Advertisement

Metabolomics study of blood pressure salt-sensitivity and hypertension

  • Mengyao Shi
    Affiliations
    Department of Epidemiology, School of Public Health and Jiangsu Key Laboratory of Preventive and Translational Medicine for Geriatric Diseases, Medical College of Soochow University, Suzhou, China

    Department of Epidemiology, Tulane University School of Public Health and Tropical Medicine, New Orleans, LA, United States
    Search for articles by this author
  • Jiang He
    Affiliations
    Department of Epidemiology, Tulane University School of Public Health and Tropical Medicine, New Orleans, LA, United States

    Department of Medicine, Tulane University School of Medicine, New Orleans, LA, United States
    Search for articles by this author
  • Changwei Li
    Affiliations
    Department of Epidemiology, Tulane University School of Public Health and Tropical Medicine, New Orleans, LA, United States
    Search for articles by this author
  • Xiangfeng Lu
    Affiliations
    Department of Epidemiology, Fuwai Hospital, National Center for Cardiovascular Diseases, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing, China

    Key Laboratory of Cardiovascular Epidemiology, Chinese Academy of Medical Sciences, Beijing, China
    Search for articles by this author
  • William J. He
    Affiliations
    Welch Center for Prevention, Epidemiology, and Clinical Research, Johns Hopkins University, Baltimore, MD, United States

    Department of Epidemiology, Johns Hopkins Bloomberg School of Public Health, Baltimore, MD, United States
    Search for articles by this author
  • Jie Cao
    Affiliations
    Department of Epidemiology, Fuwai Hospital, National Center for Cardiovascular Diseases, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing, China

    Key Laboratory of Cardiovascular Epidemiology, Chinese Academy of Medical Sciences, Beijing, China
    Search for articles by this author
  • Jing Chen
    Affiliations
    Department of Epidemiology, Tulane University School of Public Health and Tropical Medicine, New Orleans, LA, United States

    Department of Medicine, Tulane University School of Medicine, New Orleans, LA, United States
    Search for articles by this author
  • Ji-Chun Chen
    Affiliations
    Department of Epidemiology, Fuwai Hospital, National Center for Cardiovascular Diseases, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing, China

    Key Laboratory of Cardiovascular Epidemiology, Chinese Academy of Medical Sciences, Beijing, China
    Search for articles by this author
  • Lydia A. Bazzano
    Affiliations
    Department of Epidemiology, Tulane University School of Public Health and Tropical Medicine, New Orleans, LA, United States
    Search for articles by this author
  • Jian-Xin Li
    Affiliations
    Department of Epidemiology, Fuwai Hospital, National Center for Cardiovascular Diseases, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing, China

    Key Laboratory of Cardiovascular Epidemiology, Chinese Academy of Medical Sciences, Beijing, China
    Search for articles by this author
  • Hua He
    Affiliations
    Department of Epidemiology, Tulane University School of Public Health and Tropical Medicine, New Orleans, LA, United States
    Search for articles by this author
  • Dongfeng Gu
    Affiliations
    Department of Epidemiology, Fuwai Hospital, National Center for Cardiovascular Diseases, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing, China

    Key Laboratory of Cardiovascular Epidemiology, Chinese Academy of Medical Sciences, Beijing, China
    Search for articles by this author
  • Tanika N. Kelly
    Correspondence
    Corresponding author. Department of Epidemiology, Center for Public Health Genomics, Tulane University School of Public Health and Tropical Medicine, 1440 Canal Street, Suite 2000, New Orleans, LA, 70112, United States.
    Affiliations
    Department of Epidemiology, Tulane University School of Public Health and Tropical Medicine, New Orleans, LA, United States
    Search for articles by this author

      Highlights

      • One hundred and sixty one baseline metabolites were associated with BP salt-sensitivity.
      • BP salt-sensitivity was associated with changes in the values of 215 metabolites under low and/or high sodium conditions.
      • Among the 336 identified salt-sensitivity metabolites, 15 were robustly associated with hypertension.
      • The 15 salt-sensitivity metabolites explained an additional 7% of hypertension susceptibility beyond traditional risk factors.

      Abstract

      Background and aims

      Identify novel metabolite associations with blood pressure (BP) salt-sensitivity and hypertension.

      Methods and results

      The Genetic Epidemiology Network of Salt Sensitivity (GenSalt) Replication study includes 698 Chinese participants who underwent a 3-day baseline examination followed by a 7-day low-sodium feeding and 7-day high-sodium feeding. Latent mixture models identified three trajectories of blood pressure (BP) responses to the sodium interventions. We selected 50 most highly salt-sensitive and 50 most salt-resistant participants for untargeted metabolomics profiling. Multivariable adjusted mixed logistic regression models tested the associations of baseline metabolites with BP salt-sensitivity. Multivariable adjusted mixed linear regression models tested the associations of BP salt-sensitivity with metabolite changes during the sodium interventions. Identified metabolites were tested for associations with hypertension among 1249 Bogalusa Heart Study (BHS) participants using multiple logistic regression. Fifteen salt-sensitivity metabolites were associated with hypertension in the BHS. Baseline values of serine, 2-methylbutyrylcarnitine and isoleucine directly associated with high salt-sensitivity. Among them, serine indirectly associated with hypertension while 2-methylbutyrylcarnitine and isoleucine directly associated with hypertension. Baseline salt-sensitivity status predicted changes in 14 metabolites when switching to low-sodium or high-sodium interventions. Among them, glutamate, 1-carboxyethylvaline, 2-methylbutyrylcarnitine, 3-methoxytyramine sulfate, glucose, alpha-ketoglutarate, hexanoylcarnitine, gamma-glutamylisoleucine, gamma-glutamylleucine, and gamma-glutamylphenylalanine directly associated with hypertension. Conversely, serine, histidine, threonate and 5-methyluridine indirectly associated with hypertension. Together, these metabolites explained an additional 7% of hypertension susceptibility when added to a model including traditional risk factors.

      Conclusions

      Our findings contribute to the molecular characterization of BP response to sodium and provide novel biological insights into salt-sensitive hypertension.

