, M.; Bagheri-Bidakhavidi, M.; Sadeghi, O. The effect of l-arginine supplementation on body composition and performance in male athletes: A double-blinded randomized clinical trial. Eur. J. Clin. Nutr. 2017, 71, 544–548. [Google Scholar] [CrossRef] [PubMed]
  81. Wu, G.; Bazer, F.W.; Davis, T.A.; Kim, S.W.; Li, P.; Marc Rhoads, J.; Carey Satterfield, M.; Smith, S.B.; Spencer, T.E.; Yin, Y. Arginine metabolism and nutrition in growth, health and disease. Amino Acids 2009, 37, 153–168. [Google Scholar] [CrossRef] [Green Version]
Metabolites 13 00899 g001 550
Figure 1. The flow diagram describes the discovery set, validation set, and replication set (1-year follow-up), data processing, statistical analysis, and a number of significant metabolites determined in each experiment that were used for pathway enrichment analysis.
Figure 1. The flow diagram describes the discovery set, validation set, and replication set (1-year follow-up), data processing, statistical analysis, and a number of significant metabolites determined in each experiment that were used for pathway enrichment analysis.
Metabolites 13 00899 g001
Metabolites 13 00899 g002 550
Figure 2. The association between metabolites and BMI Z−score. Volcano plot: each point on the volcano plot was based on estimates of beta (β) from linear regression analysis on the x−axis versus the log10 FDR adj. p−value on the y−axis of the putatively identified metabolites. (A) Volcano plot of association of metabolites with BMI z−score in the discovery set; (B) volcano plot of association of metabolites with BMI z−score in the validation set (red dot for a positive association and blue dot for a negative association, horizontal dotted lines set false discovery rate 10% corrected p−value).
Figure 2. The association between metabolites and BMI Z−score. Volcano plot: each point on the volcano plot was based on estimates of beta (β) from linear regression analysis on the x−axis versus the log10 FDR adj. p−value on the y−axis of the putatively identified metabolites. (A) Volcano plot of association of metabolites with BMI z−score in the discovery set; (B) volcano plot of association of metabolites with BMI z−score in the validation set (red dot for a positive association and blue dot for a negative association, horizontal dotted lines set false discovery rate 10% corrected p−value).
Metabolites 13 00899 g002
Metabolites 13 00899 g003 550
Figure 3. A pie chart summarizing the range and classes of compounds (super-class annotated by HMDB) demonstrates a significant association of metabolites with BMI z-score in the discovery set of urine samples from adolescents.
Figure 3. A pie chart summarizing the range and classes of compounds (super-class annotated by HMDB) demonstrates a significant association of metabolites with BMI z-score in the discovery set of urine samples from adolescents.
Metabolites 13 00899 g003
Metabolites 13 00899 g004 550
Figure 4. Pathway analysis as generated by MetaboAnalyst 5.0 software package for metabolites significantly expressed in adolescent children. The enrichment ratio is calculated as the number of hits within a particular metabolic pathway divided by the expected number of hits. Metabolite set enrichment analysis (MSEA). Top perturbed pathways are shown. (A) Metabolic pathway analysis using the list of 221 significant features in the discovery set, and (B) a list of 16 significant features in the validation set. The color depth and column length indicate the disturbance degree of the pathway.
Figure 4. Pathway analysis as generated by MetaboAnalyst 5.0 software package for metabolites significantly expressed in adolescent children. The enrichment ratio is calculated as the number of hits within a particular metabolic pathway divided by the expected number of hits. Metabolite set enrichment analysis (MSEA). Top perturbed pathways are shown. (A) Metabolic pathway analysis using the list of 221 significant features in the discovery set, and (B) a list of 16 significant features in the validation set. The color depth and column length indicate the disturbance degree of the pathway.
Metabolites 13 00899 g004
Table
Table 1. Characteristics of the study participants.
Table 1. Characteristics of the study participants.