      Keywords

      1. Introduction

      The number of individuals with hypertension has almost doubled in the past three decades, increasing to over 1.28 billion individuals with high blood pressure (BP) in 2021 [
      • Zhou B.
      • Carrillo-Larco R.M.
      • Danaei G.
      • Riley L.M.
      • Paciorek C.J.
      • Stevens G.A.
      • et al.
      Worldwide trends in hypertension prevalence and progress in treatment and control from 1990 to 2019: a pooled analysis of 1201 population-representative studies with 104 million participants.
      ]. As a leading contributor to morbidity and mortality globally, hypertension accounts for 8.7% of total disability-adjusted life-years and 14.0% of total deaths worldwide [
      • Forouzanfar M.H.
      • Liu P.
      • Roth G.A.
      • Ng M.
      • Biryukov S.
      • Marczak L.
      • et al.
      Global burden of hypertension and systolic blood pressure of at least 110 to 115mmHg, 1990–2015.
      ,
      • Stanaway J.D.
      • Afshin A.
      • Gakidou E.
      • Lim S.S.
      • Abate D.
      • Abate K.H.
      • et al.
      Global, regional, and national comparative risk assessment of 84 behavioural, environmental and occupational, and metabolic risks or clusters of risks for 195 countries and territories, 1990-2017: a systematic analysis for the Global Burden of Disease Study 2017.
      ]. Randomized controlled trials have established a direct, causal relationship between dietary sodium intake and BP [
      • He F.J.
      • Li J.
      • MacGregor G.A.
      Effect of longer term modest salt reduction on blood pressure: cochrane systematic review and meta-analysis of randomised trials.
      ,
      • Mozaffarian D.
      • Fahimi S.
      • Singh G.M.
      • Micha R.
      • Khatibzadeh S.
      • Engell R.E.
      • et al.
      Global sodium consumption and death from cardiovascular causes.
      ,
      • Afshin A.
      • Sur P.J.
      • Fay K.A.
      • Cornaby L.
      • Ferrara G.
      • Salama J.S.
      • et al.
      Health effects of dietary risks in 195 countries, 1990–2017: a systematic analysis for the Global Burden of Disease Study 2017.
      ]. However, BP responses to sodium intake vary considerably between individuals, a phenomenon known as BP salt-sensitivity [
      • He F.J.
      • Li J.
      • MacGregor G.A.
      Effect of longer term modest salt reduction on blood pressure: cochrane systematic review and meta-analysis of randomised trials.
      ,
      • Chen L.
      • He F.J.
      • Dong Y.
      • Huang Y.
      • Harshfield G.A.
      • Zhu H.
      Sodium reduction, metabolomic profiling, and cardiovascular disease risk in untreated black hypertensives: a randomized, double-blind, placebo-controlled trial.
      ,
      • Liu Y.
      • Shi M.
      • Dolan J.
      • He J.
      Sodium sensitivity of blood pressure in Chinese populations.
      ]. Those with high salt-sensitivity of BP have increased risks of hypertension and related cardiovascular disease [
      • He J.
      • Kelly T.N.
      • Zhao Q.
      • Li H.
      • Huang J.
      • Wang L.
      • et al.
      Genome-wide association study identifies 8 novel loci associated with blood pressure responses to interventions in Han Chinese.
      ,
      • Barba G.
      • Galletti F.
      • Cappuccio F.P.
      • Siani A.
      • Venezia A.
      • Versiero M.
      • et al.
      Incidence of hypertension in individuals with different blood pressure salt-sensitivity: results of a 15-year follow-up study.
      ,
      • Weinberger M.H.
      • Fineberg N.S.
      • Fineberg S.E.
      • Weinberger M.
      Salt sensitivity, pulse pressure, and death in normal and hypertensive humans.
      ,
      • Morimoto A.
      • Uzu T.
      • Fujii T.
      • Nishimura M.
      • Kuroda S.
      • Nakamura S.
      • et al.
      Sodium sensitivity and cardiovascular events in patients with essential hypertension.
      ], suggesting that lifestyle interventions targeting this subgroup may be particularly relevant in preventing these conditions. Unfortunately, accurate measurement of BP salt-sensitivity remains costly and time-intensive, requiring a well-controlled feeding study or an acute in-patient protocol [
      • Elijovich F.
      • Weinberger M.H.
      • Anderson C.A.M.
      • Appel L.J.
      • Bursztyn M.
      • Cook N.R.
      • et al.
      Salt sensitivity of blood pressure : a scientific statement from the American Heart Association.
      ]. In contrast, molecular technologies are evolving rapidly, reducing costs and allowing for quick and comprehensive biomarker profiling. These scientific advancements provide opportunity to explore molecular phenotyping as an alternative strategy for characterizing sodium-sensitivity of BP in individuals.
      The human metabolome is the end-product of complex interactions between genomic, transcriptomic, proteomic, and environmental exposures [
      • Ussher J.R.
      • Elmariah S.
      • Gerszten R.E.
      • Dyck J.R.B.
      The emerging role of metabolomics in the diagnosis and prognosis of cardiovascular disease.
      ,
      • McGarrah R.W.
      • Crown S.B.
      • Zhang G.F.
      • Shah S.H.
      • Newgard C.B.
      Cardiovascular metabolomics.
      ,
      • Arnett D.K.
      • Claas S.A.
      Omics of blood pressure and hypertension.
      ]. Given the proximity of metabolome to phenome, metabolomics research has become increasingly utilized to identify novel biomarkers for complex phenotypes. While several reports have identified novel metabolite associations with BP [
      • Menni C.
      • Graham D.
      • Kastenmüller G.
      • Alharbi N.H.J.
      • Alsanosi S.M.
      • Mcbride M.
      • et al.
      Metabolomic identification of a novel pathway of blood pressure regulation involving hexadecanedioate.
      ,
      • Dietrich S.
      • Floegel A.
      • Weikert C.
      • Pischon T.
      • Boeing H.
      • Drogan D.
      Identification of serum metabolites associated with incident hypertension in the European prospective investigation into cancer and nutrition-potsdam study.
      ,
      • He W.J.
      • Li C.
      • Mi X.
      • Shi M.
      • Gu X.
      • Bazzano L.A.
      • et al.
      An untargeted metabolomics study of blood pressure: findings from the Bogalusa Heart Study.
      ], only one metabolomics study of salt-sensitivity of BP was conducted previously among a subsample of 103 Dietary Approaches to Stop Hypertension-Sodium (DASH-sodium) participants [
      • Cheng Y.
      • Song H.
      • Pan X.
      • Xue H.
      • Wan Y.
      • Wang T.
      • et al.
      Urinary metabolites associated with blood pressure on a low- or high-sodium diet.
      ]. This targeted analysis of 57 amino and tricarboxylic acid cycle metabolites identified several metabolites that, together, improved the classification of BP salt-sensitivity beyond race, age, and sex. Although this work was promising, discovery (untargeted) metabolomics approaches that comprehensively identify metabolites across known and unknown biological pathways are still needed. This type of research could identify a broader array of molecular biomarkers for salt-sensitivity of BP, contributing to risk stratification efforts and our general understanding of the biological pathways underlying this complex phenotype.
      The current report describes findings from the first untargeted metabolomics study of BP salt-sensitivity, which leveraged stringently collected blood pressure and urinary metabolomics data from 100 Han Chinese participants who underwent a 7-day low-sodium and 7-day high-sodium feeding protocol. Because salt-sensitivity is a recognized hypertension endophenotype [
      • He J.
      • Kelly T.N.
      • Zhao Q.
      • Li H.
      • Huang J.
      • Wang L.
      • et al.
      Genome-wide association study identifies 8 novel loci associated with blood pressure responses to interventions in Han Chinese.
      ,
      • Barba G.
      • Galletti F.
      • Cappuccio F.P.
      • Siani A.
      • Venezia A.
      • Versiero M.
      • et al.
      Incidence of hypertension in individuals with different blood pressure salt-sensitivity: results of a 15-year follow-up study.
      ], identified metabolites were further tested for associations with this complex condition in a large, bi-racial sample of 1249 participants from the Bogalusa Heart study (BHS) who underwent untargeted metabolomics profiling and stringent clinical examination during the 2013 to 2016 study visit.

      2. Method

      2.1 The GenSalt studies

      The GenSalt studies, including the original GenSalt (N = 1906) and subsequent GenSalt Replication (N = 698) studies, were separate family-based feeding studies conducted in two independent samples from rural areas of northern China using identical research protocols. The original GenSalt study was conducted from October 2003 to July 2005 while the GenSalt Replication study was conducted from April 2010 to November 2010. Community-based BP screenings were conducted to identify potential probands aged 18–60 years and their sibling, spouses and offspring aged 16 or older. Probands were defined as those with mean systolic BP between 130 and 160 mmHg, and/or mean diastolic BP between 85 and 100 mmHg and no use of antihypertension medications. Individuals who had stage 2 hypertension, secondary hypertension, and a history of clinical cardiovascular disease or diabetes, or were pregnant, heavy alcohol drinkers, taking anti-hypertension medications, or currently on a low-sodium diet were excluded from the feeding study. For the current study, we selected the 50 most salt-sensitive and 50 most salt-resistant participants of the GenSalt Replication study to undergo untargeted metabolomics profiling. Selected participants included those with the largest and smallest BP responses to the dietary sodium interventions. Institutional review boards at all participating institutions approved the GenSalt studies. Written informed consents were obtained from all GenSalt participants at baseline prior to the interventions. The consent form was general and included consent for the use of biospecimens for future molecular research.

      2.2 GenSalt baseline data collection

      Standard questionnaires were administered at baseline to GenSalt participants by trained staff, to collect information on demographic characteristics (including age and gender), lifestyle risk factors (including physical activity and alcohol drinking), and personal medical history. Physical examinations were conducted during the baseline examination to collect anthropometric and BP measures. Body weight and height were measured twice at baseline with participants in light clothing without shoes. Body mass index (BMI) was calculated as weight in kilograms divided by height in square meters and the mean of the two measures was used for analyses. Obesity was defined as BMI ≥30 kg/m2. Blood pressure (BP) was measured in the morning in triplicate following a standard protocol during each day of the 3-day baseline examination by trained and certified staff using a Hawksley random-zero sphygmomanometer. BP was measured with the participant in a sitting position following 5 min of rest. Participants were advised to avoid alcohol drinking, cigarette smoking, having coffee or tea, and exercise for at least 30 min prior to their measurements. Hypertension was defined according to the 2017 American College of Cardiology and American Heart Association (ACC/AHA) guidelines as systolic BP greater than or equal to 130 mmHg, diastolic BP greater than or equal to 80 mmHg, or use of antihypertension medication [
      • Whelton P.K.
      • Carey R.M.
      • Aronow W.S.
      • Casey D.E.
      • Collins K.J.
      • Himmelfarb C.D.
      • et al.
      2017 ACC/AHA/AAPA/ABC/ACPM/AGS/APhA/ASH/ASPC/NMA/PCNA guideline for the prevention, detection, evaluation, and management of high blood pressure in adults a report of the American College of Cardiology/American Heart Association Task Force on Clinical practice guidelines.
      ], with systolic and diastolic BP estimated as the mean of the nine baseline BP measures for each study participant.

      2.3 GenSalt dietary intervention

      After the 3-day baseline observation, the study participants received a dietary intervention including a 7-day low-sodium diet (3 g of sodium chloride or 51.3 mmol of sodium per day) followed by a 7-day high-sodium diet (18 g of sodium chloride or 307.8 mmol of sodium per day). BP was measured in the morning in triplicate following a standard protocol on days 2, 5, 6 and 7 of each intervention period using a protocol identical to that of the baseline examination. All BP observers were blinded to the dietary intervention phase. During the dietary sodium interventions, dietary potassium intake remained unchanged. Total energy intake was varied according to each participant's baseline energy intake. All study foods were cooked without salt, and prepackaged salt was added to the individual study participant's meal when it was served by the study staff. To ensure study participants' compliance to the intervention program, they were required to have their breakfast, lunch, and dinner at the study kitchen under supervision of the study staff during the entire study period. The study participants were instructed to avoid consuming any foods that were not provided by study personnel. One 24-h and two 8-h overnight urinary specimens were collected at the baseline examination and on the last 3 days of each intervention period to measure sodium and potassium. Overnight urinary sodium excretion was converted to 24-h values based on formulas developed from data obtained in a subgroup of study participants [
      GenSalt: rationale, design, methods and baseline characteristics of study participants.
      ]. The results from the 24-h urinary excretions of sodium showed excellent compliance with the study diet [
      • Gu D.
      • Zhao Q.
      • Chen J.
      • Chen J.C.
      • Huang J.
      • Bazzano L.A.
      • et al.
      Reproducibility of blood pressure responses to dietary sodium and potassium interventions: the GenSalt study.
      ].

      2.4 Selection of highly salt-sensitive and salt-resistant GenSalt participants

      Latent variable mixture modeling was used to identify subgroups sharing similar underlying trajectories of BP responses to low-sodium and high-sodium interventions, as implemented by the SAS Proc Traj procedure [
      • Jones B.L.
      • Nagin D.S.
      • Roeder K.
      A SAS procedure based on mixture models for estimating developmental trajectories.
      ]. Systolic BP responses from baseline to low-sodium intervention and responses from low-sodium to high-sodium intervention on days 2, 5, 6, and 7 of each intervention period were used for modeling. We fit models up to the optimal number of trajectories by comparing the Bayesian Information Criterion for each set of trajectories. The model with 3 trajectories fit best, and participants were classified into the highly salt-sensitive, moderately salt-sensitive, or salt-resistant trajectory groups with excellent discrimination [
      • He J.
      • Huang J.-F.
      • Li C.
      • Chen J.
      • Lu X.
      • Chen J.-C.
      • et al.
      Sodium sensitivity, sodium resistance, and incidence of hypertension.
      ]. The 50 participants with the highest probability of high salt-sensitivity and the 50 GenSalt Replication study participants with the highest probability of salt-resistance were selected for metabolomics profiling.

      2.5 GenSalt metabolomics profiling

      Untargeted, ultrahigh performance liquid chromatography-tandem mass spectroscopy (UPLC-MS/MS) was conducted by Metabolon© using GenSalt 24-h urine samples collected at three time-points, including the first day of baseline examination and the fifth days of the low-sodium and high-sodium interventions. All urine samples were stored at −80 °C since the time of collection. Untargeted metabolomics profiling resulted in the detection and semi-quantification of 1445 metabolites. These included 972 known biochemical compounds in pathways related to amino acids (n = 267), carbohydrates (n = 32), cofactors and vitamins (n = 39), energy (n = 15), lipids (n = 170), nucleotides (n = 59), peptides (n = 40), xenobiotics (n = 293), and partially characterized molecules (n = 57). An additional 473 unnamed compounds currently lacking chemical standards were also detected. These metabolites were labeled with an “X” followed by numbers (e.g., X-12345) and may be identified upon the eventual acquisition of a matching purified standard (or via classical structural analysis).
      Prior to the statistical analysis, additional quality control and manipulation of the metabolite data was undertaken. First, urine metabolomics underwent osmolality-based normalization, which is comparable to creatinine-based normalization [
      • Khamis M.M.
      • Holt T.
      • Awad H.
      • El-Aneed A.
      • Adamko D.J.
      Comparative analysis of creatinine and osmolality as urine normalization strategies in targeted metabolomics for the differential diagnosis of asthma and COPD.
      ,
      • Kennedy A.D.
      • Miller M.J.
      • Beebe K.
      • Wulff J.E.
      • Evans A.M.
      • Miller L.A.D.
      • et al.
      Metabolomic profiling of human urine as a screen for multiple inborn errors of metabolism.
      ] and is used to reduce variability in metabolite values due to differences in total urinary solute concentration [
      • Nam S.L.
      • Paulina de la Mata A.
      • Dias R.P.
      • Harynuk J.J.
      Towards standardization of data normalization strategies to improve urinary metabolomics studies by gc×gc-tofms.
      ]. Next, data filtering removed 73 metabolites that were missing or below the detection threshold in more than 80% of samples. Among the 1372 metabolites passing quality control, 76 were missing or below the detection threshold in 50%–80% of the samples. These metabolites were analyzed as ordinal variables after categorization into one of three mutually exclusive groups: 1) missing or below-the-detection-limit; 2) below the median of detectable values; or 3) greater than or equal to the median of detectable values. The remaining 1296 metabolites were analyzed as continuous variables, where half of the minimum observed value for a particular metabolite was imputed among those with missing or below-the-detection-limit values.

      2.6 The BHS

      The BHS was designed to study the natural history of hypertension and cardiovascular disease (CVD) among a biracial sample (35% African American and 65% white) of residents from Bogalusa, Louisiana. The current BHS core cohort includes 1298 participants born between 1959 and 1979 who were examined at least twice during childhood and twice during adulthood for CVD risk factors. A total of 1249 participants underwent untargeted metabolomics profiling and provided measures of BP and other important covariables in the 2013 to 2016 visit cycle.
      Data collection methods were described previously in detail [
      • Foster T.A.
      • Berenson G.S.
      Measurement error and reliability in four pediatric cross-sectional surveys of cardiovascular disease risk factor variables-The Bogalusa Heart Study.
      ]. In brief, a standard questionnaire was used to collect information on demographic characteristics, lifestyle, and medical history. Height and weight were measured in duplicate, and BP was measured in triplicate using stringent study protocols. Untargeted metabolomics profiling of serum samples from BHS participants was carried out by Metabolon [
      • He W.J.
      • Li C.
      • Mi X.
      • Shi M.
      • Gu X.
      • Bazzano L.A.
      • et al.
      An untargeted metabolomics study of blood pressure: findings from the Bogalusa Heart Study.
      ]. Metabolomics data quality control was conducted using procedures identical to those of the GenSalt study. Hypertension was again defined according to the 2017 ACC/AHA guidelines [
      • Whelton P.K.
      • Carey R.M.
      • Aronow W.S.
      • Casey D.E.
      • Collins K.J.
      • Himmelfarb C.D.
      • et al.
      2017 ACC/AHA/AAPA/ABC/ACPM/AGS/APhA/ASH/ASPC/NMA/PCNA guideline for the prevention, detection, evaluation, and management of high blood pressure in adults a report of the American College of Cardiology/American Heart Association Task Force on Clinical practice guidelines.
      ]. Written informed consent was obtained when the participants enrolled in the 2013–2016 visit cycle. The consent form included consent for the use of biospecimen for future molecular research. The study was approved by the Tulane University Biomedical Institutional Review.

      2.7 Statistical analysis

      Characteristics of study participants were presented as means and standard deviations (SDs) for continuous variables and as frequencies (percentages) for categorical variables.
      Prior to association analyses, the R package, bestnormalize, was used to normalize continuous metabolite distributions. All metabolites were then standardized to mean = 0 and SD = 1, so that the scales across metabolites were comparable. Mixed logistic regression models were used to assess whether baseline metabolites predicted salt-sensitivity after adjustment for baseline age, gender, physical activity, alcohol drinking, obesity, and hypertension. To identify metabolites that were altered differently by sodium intervention according to salt-sensitivity status, mixed linear regressions models were employed. Here, we examined the associations of salt-sensitivity status with metabolite changes in response to low-sodium intervention and also in response to high-sodium intervention, again adjusting for baseline age, gender, physical activity, alcohol drinking, obesity, and hypertension. The same mixed linear regression models were also used to estimate adjusted mean changes in metabolite values in response to low-sodium and high-sodium interventions according to salt-sensitivity status. Metabolites that achieved nominal significance (P < 0.05) among GenSalt participants were matched to Kyoto Encyclopedia of Genes and Genomes (KEGG) metabolomics pathway database using the Metaboanalyst 5.0 [
      • Pang Z.
      • Chong J.
      • Zhou G.
      • de Lima Morais D.A.
      • Chang L.
      • Barrette M.
      • et al.
      MetaboAnalyst 5.0: narrowing the gap between raw spectra and functional insights.
      ], a web-based tool, to conduct pathway enrichment and pathway topological analyses.
      Nominally significant metabolites were also each tested for association with hypertension in the BHS using multiple logistic regression models that adjusted for age, gender, race, physical activity, alcohol drinking, and obesity. Bonferroni-corrected significance thresholds were employed at this stage of the analysis (0.05/the number of metabolites moved forward from each GenSalt analysis). Metabolites achieving Bonferroni corrected significance in the BHS were categorized into quartiles and their dose response relationships with hypertension were further explored, again employing multivariable logistic regression models with adjustment for age, gender, race, physical activity, alcohol drinking, and obesity. We further assessed the added predictive value and discrimination of the identified salt-sensitivity metabolites together on hypertension status among BHS participants. To minimize the influence of the correlations between the identified metabolites, principal component analysis was done among the identified metabolites and principal components (PCs) with eigenvalue greater than one were selected. We assessed both the incremental R2 (on the liability scale) and change in c-statistics when adding the selected PCs to a model adjusting for traditional risk factors.
      Sensitivity analyses to investigate the influence of kidney function on our findings were done by further adjusting for estimated glomerular filtration rate (eGFR) in analyses of BHS participants. Furthermore, to investigate whether imputation of missing and below-the-detection-limit values influenced our findings, sensitivity analyses were carried out excluding participants with missing or below-the-detection-limit values.

      3. Results

      The participants’ flowchart was presented as Supplementary Fig. 1 and baseline characteristics were shown in Table 1. Highly salt-sensitive GenSalt participants were, on average, older and more likely to be current drinkers and have hypertension compared to their salt-resistant counterparts. BHS participants were predominantly female, and more than half were white, current drinkers, obese, and hypertensive.
      Table 1Baseline characteristics of GenSalt and BHS participants.
      GenSalt StudyBHS (N = 1249)
      Salt-Sensitive (N = 50)Salt-Resistant (N = 50)
      Age, years, mean (SD)51.8 (6.45)42.2 (10.6)48.2 (5.28)
      Male, n (%)23 (46%)26 (52%)515 (41%)
      Race, n (%)
       Asian50 (100%)50 (100%)0 (0%)
       White0 (0%)0 (0%)822 (66%)
       Black0 (0%)0 (0%)427 (34%)
      Current Drinking, n (%)15 (30%)6 (12%)696 (56%)
      Physical Activity, MET-hours per day, mean (SD)20.6 (12.0)18.4 (11.7)12.0 (10.9)
      Obesity, n (%)8 (16%)9 (18%)644 (52%)
      Hypertension, n (%)45 (90%)11 (22%)775 (62%)
      Anti-hypertension medication, n (%)0 (0%)0 (0%)433 (35%)
      BHS: Bogalusa Heart Study; SD: standard deviation; MET: metabolic equivalent.