ParameterDiscovery Set
(N = 235)		Validation Set
(N = 125)		Replication Set
(N = 81)
Age (Year) (Mean ± SD) Range14.80 ± 1.42
(11.14–16.99)		14.84 ± 1.31
(11.06–16.99)		16.08 ± 1.20
(12.9–18)
Height (Inch) (Mean ± SD) Range66.82 ± 4.10
(56.00–75.67)		66.97 ± 3.87
(57.25–75.25)		68.88 ± 3.18
(60–75)
Weight (lb) (Mean ± SD) Range158.50 ± 54.36
(72.93–334.13)		149.85 ± 45.14
(78.2–304.6)		170.64 ± 50.59
(79–320)
BMI (Kg/m2) (Mean ± SD) Range24.64 ± 7.13
(15.3–49)		23.28 ± 6.01
(15.5–44.10)		25.16 ± 6.91
(14.4–45.9)
BMI percentile (Mean ± SD) Range68.2 ± 30.43
(2–99)		66.54 ± 30.39
(2–99)		68.82 ± 29.11
(1–99)
BMI z-score (Mean ± SD) Range0.82 ± 1.23
(−2.07–3.01)		0.60 ± 1.18
(−2.04–2.84)		0.78 ± 1.23
(−3.06–2.96)
Positive change in BMI (Kg/m2) Mean/Median (Range)------0.54/0.6
(−11.6–5.8)
Obesity (n (%))
Underweight (<5 percentile)
Healthy weight (5 to <85 percentile)
Overweight (>85 to <95 percentile)
Obese (>95 percentile)		3 (1.28%)
129 (54.89%)
29 (12.34%)
74 (31.49%)		4 (3.20%)
70 (56.00%)
23 (18.40%)
28 (22.40%)		3 (3.70%)
47 (58.02%)
12 (14.81%)
19 (23.46%)
Race (n (%))
White
Black
Hispanic
Multiracial
Others		176 (74.47%)
36 (15.35%)
8 (3.39%)
12 (5.08%)
4 (1.69%)		99 (79.20%)
13 (10.40%)
4 (3.20%)
7 (5.60%)
2 (1.60%)		58 (71.60%)
13 (16.05%)
4 (4.94%)
5 (6.17%)
1 (1.23%)
County (n (%))
Franklin
Non-Franklin		121 (51.49%)
114 (48.51%)		64 (51.20%)
61 (48.80%)		44(54.32%)
37(45.68%)
Total energy intake (kcals)
(Mean/Median) Range		1883.84/1745.54
(204.05–5209.14)		1937.77/1873.41
(462.5–5549.48)		1778.03/1507.12
(236.91–4589.2)
Abbreviations: BMI, body mass index; SD, standard deviation.
Table
Table 2. Top 30 metabolic features associated with BMI z-score in the discovery set of samples.
Table 2. Top 30 metabolic features associated with BMI z-score in the discovery set of samples.
IDMetaboliteModeMZRTAdductHMDB IDSuper-ClassClassSub-ClassDiscovery Set
Estimate (95%CI)p-ValueFDR adj
neg_FT172073′-SialyllactoseNeg632.20.67M − HHMDB0000825Organic oxygen compoundsOrganooxygen compoundsCarbohydrates and carbohydrate conjugates0.05 (0.03–0.06)<0.00015.82 × 10−9
pos_FT293543′-SialyllactosePos656.20.67M + NaHMDB0000825Organic oxygen compoundsOrganooxygen compoundsCarbohydrates and carbohydrate conjugates0.06 (0.04–0.07)<0.00016.56 × 10−9
pos_FT04847Estrone sulfatePos176.060.58M + 2HHMDB0001425Lipids and lipid-like moleculesSteroids and steroid derivativesSulfated steroids−0.08 (−0.1–−0.05)<0.00011.38 × 10−7
pos_FT13024N-RibosylhistidinePos288.110.53M + HHMDB0002089Organic acids and derivativesCarboxylic acids and derivativesAmino acids, peptides, and analogues0.16 (0.11–0.21)<0.00011.38 × 10−7
pos_FT04858CitrullinePos176.10.78M + HHMDB0000904Organic acids and derivativesCarboxylic acids and derivativesAmino acids, peptides, and analogues0.06 (0.04–0.08)<0.00015.77 × 10−7
neg_FT15762PA(22:5(4Z,7Z,10Z,13Z,19Z)-O(16,17)/2:0)Neg539.246.53M − HHMDB0266570 0.08 (0.06–0.11)<0.00015.77 × 10−7
pos_FT27056Tetrahydroaldosterone-3-glucuronidePos563.246.57M + NaHMDB0010357Lipids and lipid-like moleculesSteroids and steroid derivativesSteroidal glycosides0.07 (0.05–0.1)<0.00018.64 × 10−7