      3.1 Analyses of baseline metabolite values

      Although no metabolites achieved Bonferroni corrected significance (P < 3.64 × 10−5), a total of 161 baseline metabolites were nominally associated with BP salt-sensitivity status among GenSalt participants (P < 0.05). Twenty-one were inversely associated with high salt-sensitivity (associated with salt-resistance) and 140 were positively associated with high salt-sensitivity (Supplementary Fig. 2A and Supplementary Table 1).
      Among the 161 baseline metabolites that predicted salt-sensitivity of BP in GenSalt, 71 were detected and 4 were further associated with hypertension in BHS participants, including 3 named metabolites and 1 unnamed metabolite (Table 2 and Supplementary Table 1). All of the named metabolites came from amino acid sub-pathways, including serine from the glycine, serine and threonine metabolism sub-pathway; and 2-methylbutyrylcarnitine and isoleucine from the leucine, isoleucine and valine metabolism sub-pathway. All three named metabolites conferred increased odds of high salt-sensitivity among GenSalt participants. They also conferred increased odds of hypertension in the BHS except serine, which had a negative association with hypertension (Table 2). The three known metabolites demonstrated significant graded dose response associations with hypertension when categorized into quartiles (Fig. 1).
      Table 2Associations of baseline metabolites, metabolite change in response to low-sodium and high-sodium interventions with high salt-sensitivity among GenSalt participants and corresponding metabolite associations with hypertension among BHS participants.
      MetaboliteGenSalt BaselineGenSalt Low-Sodium InterventionGenSalt High-Sodium InterventionBHS
      OR
      Adjusted for age, gender, physical activity, drinking, obesity and hypertension.
      ,
      Beta estimate corresponds to each standard deviation larger metabolite value at baseline.
      95% CIP valueBeta
      Adjusted for age, gender, physical activity, drinking, obesity and hypertension.
      ,
      Beta estimate corresponds to each standard deviation larger metabolite change during low-sodium intervention.
      SEP valueBeta
      Adjusted for age, gender, physical activity, drinking, obesity and hypertension.
      ,
      Beta estimate corresponds to each standard deviation larger metabolite change during high-sodium intervention.
      SEP valueOR
      Adjusted for age, gender, race, physical activity, drinking, and obesity.
      95% CIP value
      Glutamate0.52(0.24, 1.13)9.78E-02−0.480.212.34E-020.440.331.85E-011.42(1.22, 1.65)3.90E-06
      Serine2.13(1.20, 3.78)1.00E-02−1.060.278.84E-050.720.291.17E-020.76(0.67, 0.86)2.49E-05
      Histidine0.68(0.30, 1.54)3.60E-01−0.500.287.29E-020.640.234.70E-030.74(0.64, 0.85)1.88E-05
      1-carboxyethylvaline0.97(0.47, 2.00)9.38E-010.110.327.25E-01−0.610.304.60E-021.44(1.25, 1.66)3.83E-07
      2-methylbutyrylcarnitine (C5)2.02(1.14, 3.60)1.65E-02−0.250.253.12E-010.580.304.86E-021.32(1.14, 1.51)1.25E-04
      Isoleucine2.45(1.07, 5.60)3.36E-02−0.580.305.43E-020.290.354.08E-011.30(1.13, 1.50)2.76E-04
      3-methoxytyramine sulfate1.88(0.87, 4.06)1.10E-01−0.280.273.03E-01−0.650.313.39E-021.31(1.15, 1.50)7.16E-05
      Glucose2.01(0.89, 4.49)9.10E-02−0.110.317.12E-010.770.298.61E-031.49(1.30, 1.71)9.63E-09
      Threonate1.29(0.61, 2.74)5.03E-01−0.710.278.86E-030.390.342.46E-010.73(0.65, 0.83)2.28E-06
      Alpha-ketoglutarate1.37(0.55, 3.42)5.00E-010.470.277.38E-02−0.610.272.29E-021.33(1.17, 1.51)1.94E-05
      Hexanoylcarnitine (C6)0.89(0.38, 2.05)7.76E-01−0.390.281.52E-010.840.261.48E-031.31(1.15, 1.49)6.18E-05
      5-methyluridine (ribothymidine)2.28(0.82, 6.34)1.14E-01−0.870.281.64E-030.430.382.61E-010.71(0.62, 0.81)7.06E-07
      Gamma-glutamylisoleucine
      Indicates compounds that have not been officially confirmed based on a standard.
      1.14(0.55, 2.38)7.22E-01−0.680.291.82E-020.460.256.96E-021.50(1.29, 1.74)5.99E-08
      Gamma-glutamylleucine0.95(0.37, 2.41)9.11E-01−0.610.293.43E-020.620.261.60E-021.39(1.20, 1.62)1.55E-05
      Gamma-glutamylphenylalanine0.96(0.43, 2.13)9.16E-01−0.630.303.41E-020.740.301.45E-021.42(1.23, 1.64)1.91E-06
      Bold indicates the significance level is 0.05 for GenSalt Baseline, Low-sodium intervention and High-sodium intervention; and significance level is 3.10×10-4 for BHS.
      a Adjusted for age, gender, physical activity, drinking, obesity and hypertension.
      b Adjusted for age, gender, race, physical activity, drinking, and obesity.
      c Beta estimate corresponds to each standard deviation larger metabolite value at baseline.
      d Beta estimate corresponds to each standard deviation larger metabolite change during low-sodium intervention.
      e Beta estimate corresponds to each standard deviation larger metabolite change during high-sodium intervention.
      f Indicates compounds that have not been officially confirmed based on a standard.
      Figure 1
      Figure 1Plots depicting associations of hypertension among BHS participants with identified metabolites.

      3.2 Analyses of metabolite responses to sodium intervention

      Among GenSalt participants, salt-sensitivity status nominally associated with changes in the values of 97 metabolites when switching from usual diet to the low-sodium intervention (Supplementary Fig. 2B and Supplementary Table 2), as well as changes in the values of 153 metabolites when switching from low-sodium to high-sodium intervention (Supplementary Fig. 2C and Supplementary Table 3). None achieved Bonferroni corrected significance. Among the 215 metabolites differentially altered by low-sodium and/or high-sodium interventions according to salt-sensitivity status, 112 metabolites were detected in the BHS. Of those, 16 metabolites, including 14 named metabolites, also associated with hypertension in the BHS. Adjusted mean changes of these 14 metabolites along with isoleucine during low-sodium and high-sodium intervention according to salt-sensitivity status are presented in Fig. 2 and include: glutamate, serine, histidine, 1-carboxyethylvaline, 2-methylbutyrylcarnitine, isoleucine and 3-methoxytyramine sulfate from amino acid sub-pathways (glutamate metabolism; glycine, serine and threonine metabolism; histidine metabolism; leucine, isoleucine and valine metabolism; and tyrosine metabolism); glucose from a carbohydrate sub-pathway (glycolysis, gluconeogenesis, and pyruvate metabolism); threonate from a cofactors and vitamins sub-pathway (ascorbate and aldarate metabolism); alpha-ketoglutarate from an energy sub-pathway (TCA cycle); hexanoylcarnitine from a lipid sub-pathway (fatty acid metabolism); 5-methyluridine from a nucleotide sub-pathway (pyrimidine metabolism); along with gamma-glutamylisoleucine, gamma-glutamylleucine, and gamma-glutamylphenylalanine from a peptide sub-pathway (gamma-glutamyl amino acid).
      Figure 2
      Figure 2Adjusted mean changes in metabolite value during A.) Low-sodium intervention; and B.) High-sodium intervention.
      Compared to salt-resistance of BP, high salt-sensitivity of BP associated with relative decreases in 1-carboxyethylvaline and alpha-ketoglutarate during high-sodium intervention. For the remaining 12 metabolites, high salt-sensitivity of BP associated with relative increases metabolite values in response to high-sodium, or conversely, relative decreases in metabolite values in response to low-sodium. Each standard deviation or quartile increase in glutamate, 1-carboxyethylvaline, 2-methylbutyrylcarnitine, 3-methoxytyramine sulfate, glucose, alpha-ketoglutarate, hexanoylcarnitine, gamma-glutamylisoleucine, gamma-glutamylleucine, and gamma-glutamylphenylalanine metabolites associated with higher odds of hypertension in the BHS. In contrast, each standard deviation or quartile increase in serine, histidine, threonate, and 5-methyluridine associated with lower odds of hypertension in the BHS (Table 2 and Fig. 2).
      Results of sensitivity analyses further adjusting for eGFR were consistent with those of the main analysis (Supplementary Table 4). Among four identified metabolites with missing values imputed to half of the detection limit, sensitivity analyses excluding those participants with imputed values were also similar to those of the main analysis (Supplementary Table 5).

      3.3 Pathway analysis

      Supplementary Table 6 shows the forty KEGG biochemical pathways that were identified among salt-sensitive metabolites. Two biochemical pathways, aminoacyl-tRNA biosynthesis (P = 1.24 × 10−5, FDR = 1.05 × 10−3, pathway impact = 0.17) and arginine biosynthesis (P = 9.08 × 10−4, FDR = 3.81 × 10−2, pathway impact = 0.19), demonstrated significant enrichment for identified metabolites (Supplementary Table 6 and Supplementary Fig. 3).

      3.4 Added value of salt-sensitivity metabolites in the prediction of hypertension

      Supplementary Fig. 4 depicts the pairwise Pearson correlations of the 15 salt-sensitivity metabolites that associated with hypertension in the BHS. Aside from the strong correlations of gamma-glutamylisoleucine and gamma-glutamylphenylalanine (r = 0.81), the pairwise correlations of metabolites were generally modest to moderate (r < 0.7), with 5 PCs explaining 66.8% of the variance in the 15 metabolites (Supplementary Fig. 5). Addition of these 5 PCs to the age, gender, race, physical activity, drinking, and obesity adjusted model captured an additional 7.1% of the variance in hypertension risk (R2 on the liability scale) and improved the C-statistic from 0.73 to 0.78 (P < 0.0001; Table 3 and Supplementary Fig. 6).
      Table 3Comparison of R2 and C-statistics before and after addition of identified salt-sensitivity metabolites.
      R2C-statistics
      Model 1
      Model 1 was adjusted for age, gender, race, physical activity, drinking, and obesity.
      0.14270.7297
      Model 1
      Model 1 was adjusted for age, gender, race, physical activity, drinking, and obesity.
      + 5 PCAs
      0.21400.7782
      Difference0.07130.0485
      P value<.0001<.0001
      a Model 1 was adjusted for age, gender, race, physical activity, drinking, and obesity.