neg_FT01967CitrullineNeg174.080.81M − HHMDB0000904Organic acids and derivativesCarboxylic acids and derivativesAmino acids, peptides, and analogues0.13 (0.08–0.17)<0.00016.28 × 10−6
neg_FT15806Cortolone-3-glucuronideNeg541.266.18M − HHMDB0010320Lipids and lipid-like moleculesSteroids and steroid derivativesSteroidal glycosides0.05 (0.03–0.07)<0.00011.34 × 10−5
neg_FT01921Formiminoglutamic acidNeg173.051.31M − HHMDB0000854Organic acids and derivativesCarboxylic acids and derivativesAmino acids, peptides, and analogues0.13 (0.08–0.18)<0.00014.66 × 10−5
pos_FT04673GlycylprolinePos173.090.75M + HHMDB0000721Organic acids and derivativesCarboxylic acids and derivativesAmino acids, peptides, and analogues−0.06 (−0.09–−0.04)<0.00017.46 × 10−5
neg_FT02197GalactitolNeg181.071.81M − HHMDB0000107Organic oxygen compoundsOrganooxygen compoundsCarbohydrates and carbohydrate conjugates0.09 (0.05–0.13)<0.00017.46 × 10−5
pos_FT27104Cortolone-3-glucuronidePos565.266.1M + NaHMDB0010320Lipids and lipid-like moleculesSteroids and steroid derivativesSteroidal glycosides0.05 (0.03–0.08)<0.00019.30 × 10−5
pos_FT096304-VinylsyringolPos243.090.73M + HHMDB0301746Phenylpropanoids and polyketidesStilbenes −0.05 (−0.07–−0.03)<0.00019.96 × 10−5
neg_FT07183FludiazepamNeg301.052.36M − HHMDB0015513Organoheterocyclic compoundsBenzodiazepines1,4-benzodiazepines0.12 (0.07–0.17)<0.00011.38 × 10−4
neg_FT02892AdipoylglycineNeg202.073.23M − HHMDB0240731Organic acids and derivativesCarboxylic acids and derivativesAmino acids, peptides, and analogues0.06 (0.04–0.09)<0.00011.57 × 10−4
neg_FT111706-Hydroxymelatonin glucuronideNeg389.186.34M + ClHMDB0060786Organic oxygen compoundsOrganooxygen compoundsCarbohydrates and carbohydrate conjugates0.17 (0.1–0.24)<0.00012.29 × 10−4
neg_FT15862N-Acetylgalactosaminyl lactoseNeg544.180.71M − HHMDB0041622Organic oxygen compoundsOrganooxygen compoundsCarbohydrates and carbohydrate conjugates0.08 (0.04–0.11)<0.00012.39 × 10−4
neg_FT021483-ChlorotyrosineNeg180.061.81M + ClHMDB0001885Organic acids and derivativesCarboxylic acids and derivativesAmino acids, peptides, and analogues0.05 (0.03–0.08)<0.00012.55 × 10−4
pos_FT18573CephalexinPos370.085.82M + NaHMDB0014707Organoheterocyclic compoundsLactamsBeta lactams0.12 (0.07–0.17)<0.00013.66 × 10−4
neg_FT08222Dihyroxy-1H-indole glucuronide INeg324.073.91M − HHMDB0059997Organic oxygen compoundsOrganooxygen compoundsCarbohydrates and carbohydrate conjugates0.07 (0.04–0.1)<0.00013.66 × 10−4
neg_FT140083-alpha-hydroxy-5-alpha-androstane-17-one 3-D-glucuronideNeg465.247.84M − HHMDB0010365Lipids and lipid-like moleculesSteroids and steroid derivativesSteroidal glycosides0.06 (0.03–0.09)<0.00013.84 × 10−4
neg_FT14066Clozapine glucuronideNeg467.196.92M + ClHMDB0060901Organic oxygen compoundsOrganooxygen compoundsCarbonyl compounds0.08 (0.05–0.12)<0.00013.84 × 10−4
neg_FT01695Quinolinic acidNeg166.011.17M − HHMDB0000232Organoheterocyclic compoundsPyridines and derivativesPyridinecarboxylic acids and derivatives0.05 (0.02–0.07)<0.00013.84 × 10−4
neg_FT08456Hydroxytyrosol 3′-glucuronideNeg329.083.78M − HHMDB0240531Organic oxygen compoundsOrganooxygen compoundsCarbohydrates and carbohydrate conjugates0.09 (0.05–0.13)<0.00014.50 × 10−4
neg_FT01863GlycylprolineNeg171.070.91M − HHMDB0000721Organic acids and derivativesCarboxylic acids and derivativesAmino acids, peptides, and analogues−0.08 (−0.11–−0.04)<0.00014.68 × 10−4
neg_FT095515-Caffeoylquinic acidNeg353.086.84M − HHMDB0240477Organic oxygen compoundsOrganooxygen compoundsAlcohols and polyols0.13 (0.07–0.19)<0.00014.68 × 10−4
neg_FT00341(R)-3-Hydroxyisobutyric acidNeg103.042.02M − HHMDB0000336Organic acids and derivativesHydroxy acids and derivativesBeta hydroxy acids and derivatives0.04 (0.02–0.06)<0.00014.75 × 10−4
neg_FT023461-(Malonylamino)cyclopropanecarboxylic acidNeg186.043.48M − HHMDB0031700Organic acids and derivativesCarboxylic acids and derivativesAmino acids, peptides, and analogues0.08 (0.04–0.12)<0.00017.96 × 10−4
neg_FT1446511-beta-Hydroxyandrosterone-3-glucuronideNeg481.246.63M − HHMDB0010351Organoheterocyclic compoundsIndoles and derivativesHydroxyindoles0.05 (0.02–0.07)<0.00018.07 × 10−4
Model: metabolites = age + race + BMI z-score + total energy intake. Abbreviations: ID, metabolic feature identification; MZ, mass-charge; RT, retention time; HMDB, human metabolome database.
Table
Table 3. List of significant metabolic features associated with BMI z-score in the validation set of samples.
Table 3. List of significant metabolic features associated with BMI z-score in the validation set of samples.
IDMetaboliteModeMZRTAdductHMDB IDSuper-ClassClassSub-ClassValidation Set
Estimate (95%CI)p-ValueFDR adj
pos_FT04673GlycylprolinePos173.090.75M + HHMDB0000721Organic acids and derivativesCarboxylic acids and derivativesAmino acids, peptides, and analogues−0.105 (−0.145–0.064)<0.00012.29 × 10−4
neg_FT01967CitrullineNeg174.080.81M − HHMDB0000904Organic acids and derivativesCarboxylic acids and derivativesAmino acids, peptides, and analogues0.117 (0.065–0.168)<0.00010.002
pos_FT096304-VinylsyringolPos243.090.73M + HHMDB0301746Phenylpropanoids and polyketidesStilbenes −0.079 (−0.117–0.041)<0.00010.004
neg_FT01863GlycylprolineNeg171.070.91M − HHMDB0000721Organic acids and derivativesCarboxylic acids and derivativesAmino acids, peptides, and analogues−0.076 (−0.114–0.039)<0.00010.005
neg_FT172073′-SialyllactoseNeg632.20.67M − HHMDB0000825Organic oxygen compoundsOrganooxygen compoundsCarbohydrates and carbohydrate conjugates0.055 (0.026–0.084)0.00030.012
pos_FT04847Estrone sulfatePos176.060.58M + 2HHMDB0001425Lipids and lipid-like moleculesSteroids and steroid derivativesSulfated steroids−0.056 (−0.087–0.025)0.00040.014
pos_FT10083CarnosinePos249.090.49M + NaHMDB0000033Organic acids and derivativesPeptidomimeticsHybrid peptides−0.151 (−0.234–0.069)0.00040.014
neg_FT01921Formiminoglutamic acidNeg173.051.31M − HHMDB0000854Organic acids and derivativesCarboxylic acids and derivativesAmino acids, peptides, and analogues0.075 (0.032–0.118)0.00070.020
pos_FT04858CitrullinePos176.10.78M + HHMDB0000904Organic acids and derivativesCarboxylic acids and derivativesAmino acids, peptides, and analogues0.048 (0.019–0.077)0.0010.028