      4. Discussion

      In the first untargeted metabolomics study of salt-sensitivity of BP, we identified 161 baseline metabolites that were associated with BP salt-sensitivity. In addition, salt-sensitivity status was associated with changes in the values of 215 metabolites under low and/or high sodium conditions. Among the 336 identified salt-sensitivity metabolites, 15 were robustly associated with hypertension in a large, biracial sample of BHS participants and included 7 metabolites, nearly half, from amino acid pathways, along with 3 peptide metabolites and 1 metabolite from each of the carbohydrate, cofactor and vitamin, energy, lipid, and nucleotide pathways. An important role for amino acids in BP salt-sensitivity was further supported by pathway analyses which demonstrated significant enrichment of identified metabolites with two amino acid-related KEGG pathways, aminoacyl-tRNA and arginine biosynthesis. Together, the 15 salt-sensitivity metabolites provided additional predictive value for hypertension, significantly improving discrimination and explaining a further 7.1% of hypertension risk beyond traditional risk factors. In aggregate, our findings implicate new molecular mechanisms underlying BP response to dietary sodium intake and identify novel biomarkers of individual sodium-sensitivity and hypertension status.
      Among the seven identified salt-sensitivity metabolites from the amino acid pathway, serine and histidine showed inverse associations with hypertension. Baseline serine values positively associated with high salt-sensitivity, while salt-sensitivity status predicted changes in both metabolites during sodium interventions. Given previous reports of high salt-sensitivity as an important risk factor for hypertension [
      • He J.
      • Kelly T.N.
      • Zhao Q.
      • Li H.
      • Huang J.
      • Wang L.
      • et al.
      Genome-wide association study identifies 8 novel loci associated with blood pressure responses to interventions in Han Chinese.
      ,
      • Barba G.
      • Galletti F.
      • Cappuccio F.P.
      • Siani A.
      • Venezia A.
      • Versiero M.
      • et al.
      Incidence of hypertension in individuals with different blood pressure salt-sensitivity: results of a 15-year follow-up study.
      ], the positive baseline association of serine with high salt-sensitivity and inverse relation with hypertension status may seem counterintuitive. However, there is increasing evidence that both high salt-sensitivity and salt-resistance of BP increase hypertension risk compared to moderate salt-sensitivity status [
      • He J.
      • Huang J.-F.
      • Li C.
      • Chen J.
      • Lu X.
      • Chen J.-C.
      • et al.
      Sodium sensitivity, sodium resistance, and incidence of hypertension.
      ,
      • Nierenberg J.L.
      • Li C.
      • He J.
      • Gu D.
      • Chen J.
      • Lu X.
      • et al.
      Blood pressure genetic risk score predicts blood pressure responses to dietary sodium and potassium the GenSalt Study (Genetic epidemiology network of salt sensitivity).
      ] Since we compared highly salt-sensitive to salt-resistant participants, these findings may reflect beneficial effects of increased serine on salt-resistance of BP. Our findings of protective associations of both serine and histidine with hypertension are consistent with previous reports. For example, serine plays an important role in the production of nitric oxide and antioxidants [
      • Dietrich S.
      • Floegel A.
      • Weikert C.
      • Pischon T.
      • Boeing H.
      • Drogan D.
      Identification of serum metabolites associated with incident hypertension in the European prospective investigation into cancer and nutrition-potsdam study.
      ,
      • He W.J.
      • Li C.
      • Mi X.
      • Shi M.
      • Gu X.
      • Bazzano L.A.
      • et al.
      An untargeted metabolomics study of blood pressure: findings from the Bogalusa Heart Study.
      ,
      • Natarajan P.
      • Peloso G.M.
      • Zekavat S.M.
      • Montasser M.
      • Ganna A.
      • Chaffin M.
      • et al.
      Deep-coverage whole genome sequences and blood lipids among 16,324 individuals.
      ], which have been shown to have BP lowering effects. In addition, histidine has important anti-oxidant and anti-inflammatory properties [
      • Yu B.
      • Li A.H.
      • Muzny D.
      • Veeraraghavan N.
      • De Vries P.S.
      • Bis J.C.
      • et al.
      Association of rare loss-of-function alleles in HAL, serum histidine: levels and incident coronary heart disease.
      ]. Higher values of this metabolite have been inversely related to blood pressure [
      • Mels C.M.
      • Delles C.
      • Louw R.
      • Schutte A.E.
      Central systolic pressure and a nonessential amino acidmetabolomics profile: TheAfrican Prospective study on the Early Detection and Identification of Cardiovascular disease and Hypertension.
      ] and associated with lower incidence of coronary heart disease [
      • Yu B.
      • Li A.H.
      • Muzny D.
      • Veeraraghavan N.
      • De Vries P.S.
      • Bis J.C.
      • et al.
      Association of rare loss-of-function alleles in HAL, serum histidine: levels and incident coronary heart disease.
      ]. Although mechanistic studies are needed, we provide early evidence that serine and histidine may lower hypertension risk through sodium-related pathways.
      Among the remaining five salt-sensitivity metabolites from amino acid pathways, glutamate, 1-carboxyethylvaline, 2-methylbutyrylcarnitine, and 3-methoxytramine sulfate were differentially altered by sodium intervention according to salt-sensitivity status. Furthermore, higher baseline values of 2-methylbutyrylcarnitine and isoleucine were associated with higher odds of salt-sensitivity. All five metabolites were positively associated with hypertension in the current study. Glutamate is an excitatory neurotransmitter and its metabolism is regulated by the glutathione cycle [
      • Sedlak T.W.
      • Paul B.D.
      • Parker G.M.
      • Hester L.D.
      • Snowman A.M.
      • Taniguchi Y.
      • et al.
      The glutathione cycle shapes synaptic glutamate activity.
      ]. Excess oxidative stress induced by glutamate is a plausible biological pathway underlying the observed association [
      • Coyle J.T.
      • Puttfarcken P.
      Oxidative stress, glutamate, and neurodegenerative disorders.
      ,
      • Parfenova H.
      • Basuroy S.
      • Bhattacharya S.
      • Tcheranova D.
      • Qu Y.
      • Regan R.F.
      • et al.
      Glutamate induces oxidative stress and apoptosis in cerebral vascular endothelial cells: contributions of HO-1 and HO-2 to cytoprotection.
      ]. Although 3-methoxytyramine sulfate is poorly characterized, its parent metabolite, 3-methoxytyramine, is a dopamine metabolite. Dopamine metabolism can cause dysregulated water and salt reabsorption, which is thought to be a key feature of salt-sensitive hypertension [
      • Harris R.C.
      • Zhang M.Z.
      Dopamine, the kidney, and hypertension.
      ,
      • Contreras F.
      • Fouillioux C.
      • Bolívar A.
      • Simonovis N.
      • Hernández-Hernández R.
      • Armas-Hernandez M.J.
      • et al.
      Dopamine, hypertension and obesity.
      ]. Isoleucine, 1-carboxyethylvaline, and 2-methylbutyrylcarnitine are all involved in branched chain amino acid (BCAA) metabolism [
      • Sass J.O.
      • Ensenauer R.
      • Röschinger W.
      • Reich H.
      • Steuerwald U.
      • Schirrmacher O.
      • et al.
      2-Methylbutyryl-coenzyme A dehydrogenase deficiency: functional and molecular studies on a defect in isoleucine catabolism.
      ,
      • Neinast M.
      • Murashige D.
      • Arany Z.
      Branched chain amino acids.
      ]. Although this is the first human study to implicate BCAA metabolites in human salt-sensitivity of BP, isoleucine were previously shown to be elevated in Dahl salt-sensitive rats [
      • Wang L.
      • Hou E.
      • Wang Z.
      • Sun N.
      • He L.
      • Chen L.
      • et al.
      Analysis of metabolites in plasma reveals distinct metabolic features between Dahl salt-sensitive rats and consomic SS.13BN rats.
      ], and 2-methylbutyrylcarnitine was previously shown to change significantly during sodium restriction in humans [
      • Derkach A.
      • Sampson J.
      • Joseph J.
      • Playdon M.C.
      • Stolzenberg-Solomon R.Z.
      Effects of dietary sodium on metabolites: the dietary approaches to Stop hypertension (DASH)–Sodium feeding study.
      ]. Furthermore, the 1-carboxyethylvaline metabolite was previously associated with decreased Shannon diversity index, a marker of gut microbial health that has previously been linked to numerous cardiometabolic conditions [
      • Wilmanski T.
      • Rappaport N.
      • Earls J.C.
      • Magis A.T.
      • Manor O.
      • Lovejoy J.
      • et al.
      Blood metabolome predicts gut microbiome α-diversity in humans.
      ]. These promising findings provide novel biological insights into high salt-sensitivity in humans and are among the first to identify associations of BCAA metabolites underlying this condition. In total, our findings provide compelling evidence that these metabolites could be part of the biological pathway linking dietary sodium intake to BP.
      Three correlated peptide metabolites (all r > 0.60), gamma-glutamylisoleucine, gamma-glutamylleucine, and gamma-glutamylphenylalanine, were also differentially altered in response to sodium intervention according to salt-sensitivity status. Each of these metabolites displayed a direct association with hypertension. In general, gamma glutamyl amino acids are mainly biosynthesized during glutathione degradation [
      • Bachhawat A.K.
      • Yadav S.
      The glutathione cycle: glutathione metabolism beyond the γ-glutamyl cycle.
      ] and regulated by gamma-glutamyl-transferase, an enzyme previously linked to hypertension and purported to exert its influence through oxidative stress and inflammation pathways [
      • Kunutsor S.K.
      • Apekey T.A.
      • Cheung B.M.Y.
      Gamma-glutamyltransferase and risk of hypertension: a systematic review and dose-responsemeta-analysis of prospective evidence.
      ]. Previously, Derkach and colleagues reported significant changes in gamma-glutamyl metabolites after sodium restriction [
      • Derkach A.
      • Sampson J.
      • Joseph J.
      • Playdon M.C.
      • Stolzenberg-Solomon R.Z.
      Effects of dietary sodium on metabolites: the dietary approaches to Stop hypertension (DASH)–Sodium feeding study.
      ]. We add to these results by further implicating gamma-glutamyl amino acid metabolites in salt-sensitive hypertension.
      The five remaining salt-sensitivity metabolites identified by the current study were all differentially altered by dietary sodium intervention according to salt-sensitivity status. Among them, three directly associated with hypertension, including glucose from the carbohydrate pathway, alpha-ketoglutarate from the energy pathway and hexanoylcarnitine from the lipid pathway, and two indirectly associated with hypertension, including threonate from the cofactors and vitamins pathway and 5-methyluridine from the nucleotide pathway. Our findings are in line with previous reports implicating elevated glucose as a risk factor for BP salt-sensitivity and hypertension [
      • Elijovich F.
      • Weinberger M.H.
      • Anderson C.A.M.
      • Appel L.J.
      • Bursztyn M.
      • Cook N.R.
      • et al.
      Salt sensitivity of blood pressure : a scientific statement from the American Heart Association.
      ,
      • Derkach A.
      • Sampson J.
      • Joseph J.
      • Playdon M.C.
      • Stolzenberg-Solomon R.Z.
      Effects of dietary sodium on metabolites: the dietary approaches to Stop hypertension (DASH)–Sodium feeding study.
      ,
      • Ferrannini E.
      • Cushman W.C.
      Diabetes and hypertension: the bad companions.
      ]. In contrast, alpha-ketoglutarate and hexanoylcarnitine have not been related to BP phenotypes previously. Alpha-ketoglutarate is a rate-determining metabolite in the tricarboxylic acid (TCA) and is a precursor for glutamate and glutamine [
      • Wu N.
      • Yang M.
      • Gaur U.
      • Xu H.
      • Yao Y.
      • Li D.
      Alpha-ketoglutarate: physiological functions and applications.
      ,
      • Zdzisińska B.
      • Żurek A.
      • Kandefer-Szerszeń M.
      Alpha-ketoglutarate as a molecule with pleiotropic activity: well-known and novel possibilities of therapeutic use.
      ]. Hexanoylcarnitine has been associated with adiposity phenotypes [
      • Muniandy M.
      • Velagapudi V.
      • Hakkarainen A.
      • Lundbom J.
      • Lundbom N.
      • Rissanen A.
      • et al.
      Plasma metabolites reveal distinct profiles associating with different metabolic risk factors in monozygotic twin pairs.
      ] and cardiovascular disease symptoms [
      • Vorkas P.A.
      • Shalhoub J.
      • Lewis M.R.
      • Spagou K.
      • Want E.J.
      • Nicholson J.K.
      • et al.
      Metabolic phenotypes of carotid atherosclerotic plaques relate to stroke risk: an exploratory study.
      ,
      • Kukharenko A.
      • Brito A.
      • Kozhevnikova M.V.
      • Moskaleva N.
      • Markin P.A.
      • Bochkareva N.
      • et al.
      Relationship between the plasma acylcarnitine profile and cardiometabolic risk factors in adults diagnosed with cardiovascular diseases.
      ]. Furthermore, its function as an oxidation intermediate may provide a likely biological explanation for the observed association with high salt-sensitivity and BP [
      • Vorkas P.A.
      • Shalhoub J.
      • Lewis M.R.
      • Spagou K.
      • Want E.J.
      • Nicholson J.K.
      • et al.
      Metabolic phenotypes of carotid atherosclerotic plaques relate to stroke risk: an exploratory study.
      ]. Among the two metabolites inversely associated with hypertension, threonate is an ascorbate metabolite synthesized from vitamin C degradation [
      • Wang H.
      • Jiang J.
      • Hu P.
      Determination of l-threonate in human plasma and urine by high performance liquid chromatography-tandem mass spectrometry.
      ], and higher levels have been related to increased intake of green leafy vegetables [
      • Playdon M.C.
      • Sampson J.N.
      • Cross A.J.
      • Sinha R.
      • Guertin K.A.
      • Moy K.A.
      • et al.
      Comparing metabolite profiles of habitual diet in serum and urine.
      ]. Very limited research on the role of 5-methyluridine, a common modification during the maturation of transfer RNA [
      • Chang Y.H.
      • Nishimura S.
      • Oishi H.
      • Kelly V.P.
      • Kuno A.
      • Takahashi S.
      TRMT2A is a novel cell cycle regulator that suppresses cell proliferation.
      ], in human disease has been conducted, and we are the first to report its association with a BP related trait. In aggregate, these findings support a role of glucose in salt-sensitive hypertension and propose novel biological pathways underlying this condition.
      The current study has several strengths. To our knowledge, this is the first untargeted metabolomics study of BP salt-sensitivity conducted to date, allowing us to comprehensively assess the relation between metabolites across multiple biological pathways and salt-sensitivity of BP. Furthermore, GenSalt uses a rigorous protocol for measuring sodium-sensitivity, including a stringent dietary sodium feeding and multiple measures of BP on multiple days of the baseline examination and each sodium intervention period. The interventional GenSalt design and repeated metabolite measures collected at baseline and the end of each intervention phase further provided a unique opportunity to directly examine changes in metabolites in response to sodium, providing novel mechanistic insights into the relation between dietary sodium intake and BP. In addition, by utilizing data from the BHS, we were able to further examine the relevance of BP salt-sensitivity metabolites in hypertension in a racially diverse sample, more consistent with the demographics of the US population. Certain limitations should also be mentioned. The small sample size of GenSalt metabolomics study subsample may have limited our power to uncover metabolites with relatively small effect sizes. Furthermore, despite robust sensitivity analyses in the BHS, we cannot rule out confounding due to estimated glomerular filtration rate and other variables that were not measured in the GenSalt study. In addition, metabolomics profiling in GenSalt was conducted using urinary biospecimens but only serum metabolomics data was available in the BHS. Although some studies have reported correlations of metabolite values across urine and serum [
      • Playdon M.C.
      • Sampson J.N.
      • Cross A.J.
      • Sinha R.
      • Guertin K.A.
      • Moy K.A.
      • et al.
      Comparing metabolite profiles of habitual diet in serum and urine.
      ,
      • Lau C.H.E.
      • Siskos A.P.
      • Maitre L.
      • Robinson O.
      • Athersuch T.J.
      • Want E.J.
      • et al.
      Determinants of the urinary and serum metabolome in children from six European populations.
      ], lack of association with hypertension could have resulted from differences in biospecimens. Moreover, if certain metabolite findings were ancestry specific, their relevance to hypertension could not be adequately assessed in the black and white BHS participants. These differences might have led to false negative results. On the other hand, the metabolites robustly identified across studies may be very generalizable to diverse populations. Finally, our findings differed from those of the previous targeted metabolomics study of salt-sensitivity conducted among DASH-sodium participants [
      • Cheng Y.
      • Song H.
      • Pan X.
      • Xue H.
      • Wan Y.
      • Wang T.
      • et al.
      Urinary metabolites associated with blood pressure on a low- or high-sodium diet.
      ]. Among the 5 metabolites identified in DASH-sodium (including β-aminoisobutyric acid, homocysteine, citrulline, lysine and cystine), only citrulline, lysine and cystine were detected and tested in GenSalt. These three metabolites did not attain nominal significance in our study. Most DASH-sodium participants (>95%) were white or Black, with typical American diets, while our study participants were all of Chinese Han ancestry and followed dietary patterns common to northern, rural China. Metabolites levels are known to be influenced by ancestries, diet and other environmental factors which could explained the inconsistent findings observed [
      • Patel M.J.
      • Batch B.C.
      • Svetkey L.P.
      • Bain J.R.
      • Turer C.B.
      • Haynes C.
      • et al.
      Race and sex differences in small-molecule metabolites and metabolic hormones in overweight and obese adults.
      ,
      • Menni C.
      • Zhai G.
      • MacGregor A.
      • Prehn C.
      • Römisch-Margl W.
      • Suhre K.
      • et al.
      Targeted metabolomics profiles are strongly correlated with nutritional patterns in women.
      ,
      • Nicholson G.
      • Rantalainen M.
      • Maher A.D.
      • Li J.V.
      • Malmodin D.
      • Ahmadi K.R.
      • et al.
      Human metabolic profiles are stably controlled by genetic and environmental variation.
      ]. However, we cannot rule out the possibility of false positive or false negative findings, given the small sample sizes in both studies.
      In total, the current study identified 15 salt-sensitivity related metabolites that may play a role in the development of hypertension. Together, these metabolites enhanced the prediction of hypertension over models that included traditional risk factors alone. These data suggest that metabolomics information could represent a valuable tool in efforts to molecularly characterize sodium-sensitivity of BP. Furthermore, our findings provide novel insights into the biological mechanisms underlying salt-sensitive hypertension, supporting a role for BCAAs, gamma glutamyl amino acids, and both fatty acid and pyrimidine metabolism underlying this complex condition.

      Funding

      The Genetic Epidemiology Network of Salt Sensitivity study (JH) was supported by the National Heart, Lung, and Blood Institute of the NIH under award numbers U01HL072507 , R01HL087263 , and R01HL090682 , and the National Institute of General Medical Sciences of the NIH under award number P20GM109036 . The Bogalusa Heart Study (LAB) was supported by the National Institute on Aging of the NIH under award number RF1AG041200 , and the National Institute of General Medical Sciences of the NIH under award number P20GM109036 .

      Declaration of competing interest

      The authors have no competing interests to declare that are relevant to the content of this article.

      Acknowledgments

      We are grateful for the contribution of all staff members who were involved in conducting the Genetic Epidemiology Network of Salt Sensitivity (GenSalt) Replication study and the Bogalusa Heart Study. We extend our gratitude to the participants of GenSalt Replication study and Bogalusa Heart Study.
      The Genetic Epidemiology Network of Salt Sensitivity study (JH) was supported by the National Heart, Lung, and Blood Institute of the NIH under award numbers U01HL072507, R01HL087263, and R01HL090682, and the National Institute of General Medical Sciences of the NIH under award number P20GM109036. The Bogalusa Heart Study (LAB) was supported by the National Institute on Aging of the NIH under award numbers RF1AG041200 R21AG051914, and the National Institute of General Medical Sciences of the NIH under award number P20GM109036.

      Appendix A. Supplementary data

      The following is the Supplementary data to this article:

      References

        • Zhou B.
        • Carrillo-Larco R.M.
        • Danaei G.
        • Riley L.M.
        • Paciorek C.J.
        • Stevens G.A.
        • et al.
        Worldwide trends in hypertension prevalence and progress in treatment and control from 1990 to 2019: a pooled analysis of 1201 population-representative studies with 104 million participants.
        Lancet (London, England). 2021; 398: 957https://doi.org/10.1016/S0140-6736(21)01330-1
        • Forouzanfar M.H.
        • Liu P.
        • Roth G.A.
        • Ng M.
        • Biryukov S.
        • Marczak L.
        • et al.
        Global burden of hypertension and systolic blood pressure of at least 110 to 115mmHg, 1990–2015.
        JAMA, J Am Med Assoc. 2017; 317: 165-182https://doi.org/10.1001/jama.2016.19043
        • Stanaway J.D.
        • Afshin A.
        • Gakidou E.
        • Lim S.S.
        • Abate D.
        • Abate K.H.
        • et al.
        Global, regional, and national comparative risk assessment of 84 behavioural, environmental and occupational, and metabolic risks or clusters of risks for 195 countries and territories, 1990-2017: a systematic analysis for the Global Burden of Disease Study 2017.
        Lancet. 2018; 392: 1923-1994https://doi.org/10.1016/S0140-6736(18)32225-6
        • He F.J.
        • Li J.
        • MacGregor G.A.
        Effect of longer term modest salt reduction on blood pressure: cochrane systematic review and meta-analysis of randomised trials.
        BMJ. 2013; 346https://doi.org/10.1136/bmj.f1325
        • Mozaffarian D.
        • Fahimi S.
        • Singh G.M.
        • Micha R.
        • Khatibzadeh S.
        • Engell R.E.
        • et al.
        Global sodium consumption and death from cardiovascular causes.
        N Engl J Med. 2014; 371: 624-634https://doi.org/10.1056/NEJMoa1304127
        • Afshin A.
        • Sur P.J.
        • Fay K.A.
        • Cornaby L.
        • Ferrara G.
        • Salama J.S.
        • et al.
        Health effects of dietary risks in 195 countries, 1990–2017: a systematic analysis for the Global Burden of Disease Study 2017.
        Lancet. 2019; 393: 1958-1972https://doi.org/10.1016/S0140-6736(19)30041-8
        • Chen L.
        • He F.J.
        • Dong Y.
        • Huang Y.
        • Harshfield G.A.
        • Zhu H.
        Sodium reduction, metabolomic profiling, and cardiovascular disease risk in untreated black hypertensives: a randomized, double-blind, placebo-controlled trial.
        Hypertension. 2019; 74: 194-200https://doi.org/10.1161/HYPERTENSIONAHA.119.12880
        • Liu Y.
        • Shi M.
        • Dolan J.
        • He J.
        Sodium sensitivity of blood pressure in Chinese populations.
        J Hum Hypertens. 2019; https://doi.org/10.1038/s41371-018-0152-0
        • He J.
        • Kelly T.N.
        • Zhao Q.
        • Li H.
        • Huang J.
        • Wang L.
        • et al.
        Genome-wide association study identifies 8 novel loci associated with blood pressure responses to interventions in Han Chinese.
        Circ Cardiovasc Genet. 2013; 6: 598-607https://doi.org/10.1161/CIRCGENETICS.113.000307
        • Barba G.
        • Galletti F.
        • Cappuccio F.P.
        • Siani A.
        • Venezia A.
        • Versiero M.
        • et al.
        Incidence of hypertension in individuals with different blood pressure salt-sensitivity: results of a 15-year follow-up study.
        J Hypertens. 2007; 25: 1465-1471https://doi.org/10.1097/HJH.0b013e3281139ebd
        • Weinberger M.H.
        • Fineberg N.S.
        • Fineberg S.E.
        • Weinberger M.
        Salt sensitivity, pulse pressure, and death in normal and hypertensive humans.
        Hypertension. 2001; 37 (Lippincott Williams and Wilkins): 429-432https://doi.org/10.1161/01.hyp.37.2.429
        • Morimoto A.
        • Uzu T.
        • Fujii T.
        • Nishimura M.
        • Kuroda S.
        • Nakamura S.
        • et al.
        Sodium sensitivity and cardiovascular events in patients with essential hypertension.
        Lancet. 1997; 350: 1734-1737https://doi.org/10.1016/S0140-6736(97)05189-1
        • Elijovich F.
        • Weinberger M.H.
        • Anderson C.A.M.
        • Appel L.J.
        • Bursztyn M.
        • Cook N.R.
        • et al.
        Salt sensitivity of blood pressure : a scientific statement from the American Heart Association.
        Hypertension. 2016; 68: e7-e46https://doi.org/10.1161/HYP.0000000000000047
        • Ussher J.R.
        • Elmariah S.
        • Gerszten R.E.
        • Dyck J.R.B.
        The emerging role of metabolomics in the diagnosis and prognosis of cardiovascular disease.
        J Am Coll Cardiol. 2016; 68: 2850-2870https://doi.org/10.1016/j.jacc.2016.09.972
        • McGarrah R.W.
        • Crown S.B.
        • Zhang G.F.
        • Shah S.H.
        • Newgard C.B.
        Cardiovascular metabolomics.
        Circ Res. 2018; 122: 1238-1258https://doi.org/10.1161/CIRCRESAHA.117.311002
        • Arnett D.K.
        • Claas S.A.
        Omics of blood pressure and hypertension.
        Circ Res. 2018; 122: 1409-1419https://doi.org/10.1161/CIRCRESAHA.118.311342
        • Menni C.
        • Graham D.
        • Kastenmüller G.
        • Alharbi N.H.J.
        • Alsanosi S.M.
        • Mcbride M.
        • et al.
        Metabolomic identification of a novel pathway of blood pressure regulation involving hexadecanedioate.
        Hypertension. 2015; 66: 422-429https://doi.org/10.1161/HYPERTENSIONAHA.115.05544
        • Dietrich S.
        • Floegel A.
        • Weikert C.
        • Pischon T.
        • Boeing H.
        • Drogan D.
        Identification of serum metabolites associated with incident hypertension in the European prospective investigation into cancer and nutrition-potsdam study.
        Hypertension. 2016; 68: 471-477https://doi.org/10.1161/HYPERTENSIONAHA.116.07292
        • He W.J.
        • Li C.
        • Mi X.
        • Shi M.
        • Gu X.
        • Bazzano L.A.
        • et al.
        An untargeted metabolomics study of blood pressure: findings from the Bogalusa Heart Study.
        J Hypertens. 2020; 38: 1302-1311https://doi.org/10.1097/HJH.0000000000002363
        • Cheng Y.
        • Song H.
        • Pan X.
        • Xue H.
        • Wan Y.
        • Wang T.
        • et al.
        Urinary metabolites associated with blood pressure on a low- or high-sodium diet.
        Theranostics. 2018; 8: 1468-1480https://doi.org/10.7150/thno.22018
        • Whelton P.K.
        • Carey R.M.
        • Aronow W.S.
        • Casey D.E.
        • Collins K.J.
        • Himmelfarb C.D.
        • et al.
        2017 ACC/AHA/AAPA/ABC/ACPM/AGS/APhA/ASH/ASPC/NMA/PCNA guideline for the prevention, detection, evaluation, and management of high blood pressure in adults a report of the American College of Cardiology/American Heart Association Task Force on Clinical practice guidelines.
        Hypertension. 2018; 71: E13-E115https://doi.org/10.1161/HYP.0000000000000065
      1. GenSalt: rationale, design, methods and baseline characteristics of study participants.
        J Hum Hypertens. 2007; 21: 639-646https://doi.org/10.1038/sj.jhh.1002207
        • Gu D.
        • Zhao Q.
        • Chen J.
        • Chen J.C.
        • Huang J.
        • Bazzano L.A.
        • et al.
        Reproducibility of blood pressure responses to dietary sodium and potassium interventions: the GenSalt study.
        Hypertension. 2013; 62: 499-505https://doi.org/10.1161/HYPERTENSIONAHA.113.01034
        • Jones B.L.
        • Nagin D.S.
        • Roeder K.
        A SAS procedure based on mixture models for estimating developmental trajectories.
        Socio Methods Res. 2001; 29: 374-393https://doi.org/10.1177/0049124101029003005
        • He J.
        • Huang J.-F.
        • Li C.
        • Chen J.
        • Lu X.
        • Chen J.-C.
        • et al.
        Sodium sensitivity, sodium resistance, and incidence of hypertension.
        Hypertension. 2021; : 155-164https://doi.org/10.1161/hypertensionaha.120.16758
        • Khamis M.M.
        • Holt T.
        • Awad H.
        • El-Aneed A.
        • Adamko D.J.
        Comparative analysis of creatinine and osmolality as urine normalization strategies in targeted metabolomics for the differential diagnosis of asthma and COPD.
        Metabolomics. 2018; 14https://doi.org/10.1007/s11306-018-1418-9
        • Kennedy A.D.
        • Miller M.J.
        • Beebe K.
        • Wulff J.E.
        • Evans A.M.
        • Miller L.A.D.
        • et al.
        Metabolomic profiling of human urine as a screen for multiple inborn errors of metabolism.
        Genet Test Mol Biomarkers. 2016; 20: 485-495https://doi.org/10.1089/gtmb.2015.0291
        • Nam S.L.
        • Paulina de la Mata A.
        • Dias R.P.
        • Harynuk J.J.
        Towards standardization of data normalization strategies to improve urinary metabolomics studies by gc×gc-tofms.
        Metabolites. 2020; 10: 1-13https://doi.org/10.3390/metabo10090376
        • Foster T.A.
        • Berenson G.S.
        Measurement error and reliability in four pediatric cross-sectional surveys of cardiovascular disease risk factor variables-The Bogalusa Heart Study.
        J Chron Dis. 1987; 40: 13-21https://doi.org/10.1016/0021-9681(87)90092-0
        • Pang Z.
        • Chong J.
        • Zhou G.
        • de Lima Morais D.A.
        • Chang L.
        • Barrette M.
        • et al.
        MetaboAnalyst 5.0: narrowing the gap between raw spectra and functional insights.
        Nucleic Acids Res. 2021; 49: 388-396https://doi.org/10.1093/nar/gkab382
        • Nierenberg J.L.
        • Li C.
        • He J.
        • Gu D.
        • Chen J.
        • Lu X.
        • et al.
        Blood pressure genetic risk score predicts blood pressure responses to dietary sodium and potassium the GenSalt Study (Genetic epidemiology network of salt sensitivity).
        Hypertension. 2017; 70: 1106-1112https://doi.org/10.1161/HYPERTENSIONAHA.117.10108
        • Natarajan P.
        • Peloso G.M.
        • Zekavat S.M.
        • Montasser M.
        • Ganna A.
        • Chaffin M.
        • et al.
        Deep-coverage whole genome sequences and blood lipids among 16,324 individuals.
        Nat Commun. 2018; 9: 1-12https://doi.org/10.1038/s41467-018-05747-8
        • Yu B.
        • Li A.H.
        • Muzny D.
        • Veeraraghavan N.
        • De Vries P.S.
        • Bis J.C.
        • et al.
        Association of rare loss-of-function alleles in HAL, serum histidine: levels and incident coronary heart disease.
        Circ Cardiovasc Genet. 2015; 8: 351-355https://doi.org/10.1161/CIRCGENETICS.114.000697
        • Mels C.M.
        • Delles C.
        • Louw R.
        • Schutte A.E.
        Central systolic pressure and a nonessential amino acidmetabolomics profile: TheAfrican Prospective study on the Early Detection and Identification of Cardiovascular disease and Hypertension.
        J Hypertens. 2019; 37: 1157-1166https://doi.org/10.1097/HJH.0000000000002040
        • Sedlak T.W.
        • Paul B.D.
        • Parker G.M.
        • Hester L.D.
        • Snowman A.M.
        • Taniguchi Y.
        • et al.
        The glutathione cycle shapes synaptic glutamate activity.
        Proc Natl Acad Sci U S A. 2019; 116: 2701-2706https://doi.org/10.1073/pnas.1817885116
        • Coyle J.T.
        • Puttfarcken P.
        Oxidative stress, glutamate, and neurodegenerative disorders.
        Science. 1993; 262 (80-): 689-695https://doi.org/10.1126/science.7901908
        • Parfenova H.
        • Basuroy S.
        • Bhattacharya S.
        • Tcheranova D.
        • Qu Y.
        • Regan R.F.
        • et al.
        Glutamate induces oxidative stress and apoptosis in cerebral vascular endothelial cells: contributions of HO-1 and HO-2 to cytoprotection.
        Am J Physiol Cell Physiol. 2006; 290https://doi.org/10.1152/ajpcell.00386.2005
        • Harris R.C.
        • Zhang M.Z.
        Dopamine, the kidney, and hypertension.
        Curr Hypertens Rep. 2012; 14: 138-143https://doi.org/10.1007/s11906-012-0253-z
        • Contreras F.
        • Fouillioux C.
        • Bolívar A.
        • Simonovis N.
        • Hernández-Hernández R.
        • Armas-Hernandez M.J.
        • et al.
        Dopamine, hypertension and obesity.
        J Hum Hypertens. 2002; 16: S13-S17https://doi.org/10.1038/sj.jhh.1001334
        • Sass J.O.
        • Ensenauer R.
        • Röschinger W.
        • Reich H.
        • Steuerwald U.
        • Schirrmacher O.
        • et al.
        2-Methylbutyryl-coenzyme A dehydrogenase deficiency: functional and molecular studies on a defect in isoleucine catabolism.
        Mol Genet Metabol. 2008; 93: 30-35https://doi.org/10.1016/j.ymgme.2007.09.002
        • Neinast M.
        • Murashige D.
        • Arany Z.
        Branched chain amino acids.
        Annu Rev Physiol. 2019; 81: 139-164https://doi.org/10.1146/annurev-physiol-020518-114455
        • Wang L.
        • Hou E.
        • Wang Z.
        • Sun N.
        • He L.
        • Chen L.
        • et al.
        Analysis of metabolites in plasma reveals distinct metabolic features between Dahl salt-sensitive rats and consomic SS.13BN rats.
        Biochem Biophys Res Commun. 2014; 450: 863-869https://doi.org/10.1016/j.bbrc.2014.06.089
        • Derkach A.
        • Sampson J.
        • Joseph J.
        • Playdon M.C.
        • Stolzenberg-Solomon R.Z.
        Effects of dietary sodium on metabolites: the dietary approaches to Stop hypertension (DASH)–Sodium feeding study.
        Am J Clin Nutr. 2017; 106: 1131-1141https://doi.org/10.3945/ajcn.116.150136
        • Wilmanski T.
        • Rappaport N.
        • Earls J.C.
        • Magis A.T.
        • Manor O.
        • Lovejoy J.
        • et al.
        Blood metabolome predicts gut microbiome α-diversity in humans.
        Nat Biotechnol. 2019; 37: 1217-1228https://doi.org/10.1038/s41587-019-0233-9
        • Bachhawat A.K.
        • Yadav S.
        The glutathione cycle: glutathione metabolism beyond the γ-glutamyl cycle.
        IUBMB Life. 2018; 70: 585-592https://doi.org/10.1002/iub.1756
        • Kunutsor S.K.
        • Apekey T.A.
        • Cheung B.M.Y.
        Gamma-glutamyltransferase and risk of hypertension: a systematic review and dose-responsemeta-analysis of prospective evidence.
        J Hypertens. 2015; 33: 2373-2381https://doi.org/10.1097/HJH.0000000000000763
        • Ferrannini E.
        • Cushman W.C.
        Diabetes and hypertension: the bad companions.
        Lancet. 2012; 380: 601-610https://doi.org/10.1016/S0140-6736(12)60987-8
        • Wu N.
        • Yang M.
        • Gaur U.
        • Xu H.
        • Yao Y.
        • Li D.
        Alpha-ketoglutarate: physiological functions and applications.
        Biomol Ther. 2016; 24: 1-8https://doi.org/10.4062/biomolther.2015.078
        • Zdzisińska B.
        • Żurek A.
        • Kandefer-Szerszeń M.
        Alpha-ketoglutarate as a molecule with pleiotropic activity: well-known and novel possibilities of therapeutic use.
        Arch Immunol Ther Exp (Warsz). 2017; 65: 21-36https://doi.org/10.1007/s00005-016-0406-x
        • Muniandy M.
        • Velagapudi V.
        • Hakkarainen A.
        • Lundbom J.
        • Lundbom N.
        • Rissanen A.
        • et al.
        Plasma metabolites reveal distinct profiles associating with different metabolic risk factors in monozygotic twin pairs.
        Int J Obes. 2019; 43: 487-502https://doi.org/10.1038/s41366-018-0132-z
        • Vorkas P.A.
        • Shalhoub J.
        • Lewis M.R.
        • Spagou K.
        • Want E.J.
        • Nicholson J.K.
        • et al.
        Metabolic phenotypes of carotid atherosclerotic plaques relate to stroke risk: an exploratory study.
        Eur J Vasc Endovasc Surg. 2016; 52: 5-10https://doi.org/10.1016/j.ejvs.2016.01.022
        • Kukharenko A.
        • Brito A.
        • Kozhevnikova M.V.
        • Moskaleva N.
        • Markin P.A.
        • Bochkareva N.
        • et al.
        Relationship between the plasma acylcarnitine profile and cardiometabolic risk factors in adults diagnosed with cardiovascular diseases.
        Clin Chim Acta. 2020; 507: 250-256https://doi.org/10.1016/j.cca.2020.04.035
        • Wang H.
        • Jiang J.
        • Hu P.
        Determination of l-threonate in human plasma and urine by high performance liquid chromatography-tandem mass spectrometry.
        J Chromatogr B Anal Technol Biomed Life Sci. 2006; 834: 155-162https://doi.org/10.1016/j.jchromb.2006.02.057
        • Playdon M.C.
        • Sampson J.N.
        • Cross A.J.
        • Sinha R.
        • Guertin K.A.
        • Moy K.A.
        • et al.
        Comparing metabolite profiles of habitual diet in serum and urine.
        Am J Clin Nutr. 2016; 104: 776-789https://doi.org/10.3945/ajcn.116.135301
        • Chang Y.H.
        • Nishimura S.
        • Oishi H.
        • Kelly V.P.
        • Kuno A.
        • Takahashi S.
        TRMT2A is a novel cell cycle regulator that suppresses cell proliferation.
        Biochem Biophys Res Commun. 2019; 508: 410-415https://doi.org/10.1016/j.bbrc.2018.11.104
        • Lau C.H.E.
        • Siskos A.P.
        • Maitre L.
        • Robinson O.
        • Athersuch T.J.
        • Want E.J.
        • et al.
        Determinants of the urinary and serum metabolome in children from six European populations.
        BMC Med. 2018; 16https://doi.org/10.1186/s12916-018-1190-8
        • Patel M.J.
        • Batch B.C.
        • Svetkey L.P.
        • Bain J.R.
        • Turer C.B.
        • Haynes C.
        • et al.
        Race and sex differences in small-molecule metabolites and metabolic hormones in overweight and obese adults.
        OMICS. 2013; 17: 627https://doi.org/10.1089/OMI.2013.0031
        • Menni C.
        • Zhai G.
        • MacGregor A.
        • Prehn C.
        • Römisch-Margl W.
        • Suhre K.
        • et al.
        Targeted metabolomics profiles are strongly correlated with nutritional patterns in women.
        Metabolomics. 2013; 9: 506-514https://doi.org/10.1007/S11306-012-0469-6/TABLES/3
        • Nicholson G.
        • Rantalainen M.
        • Maher A.D.
        • Li J.V.
        • Malmodin D.
        • Ahmadi K.R.
        • et al.
        Human metabolic profiles are stably controlled by genetic and environmental variation.
        Mol Syst Biol. 2011; 7: 525https://doi.org/10.1038/MSB.2011.57