neg_FT007884-HydroxyprolineNeg130.051.4M − HHMDB0000725Organic acids and derivativesCarboxylic acids and derivativesAmino acids, peptides, and analogues0.082 (0.031–0.132)0.0020.033
pos_FT09878Hydroxyprolyl-AsparaginePos246.10.69M + HHMDB0028858Organic acids and derivativesCarboxylic acids and derivativesAmino acids, peptides, and analogues−0.059 (−0.097–0.022)0.0020.033
pos_FT135922-HexenoylcarnitinePos296.126.85M + KHMDB0013161Lipids and lipid-like moleculesFatty AcylsFatty acid esters0.104 (0.04–0.167)0.0020.033
pos_FT03157L-GlutaminePos147.070.52M + HHMDB0000641Organic acids and derivativesCarboxylic acids and derivativesAmino acids, peptides, and analogues−0.065 (−0.106–0.024)0.0020.033
neg_FT05602InosineNeg267.070.67M − HHMDB0000195Nucleosides, nucleotides, and analoguesPurine nucleosides 0.038 (0.011–0.064)0.0050.073
neg_FT08407N-(2-Hydroxyphenyl)acetamide glucuronideNeg328.063.81M − HHMDB0240542Organic oxygen compoundsOrganooxygen compoundsCarbohydrates and carbohydrate conjugates0.081 (0.023–0.139)0.0060.092
pos_FT15625GalactosylhydroxylysinePos325.160.5M + HHMDB0000600Organic acids and derivativesCarboxylic acids and derivativesAmino acids, peptides, and analogues−0.048 (−0.084–0.013)0.0070.096
Model: metabolites = age + race + BMI z-score + total energy intake. Abbreviations: ID: metabolic feature identification; MZ, mass-charge; RT, retention time; HMDB, human metabolome database.
Table
Table 4. List of significant metabolic features associated with positive change in BMI using the validation set feature in the replication set (1-year follow-up).
Table 4. List of significant metabolic features associated with positive change in BMI using the validation set feature in the replication set (1-year follow-up).
IDMetabolitesModeMZRTAdductHMDB IDSuper-ClassClassSub-ClassReplication Set
Estimate (95%CI)p-ValueFDR adj
neg_FT01863GlycylprolineNeg173.090.75M + HHMDB0000721Organic acids and derivativesCarboxylic acids and derivativesAmino acids, peptides, and analogues−0.018 (−0.029–0.007)0.0020.031
neg_FT172073′-SialyllactoseNeg632.20.67M − HHMDB0000825Organic oxygen compoundsOrganooxygen compoundsCarbohydrates and carbohydrate conjugates0.009 (0.002–0.016)0.0060.038
neg_FT01921Formiminoglutamic acidNeg173.051.31M − HHMDB0000854Organic acids and derivativesCarboxylic acids and derivativesAmino acids, peptides, and analogues0.016 (0.004–0.028)0.0080.038
pos_FT04673GlycylprolinePos171.070.91M − HHMDB0000721Organic acids and derivativesCarboxylic acids and derivativesAmino acids, peptides, and analogues−0.014 (−0.025–0.003)0.010.038
neg_FT007884-HydroxyprolineNeg130.051.4M − HHMDB0000725Organic acids and derivativesCarboxylic acids and derivativesAmino acids, peptides, and analogues0.016 (0.003–0.03)0.0160.043
pos_FT04858CitrullinePos174.080.81M − HHMDB0000904Organic acids and derivativesCarboxylic acids and derivativesAmino acids, peptides, and analogues0.01 (0.002–0.018)0.0130.043
pos_FT096304-VinylsyringolPos243.090.73M + HHMDB0301746Phenylpropanoids and polyketidesStilbenes −0.01 (−0.02–0.001)0.0220.049
neg_FT01967CitrullineNeg176.10.78M + HHMDB0000904Organic acids and derivativesCarboxylic acids and derivativesAmino acids, peptides, and analogues0.012 (0.001–0.023)0.0250.049
neg_FT05602InosineNeg267.070.67M − HHMDB0000195Nucleosides, nucleotides, and analoguesPurine nucleosides 0.005 (0.0004–0.01)0.0330.058
pos_FT10083CarnosinePos249.090.49M + NaHMDB0000033Organic acids and derivativesPeptidomimeticsHybrid peptides−0.015 (−0.031–0.0004)0.0560.082
neg_FT08407N-(2-Hydroxyphenyl)acetamide glucuronideNeg328.063.81M − HHMDB0240542Organic oxygen compoundsOrganooxygen compoundsCarbohydrates and carbohydrate conjugates0.015 (−0.0004–0.032)0.0560.082
pos_FT03157L-GlutaminePos147.070.52M + HHMDB0000641Organic acids and derivativesCarboxylic acids and derivativesAmino acids, peptides, and analogues−0.006 (−0.014–0.0001)0.0860.114
pos_FT04847Estrone sulfatePos176.060.58M + 2HHMDB0001425Lipids and lipid-like moleculesSteroids and steroid derivativesSulfated steroids−0.004 (−0.011–0.001)0.1390.171
pos_FT09878Hydroxyprolyl-AsparaginePos246.10.69M + HHMDB0028858Organic acids and derivativesCarboxylic acids and derivativesAmino acids, peptides, and analogues−0.001 (−0.009–0.006)0.7510.858
pos_FT15625GalactosylhydroxylysinePos325.160.5M + HHMDB0000600Organic acids and derivativesCarboxylic acids and derivativesAmino acids, peptides, and analogues−0.0008 (−0.008–0.007)0.830.885
pos_FT135922-HexenoylcarnitinePos296.126.85M + KHMDB0013161Lipids and lipid-like moleculesFatty AcylsFatty acid esters0.00 (−0.017–0.017)11
Model: metabolites = age + race + BMI z-score + total energy intake. Abbreviations: ID, metabolic feature identification; MZ, mass-charge; RT, retention time; HMDB, human metabolome database.
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Singh, A.; Kinnebrew, G.; Hsu, P.-C.; Weng, D.Y.; Song, M.-A.; Reisinger, S.A.; McElroy, J.P.; Keller-Hamilton, B.; Ferketich, A.K.; Freudenheim, J.L.; et al. Untargeted Metabolomics and Body Mass in Adolescents: A Cross-Sectional and Longitudinal Analysis. Metabolites 2023, 13, 899. https://doi.org/10.3390/metabo13080899

AMA Style

Singh A, Kinnebrew G, Hsu P-C, Weng DY, Song M-A, Reisinger SA, McElroy JP, Keller-Hamilton B, Ferketich AK, Freudenheim JL, et al. Untargeted Metabolomics and Body Mass in Adolescents: A Cross-Sectional and Longitudinal Analysis. Metabolites. 2023; 13(8):899. https://doi.org/10.3390/metabo13080899

Chicago/Turabian Style

Singh, Amarnath, Garrett Kinnebrew, Ping-Ching Hsu, Daniel Y. Weng, Min-Ae Song, Sarah A. Reisinger, Joseph P. McElroy, Brittney Keller-Hamilton, Amy K. Ferketich, Jo L. Freudenheim, and et al. 2023. "Untargeted Metabolomics and Body Mass in Adolescents: A Cross-Sectional and Longitudinal Analysis" Metabolites 13, no. 8: 899. https://doi.org/10.3390/metabo13080899

APA Style

Singh, A., Kinnebrew, G., Hsu, P. -C., Weng, D. Y., Song, M. -A., Reisinger, S. A., McElroy, J. P., Keller-Hamilton, B., Ferketich, A. K., Freudenheim, J. L., & Shields, P. G. (2023). Untargeted Metabolomics and Body Mass in Adolescents: A Cross-Sectional and Longitudinal Analysis. Metabolites, 13(8), 899. https://doi.org/10.3390/metabo13080899

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop