Advertisement

Metabolic profiling of charged metabolites in association with menopausal status in Japanese community-dwelling midlife women: Tsuruoka Metabolomic Cohort Study

  • Keiko Watanabe
    Affiliations
    Department of Obstetrics and Gynecology, Keio University School of Medicine, 35 Shinanomachi, Shinjuku-ku, Tokyo 160-8582, Japan

    Department of Preventive Medicine and Public Health, Keio University School of Medicine, 35 Shinanomachi, Shinjuku-ku, Tokyo 160-8582, Japan
    Search for articles by this author
  • Miho Iida
    Affiliations
    Department of Obstetrics and Gynecology, Keio University School of Medicine, 35 Shinanomachi, Shinjuku-ku, Tokyo 160-8582, Japan

    Department of Preventive Medicine and Public Health, Keio University School of Medicine, 35 Shinanomachi, Shinjuku-ku, Tokyo 160-8582, Japan

    Institute for Advanced Biosciences, Keio University, Baba-cho 14-1, Tsuruoka City, Yamagata 997-0035, Japan
    Search for articles by this author
  • Sei Harada
    Affiliations
    Department of Preventive Medicine and Public Health, Keio University School of Medicine, 35 Shinanomachi, Shinjuku-ku, Tokyo 160-8582, Japan

    Institute for Advanced Biosciences, Keio University, Baba-cho 14-1, Tsuruoka City, Yamagata 997-0035, Japan
    Search for articles by this author
  • Suzuka Kato
    Affiliations
    Department of Preventive Medicine and Public Health, Keio University School of Medicine, 35 Shinanomachi, Shinjuku-ku, Tokyo 160-8582, Japan
    Search for articles by this author
  • Kazuyo Kuwabara
    Affiliations
    Department of Preventive Medicine and Public Health, Keio University School of Medicine, 35 Shinanomachi, Shinjuku-ku, Tokyo 160-8582, Japan
    Search for articles by this author
  • Ayako Kurihara
    Affiliations
    Department of Preventive Medicine and Public Health, Keio University School of Medicine, 35 Shinanomachi, Shinjuku-ku, Tokyo 160-8582, Japan
    Search for articles by this author
  • Ayano Takeuchi
    Affiliations
    Department of Preventive Medicine and Public Health, Keio University School of Medicine, 35 Shinanomachi, Shinjuku-ku, Tokyo 160-8582, Japan
    Search for articles by this author
  • Daisuke Sugiyama
    Affiliations
    Department of Preventive Medicine and Public Health, Keio University School of Medicine, 35 Shinanomachi, Shinjuku-ku, Tokyo 160-8582, Japan

    Faculty of Nursing and Medical Care Graduate School of Health Management, Keio University, 4411 Endo, Fujisawa-shi, Kanagawa 252-0883 Japan
    Search for articles by this author
  • Tomonori Okamura
    Affiliations
    Department of Preventive Medicine and Public Health, Keio University School of Medicine, 35 Shinanomachi, Shinjuku-ku, Tokyo 160-8582, Japan
    Search for articles by this author
  • Asako Suzuki
    Affiliations
    Institute for Advanced Biosciences, Keio University, Baba-cho 14-1, Tsuruoka City, Yamagata 997-0035, Japan
    Search for articles by this author
  • Kaori Amano
    Affiliations
    Institute for Advanced Biosciences, Keio University, Baba-cho 14-1, Tsuruoka City, Yamagata 997-0035, Japan
    Search for articles by this author
  • Akiyoshi Hirayama
    Affiliations
    Institute for Advanced Biosciences, Keio University, Baba-cho 14-1, Tsuruoka City, Yamagata 997-0035, Japan
    Search for articles by this author
  • Masahiro Sugimoto
    Affiliations
    Institute for Advanced Biosciences, Keio University, Baba-cho 14-1, Tsuruoka City, Yamagata 997-0035, Japan
    Search for articles by this author
  • Tomoyoshi Soga
    Affiliations
    Institute for Advanced Biosciences, Keio University, Baba-cho 14-1, Tsuruoka City, Yamagata 997-0035, Japan

    Faculty of Environment and Information Studies, Keio University, 5322 Endo, Fujisawa-shi, Kanagawa 252-0882 Japan
    Search for articles by this author
  • Masaru Tomita
    Affiliations
    Institute for Advanced Biosciences, Keio University, Baba-cho 14-1, Tsuruoka City, Yamagata 997-0035, Japan

    Faculty of Environment and Information Studies, Keio University, 5322 Endo, Fujisawa-shi, Kanagawa 252-0882 Japan
    Search for articles by this author
  • Yusuke Kobayashi
    Affiliations
    Department of Obstetrics and Gynecology, Keio University School of Medicine, 35 Shinanomachi, Shinjuku-ku, Tokyo 160-8582, Japan
    Search for articles by this author
  • Kouji Banno
    Affiliations
    Department of Obstetrics and Gynecology, Keio University School of Medicine, 35 Shinanomachi, Shinjuku-ku, Tokyo 160-8582, Japan
    Search for articles by this author
  • Daisuke Aoki
    Affiliations
    Department of Obstetrics and Gynecology, Keio University School of Medicine, 35 Shinanomachi, Shinjuku-ku, Tokyo 160-8582, Japan
    Search for articles by this author
  • Toru Takebayashi
    Correspondence
    Corresponding author at: Department of Preventive Medicine and Public Health, Keio University School of Medicine, 35 Shinanomachi, Shinjuku-ku, Tokyo 160-8582, Japan.
    Affiliations
    Department of Preventive Medicine and Public Health, Keio University School of Medicine, 35 Shinanomachi, Shinjuku-ku, Tokyo 160-8582, Japan

    Institute for Advanced Biosciences, Keio University, Baba-cho 14-1, Tsuruoka City, Yamagata 997-0035, Japan
    Search for articles by this author
Open AccessPublished:October 09, 2021DOI:https://doi.org/10.1016/j.maturitas.2021.10.004

      Highlights

      • Evidence suggest that amino acids play key roles in obesity-related diseases.
      • Few studies have focused on amino acids and related metabolites in the specific context of menopause.
      • The association between menopausal status and charged metabolites was explored in over 1000 midlife women.
      • Multiple charged metabolites were associated with menopausal status.
      • Amongst the differentially expressed metabolites were those relevant to atherosclerosis.

      Abstract

      Background

      Emerging evidence has shown that charged metabolites, such as amino acids, may play an important role in the pathogenesis of various metabolic disorders, many of which women in the postmenopausal period are at high risk of developing. This study examined the metabolic profile of middle-aged Japanese women to investigate alterations in charged metabolites induced by menopausal transition.

      Methods

      The participants were 1193 female residents aged 40–60 at the baseline survey of the Tsuruoka Metabolomics Cohort Study. We investigated the cross-sectional association of menopausal status with 94 metabolomic biomarkers assayed in fasting plasma samples via capillary electrophoresis time-of-flight mass spectrometry using linear regression analysis.

      Results

      Among the participants, 529 were premenopausal, 132 were in menopausal transition (MT), and 532 were postmenopausal. Significant differences were found in age, blood pressure, glucose and lipid levels, and smoking and drinking habits among the three groups. The concentrations of 5 metabolites in the MT group and 15 metabolites in the postmenopausal group were significantly higher than those in the premenopausal group after adjusting for confounding factors. When classified into pathways, these metabolites were related to the tricarboxylic cycle, urea cycle, and homocysteine metabolism, some of which are linked to arteriosclerosis.

      Conclusion

      Multiple charged metabolites were associated with women's menopausal status, showing a gradual increase as women shifted from pre-, to peri-, to postmenopause. These findings might reflect the early changes behind the increased risk of dyslipidemia, diabetes, cardiovascular disease, and osteoporosis in later life.

      Keywords

      1. Introduction

      As life expectancy increases annually around the world, responding to disorders associated with aging has become an important public health challenge. Menopause, an aging process specific to women, is a major health milestone that influences women's future health.  As various health risks increase rapidly after the cessation of menstruation, elucidating the changes in biological systems caused by menopause is important for delivering effective prevention and treatment options.
      Although decline in estrogen levels is the major causative factor, lifestyle and genetic factors should also be considered when analyzing the impact of menopause on increased health risks. Recently, metabolomics has been widely adopted in epidemiologic research to identify early traits of diseases that reflect both genetic variations and the impact of the environment in community-based settings [
      • Yu B
      • Zanetti KA
      • Temprosa M
      • Albanes D
      • Appel N
      • Barrera CB
      • Ben-Shlomo Y
      • Boerwinkle E
      • Casas JP
      • Clish C
      • Dale C
      • Dehghan A
      • Derkach A
      • Eliassen AH
      • Elliott P
      • Fahy E
      • Gieger C
      • Gunter MJ
      • Harada S
      • Harris T
      • Herr DR
      • Herrington D
      • Hirschhorn JN
      • Hoover E
      • Hsing AW
      • Johansson M
      • Kelly RS
      • Khoo CM
      • Kivimäki M
      • Kristal BS
      • Langenberg C
      • Lasky-Su J
      • Lawlor DA
      • Lotta LA
      • Mangino M
      • Le Marchand L
      • Mathé E
      • Matthews CE
      • Menni C
      • Mucci LA
      • Murphy R
      • Oresic M
      • Orwoll E
      • Ose J
      • Pereira AC
      • Playdon MC
      • Poston L
      • Price J
      • Qi Q
      • Rexrode K
      • Risch A
      • Sampson J
      • Seow WJ
      • Sesso HD
      • Shah SH
      • Shu XO
      • Smith GCS
      • Sovio U
      • Stevens VL
      • Stolzenberg-Solomon R
      • Takebayashi T
      • Tillin T
      • Travis R
      • Tzoulaki I
      • Ulrich CM
      • Vasan RS
      • Verma M
      • Wang Y
      • Wareham NJ
      • Wong A
      • Younes N
      • Zhao H
      • Zheng W
      • Moore SC
      The Consortium of Metabolomics Studies (COMETS): metabolomics in 47 Prospective Cohort Studies.
      ]. In studies which investigated the relationship between metabolites and menopausal status [
      • Darst BF
      • Koscik RL
      • Hogan KJ
      • Johnson SC
      • Engelman CD
      Longitudinal plasma metabolomics of aging and sex.
      ,
      • Ke C
      • Hou Y
      • Zhang H
      • Yang K
      • Wang J
      • Guo B
      • Zhang F
      • Li H
      • Zhou X
      • Li Y
      • Li K
      Plasma metabolic profiles in women are menopause dependent.
      ], alterations of lipid metabolites were implied as the underlying mechanism in increased cardiometabolic risk in the postmenopausal period. Although the link between menopause and lipid metabolism are well known, few have explored the importance of charged metabolites and their metabolism during the menopausal transition (MT). Amino acid metabolism is known to be associated with health problems such as obesity, metabolic syndrome, and cardiovascular disease (CVD) in relation to age and gender [
      • Auro K
      • Joensuu A
      • Fischer K
      • Kettunen J
      • Salo P
      • Mattsson H
      • Niironen M
      • Kaprio J
      • Eriksson JG
      • Lehtimäki T
      • Raitakari O
      • Jula A
      • Tiitinen A
      • Jauhiainen M
      • Soininen P
      • Kangas AJ
      • Kähönen M
      • Havulinna AS
      • Ala-Korpela M
      • Salomaa V
      • Metspalu A
      • Perola M
      A metabolic view on menopause and ageing.
      ,
      • Rist MJ
      • Roth A
      • Frommherz L
      • Weinert CH
      • Krüger R
      • Merz B
      • Bunzel D
      • Mack C
      • Egert B
      • Bub A
      • Görling B
      • Tzvetkova P
      • Luy B
      • Hoffmann I
      • Kulling SE
      • Watzl B
      Metabolite patterns predicting sex and age in participants of the Karlsruhe Metabolomics and Nutrition (KarMeN) study.
      ], and disorders of amino acid metabolism have also been reported to precede the development of diabetes [
      • Wang TJ
      • Larson MG
      • Vasan RS
      • Cheng S
      • Rhee EP
      • McCabe E
      • Lewis GD
      • Fox CS
      • Jacques PF
      • Fernandez C
      • O’Donnell CJ
      • Carr SA
      • Mootha VK
      • Florez JC
      • Souza A
      • Melander O
      • Clish CB
      • Gerszten RE
      Metabolite profiles and the risk of developing diabetes.
      ]. Diabetes is noted to have a greater impact on CVD risk in women than men [
      • Dal Canto E
      • Ceriello A
      • Rydén L
      • Ferrini M
      • Hansen TB
      • Schnell O
      • Standl E
      • Beulens JW
      Diabetes as a cardiovascular risk factor: an overview of global trends of macro and micro vascular complications.
      ], but whether menopause has an effect on diabetes risk remains controversial [
      • Soriguer F
      • Morcillo S
      • Hernando V
      • Valdés S
      • Ruiz de Adana MS
      • Olveira G
      • Fuentes EG
      • González I
      • Tapia MJ
      • Esteva I
      • Rojo-Martínez G
      Type 2 diabetes mellitus and other cardiovascular risk factors are no more common during menopause: longitudinal study.
      ,
      • Brand JS
      • van der Schouw YT
      • Onland-Moret NC
      • et al.
      Age at menopause, reproductive life span, and type 2 diabetes risk: results from the EPIC-InterAct study.
      ]. Understanding the changes in charged metabolites during MT may contribute to the understanding of the effects of menopause on glucose metabolism and related health disturbances.
      Several platforms are used to analyze a wide range of metabolites, including nuclear magnetic resonance (NMR) spectroscopy, gas chromatography–mass spectrometry (MS), liquid chromatography-MS, and capillary electrophoresis (CE)-MS, with each having its benefits and drawbacks. Among these, CE-MS is known for its ability to accurately quantify charged metabolites at low concentrations. This study aimed to investigate the association with menopause on metabolite levels among middle-aged women, placing an emphasis on charged metabolites that could play important roles in various diseases induced by menopause.

      2. Methods

      2.1 Study population

      Data from the Tsuruoka Metabolomics Cohort Study (TMCS) were used. TMCS is an ongoing prospective cohort study in Japan enrolling 11,002 participants aged 35–74 living or working in Tsuruoka City, designed to discover metabolomic biomarkers for common diseases and disorders related to genetic and environmental factors [
      • Harada S
      • Takebayashi T
      • Kurihara A
      • Akiyama M
      • Suzuki A
      • Hatakeyama Y
      • Sugiyama D
      • Kuwabara K
      • Takeuchi A
      • Okamura T
      • Nishiwaki Y
      • Tanaka T
      • Hirayama A
      • Sugimoto M
      • Soga T
      • Tomita M
      Metabolomic profiling reveals novel biomarkers of alcohol intake and alcohol-induced liver injury in community-dwelling men.
      ,
      • Iida M
      • Harada S
      • Kurihara A
      • Fukai K
      • Kuwabara K
      • Sugiyama D
      • Takeuchi A
      • Okamura T
      • Akiyama M
      • Nishiwaki Y
      • Suzuki A
      • Hirayama A
      • Sugimoto M
      • Soga T
      • Tomita M
      • Banno K
      • Aoki D
      • Takebayashi T
      Profiling of plasma metabolites in postmenopausal women with metabolic syndrome.
      ]. The baseline survey was conducted between April 2012 and March 2015, and the participants were recruited among attendees of annual municipal or workplace health check-up programs. Among the 5827 women (59.3±10.2 years) enrolled, 1193 women aged 40 to 60 at baseline were selected to explore the cross-sectional association of menopausal status with metabolomic biomarkers. We excluded subjects with histories of CVD, cancer, iatrogenic menopause, hormone replacement therapy within a year, and those on any types of regular medication, as these factors can greatly affect the metabolome (Fig. 1).
      Fig 1:
      Fig. 1Flow diagram of included and excluded participants.
      The study was approved by the Medical Ethics Committee of the Keio University School of Medicine, Tokyo, Japan (approval no. 20110264), and all participants provided written informed consent.

      2.2 Assessment of menopausal status

      All participants completed a self-administered questionnaire. Menstrual history questionnaires included current status (continuing/stopping/stopped), the first day of the last menstrual period (LMP) and menstrual cycle if menstruation had not stopped, and age at menopause (AAM) and causes of menopause (natural/medical/other) if postmenopausal. Women were classified into three groups, premenopausal, MT, and postmenopausal, according to their current menstrual status. The LMP, menstrual cycle, and AAM were also confirmed so that the postmenopausal group included subjects who had ceased menstruation for at least 1 year, and the MT with an amenorrhea period of at least 60 days but less than 1 year, an irregular menstrual cycle as characterized in the 2001 Stages of Reproductive Aging Workshop staging system as Stage −1, or a late MT phase. Gynecological medical history was also collected, including the use of hormonal therapy. Blood estradiol (E2) and follicle-stimulating hormone (FSH) levels were measured to assess the validity of the self-administered questionnaire.

      2.3 Data and sample collection

      Detailed information on lifestyle parameters were collected through questionnaires, such as smoking, drinking, and physical activity. Data regarding anthropometry, clinical biochemistry, and blood specimens for metabolomic profiling were also obtained. Height and weight were measured in light clothing, and body mass index (BMI) was calculated. Blood pressure (BP) was measured twice on one occasion in the sitting position using an automated sphygmomanometer, and the mean of the two measurements was used for analysis. Hypertension was defined as systolic BP (SBP) ≥ 140 mmHg or diastolic BP ≥ 90 mmHg.
      All blood samples were collected after 12 h of overnight fasting to avoid dietary and circadian rhythm variations. Details regarding the blood sample collection protocols are available elsewhere [
      • Harada S
      • Hirayama A
      • Chan Q
      • Kurihara A
      • Fukai K
      • Iida M
      • Kato S
      • Sugiyama D
      • Kuwabara K
      • Takeuchi A
      • Akiyama M
      • Okamura T
      • Ebbels TMD
      • Elliott P
      • Tomita M
      • Sato A
      • Suzuki C
      • Sugimoto M
      • Soga T
      • Takebayashi T
      Reliability of plasma polar metabolite concentrations in a large-scale cohort study using capillary electrophoresis-mass spectrometry.
      ]. Fasting plasma samples were used for metabolome profiling via CE-time-of-flight MS (TOFMS). CE-TOFMS analysis of cationic and anionic metabolites was performed as previously described [
      • Sugimoto M
      • Wong DT
      • Hirayama A
      • Soga T
      • Tomita M
      Capillary electrophoresis mass spectrometry-based saliva metabolomics identified oral, breast and pancreatic cancer-specific profiles.
      ]. Raw data were processed using our proprietary software (MasterHands) [
      • Sugimoto M
      • Wong DT
      • Hirayama A
      • Soga T
      • Tomita M
      Capillary electrophoresis mass spectrometry-based saliva metabolomics identified oral, breast and pancreatic cancer-specific profiles.
      ]. The absolute concentrations of 94 charged metabolites (54 cations and 40 anions) that were expected to be detected in ≥ 20% of plasma samples were measured. Serum levels of E2, FSH, and glycated hemoglobin (HbA1c) were determined via immunoassay. Total cholesterol (TC), triglyceride, and fasting plasma glucose (FPG) levels were analyzed using enzymatic methods. High-density lipoprotein-cholesterol (HDL-C) levels were measured by a direct method. Low-density lipoprotein-cholesterol (LDL-C) levels were calculated using the Friedwald equation. Dyslipidemia was defined as one or more of the following findings: LDL-C ≥ 140 mg/dL, HDL-C < 40 mg/dL, or triglyceride ≥ 150 mg/dL. Diabetes was defined as either HbA1c (National Glycohemoglobin Standardization Program) ≥ 6.5% or FPG ≥ 126 mg/dL.

      2.4 Statistical analysis

      The participant characteristics by menopausal status were statistically compared via one-way ANOVA or the Kruskal-Wallis test for continuous variables and the chi-squared test for categorical variables. All metabolite concentrations were log-transformed and treated as continuous variables. For samples with metabolite levels below the limit of detection (LOD), values were substituted using half of the LODs. To examine and compare concentration differences according to age and menopausal status, heatmaps were constructed using MetaboAnalyst 4.0 (Quebec, Canada). To investigate the relationship between 94 metabolites and menopausal status, linear regression analysis was performed among the three groups using each metabolite concentration as the dependent variable and the following factors as covariates: age, BMI, SBP, smoking and drinking status. Six conventional lipid and glucose markers (TC, LDL-C, HDL-C, triglycerides, FPG, and HbA1c), known to be altered by the menopausal status, were also included in the analysis to compare and interpret the results of charged metabolites. We calculated P values using the Benjamini and Hochberg false discovery rate method [
      • Benjamini Y
      • Yekutieli D
      The control of the false discovery rate in multiple testing under dependency.
      ], a commonly used approach for testing multiple hypotheses [
      • Narum SR
      Beyond Bonferroni: Less conservative analyses for conservation genetics.
      ]. Adjusted average concentrations were calculated for each metabolite by group using the aforementioned covariates. Age stratification analysis was performed (<50 and ≥50 years) to further eliminate confounding by age. Because no postmenopausal subjects aged 40–41 years and no premenopausal subjects aged ≥ 56 years were identified, the stratification analysis was performed using the data of women aged 42–55 years. Various sensitivity analyses were performed to exclude the effects of other confounders, such as physical activity adjustment, limiting data to subjects with 18.5 ≤ BMI < 25, excluding women with dyslipidemia and/or diabetes, current smokers, regular drinkers, or regular supplement users.
      All statistical analyses were performed using R.3.6.2 (R Core Team 2019, R Foundation for Statistical Computing, Vienna, Austria).

      3. Results

      3.1 Population characteristics

      Among the 1193 participants, 529 (44.3%) were premenopausal, 132 (11.1%) were in MT, and 532 (44.6%) were postmenopausal. The characteristics of the participants are presented in Table 1. The average ages in the premenopausal, MT, and postmenopausal groups were 44.9, 50.3, and 56.1 years, respectively (p < 0.01). AAM in postmenopausal women was 50.1±3.4 years, consistent with the national average (50.0±0.1 years) [
      • Hirota K
      • Honjo H
      • Shintani M
      Factors affecting menopause.
      ]. Significant differences were found in BP, FPG and lipid levels, and smoking and drinking status among the groups. E2 and FSH levels confirmed that the classification of menopausal status was appropriate. There were no significant differences in BMI or the prevalence of diabetes. The prevalence of hypertension and dyslipidemia were higher in the postmenopausal group.
      Table 1Characteristics of the study population (n= 1193).
      PremenopauseMenopausal transitionPostmenopausePf
      (n= 529)(n= 132)(n= 532)
      Age, years44.9 ± 3.550.3 ± 2.556.1 ± 3.2<0.01
      40–44 yearsa257(48.6%)3(2.3%)2(0.4%)
      45–49 yearsa213(40.3%)44(33.3%)12(2.3%)
      50–55 yearsa59(11.2%)84(63.6%)185(34.8%)
      56–60 yearsa0(0%)1(0.8%)333(62.6%)
      Age at menopause, years--50.1 ± 3.4
      Estradiol, pg/mLb, c114.0(64.7, 186.6)40.1(2.65, 135.0)2.65(2.65, 2.65)<0.01
      FSH, mIU/mLb, d6.22(3.92, 9.82)43.0(16.6, 68.9)67.6(53.8, 83.7)<0.01
      Body mass index, kg/m221.8 ± 3.322.3 ± 3.622.2 ± 3.0N.S.
      Systolic blood pressure, mmHg118.0 ± 14.7124.5 ± 19.4122.2 ± 17.5<0.01
      Fasting plasma glucose, mg/dL91.0 ± 8.993.0 ± 11.994.6 ± 9.4<0.01
      HbA1c (NGSP), %5.4 ± 0.35.5 ± 0.35.6 ± 0.3<0.01
      Total cholesterol, mg/dL196.9 ± 30.8215.2 ± 31.9222.3 ± 33.6<0.01
      Triglyceride, mg/dLb64.0(48.0, 83.0)68.5(53.5, 107.3)78.0(59.0, 107.0)<0.01
      HDL-cholesterol, mg/dL73.3 ± 16.376.8 ± 18.773.2 ± 16.5N.S.
      LDL-cholesterol, mg/dL109.0 ± 28.6128.6 ± 28.8130.9 ± 31.6<0.01
      Daily physical activity, METs × h/daye25.6 ± 13.325.8 ± 11.527.6 ± 14.5N.S.
      Current smoker, yesa51(9.6%)10(7.6%)29(5.5%)<0.05
      Any current alcohol intake, yesa228(43.1%)57(43.2%)179(33.7%)<0.01
      Hypertension, yesa53(10.0%)32(24.2%)84(15.8%)<0.01
      Diabetes mellitus, yesa3(0.6%)1(0.8%)1(0.2%)N.S.
      Dyslipidemia, yesa92(17.4%)40(30.3%)228(42.9%)<0.01
      FSH, Follicle-stimulating hormone; HDL, high-density lipoprotein; LDL, low-density lipoprotein; NGSP, National Glycohemoglobin Standardization Program; N.S., not significant.
      Data are reported as the mean (standard deviation) unless stated otherwise.
      aReported as numbers (percentage).
      bReported as the median (interquartile range).
      cInformation on estradiol was missing for 37 participants.
      dInformation on FSH was missing for 38 participants.
      eInformation on daily physical activity was missing for one participant.
      fAnalysis of variance was used for comparisons of group means. The Kruskal-Wallis test was used for comparisons of age, triglyceride, fasting plasma glucose, hemoglobin A1c, and daily physical activity. Fisher's exact test was used to compare proportions. P < 0.05 was considered significant.

      3.2 Heatmaps of metabolomic differences associated with age and menopausal status

      Fig. 2 presents the heatmaps of metabolomic differences according to age and menopausal status. In each heatmap, log-transformed metabolomic concentrations were scaled to standard scores, and the mean concentration by age or menopausal status was illustrated as the central value of each color palette. As the plasma concentration altered with age, a notable shift in colors were observed around the age of 50. The heatmap constructed according to menopausal status also revealed a trend with similar features. Namely, as women transitioned from premenopause to MT and postmenopause, metabolite concentrations altered from low to high. To eliminate the influence of age, data for women aged 45–55 years were used to create a heatmap presenting metabolomic differences in 272 premenopausal, 128 MT, and 197 postmenopausal women, and similar results were obtained (see Supplementary Figure 1).
      Fig 2:
      Fig. 2Heatmaps presenting (a) age-related and (b) menopausal status-related changes in metabolite levels.
      (a) The x-axis presents age in 1-year increments between 40 and 60 years. The y-axis presents individual metabolites. The color was standardized to a mean of 0 and a variance of 1 for each metabolite, with blue indicating low density and red indicating high density. At an age of approximately 50 years, the color changed for several metabolites from blue to red, indicating a rapid change in concentration at this age.
      (b) The x-axis presents the status of menopause, presenting three groups from the left: premenopause, menopausal transition, and postmenopause. The y-axis is the same as that in (a).
      SDMA, symmetric dimethylarginine, ADMA, asymmetric dimethylarginine, CSSG, cysteine-glutathione disulfide. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

      3.3 Charged metabolites related with the menopausal status

      Linear regression analyses identified multiple charged metabolites associated with the menopausal status. With premenopausal women being the reference, univariate analysis identified 11 charged metabolites with significantly different levels in the MT group and 56 in the postmenopausal group. When age was added to the model, the numbers of significant metabolites in the aforementioned groups decreased to 5 and 15, respectively. The same results were observed in the model adjusted for age, BMI, SBP, smoking and drinking habits. The charged metabolites exhibiting significant differences between the pre- and postmenopausal groups were ornithine, taurine, glutamine, lysine, carnitine, hydroxyproline, trans-aconitate, betaine, cis-aconite, isocitrate, succinate, cystine, proline betaine, arginine, and valine, all of which had higher levels in the postmenopausal group than in the premenopausal group (Table 2). When classified by pathways, these metabolites were related to the tricarboxylic (TCA) cycle, urea cycle, and homocysteine metabolism, some of which are associated with the risk of arteriosclerosis. Five substances (ornithine, glutamine, lysine, isocitrate, and taurine) had significantly higher concentrations in the MT group than in the premenopausal group, all of which were included among the 15 metabolites showing significant differences between the pre- and post-menopausal groups (see Table 3). Fold change values of the fully adjusted model in Table 2 were larger than those in Table 3, suggesting that the metabolite concentrations in the MT group were intermediate between those in the pre- and postmenopausal groups. However, when MT and postmenopausal groups were statistically compared, none of the 15 metabolites were found significantly different between the two (Supplementary Table 1).
      Table 2Metabolites with significant differences in concentrations between the pre- and postmenopausal groups.
      PathwaysMetabolitesModel 1Model 2Model 3
      Fold change95% CIPaFDRbFold change95% CIPaFDRbFold change95% CIPaFDRb
      Urea cycle; arginine and proline metabolismOrnithine1.201.16–1.235.50E-341.80E-321.171.12–1.254.50E-094.50E-071.181.11–1.245.10E-095.10E-07
      Methionine, cysteine, SAM, and taurine metabolismTaurine1.141.11–1.161.30E-211.50E-201.141.09–1.201.00E-075.20E-061.151.09–1.208.60E-084.30E-06
      Glutamate metabolismGlutamine1.091.08–1.121.40E-312.80E-301.071.04–1.111.90E-065.90E-051.071.04–1.102.20E-067.30E-05
      Lysine metabolismLysine1.131.11–1.152.60E-263.70E-251.111.06–1.152.40E-065.90E-051.101.06–1.153.10E-067.80E-05
      Carnitine metabolismCarnitine1.131.11–1.157.90E-332.00E-311.081.04–1.132.20E-053.60E-041.091.05–1.131.80E-053.00E-04
      Urea cycle; arginine and proline metabolismHydroxyproline1.061.02–1.110.0030.0081.161.08–1.256.80E-057.60E-041.171.08–1.253.80E-055.40E-04
      TCA cycletrans-Aconitate1.461.31–1.635.20E-112.70E-101.551.26–1.925.50E-057.60E-041.551.25–1.916.90E-058.70E-04
      Glycine, serine, and threonine metabolismBetaine1.151.12–1.192.60E-202.60E-191.121.06–1.199.20E-059.20E-041.121.06–1.188.80E-059.10E-04
      TCA cyclecis-Aconitate1.121.08–1.151.90E-111.10E-101.121.05–1.195.50E-040.0051.111.04–1.186.80E-040.006
      TCA cycleIsocitrate1.071.02–1.120.0030.0061.161.07–1.274.30E-040.0041.161.06–1.266.90E-040.006
      TCA cycleSuccinate1.061.03–1.091.40E-054.10E-051.091.04–1.161.00E-030.0081.091.03–1.151.40E-030.01
      Methionine, cysteine, SAM, and taurine metabolismCystine1.161.14–1.177.60E-507.60E-481.051.02–1.092.70E-030.021.051.02–1.092.60E-030.02
      Food component/plantProline betaine1.311.14–1.511.90E-040.0011.481.12–1.935.30E-030.031.481.13–1.945.10E-030.03
      Urea cycle; arginine and proline metabolismArginine1.091.06–1.126.90E-124.00E-111.071.02–1.134.80E-030.031.071.02–1.125.50E-030.03
      Leucine, isoleucine, and valine metabolismValine1.061.04–1.081.90E-108.80E-101.051.01–1.086.80E-030.041.051.01–1.095.80E-030.03
      SterolLDL-C1.211.17–1.256.40E-311.10E-291.151.08–1.223.90E-067.90E-051.141.08–1.217.10E-061.40E-04
      SterolTotal-C1.131.11–1.154.40E-362.20E-341.071.04–1.126.30E-057.60E-041.071.04–1.119.10E-059.10E-04
      LDL-C, low-density lipoprotein cholesterol; SAM, S-adenosylmethionine; TCA, tricarboxylic acid; Total-C, total cholesterol.
      Model 1: unadjusted. Model 2: adjusted for age. Model 3: adjusted for age, body mass index, systolic blood pressure, current smoking habits (yes or no), and drinking habits (yes or no).
      a Raw P values are shown.
      b False discovery rate p  values are shown. p < 0.05 was considered significant.
      Table 3Substances with significant differences in concentrations between the premenopause and menopausal transition groups.
      Model 1Model 2Model 3
      PathwaysMetabolitesFold change95% CIPaFDRbFold change95% CIPaFDRbFold change95% CIPaFDRb
      Urea cycle; arginine and proline metabolismOrnithine1.141.09–1.209.50E-094.80E-071.131.07–1.191.50E-061.50E-041.131.07–1.192.20E-062.20E-04
      Glutamate metabolismGlutamine1.071.04–1.093.70E-081.20E-061.061.03–1.083.10E-051.10E-031.061.03–1.091.10E-055.30E-04
      Lysine metabolismLysine1.091.06–1.132.40E-076.00E-061.081.04–1.133.40E-051.10E-031.081.04–1.133.50E-051.20E-03
      TCA cycleIsocitrate1.121.04–1.201.60E-031.50E-021.161.08–1.261.20E-042.40E-031.161.07–1.252.40E-044.80E-03
      Methionine, cysteine, SAM, and taurine metabolismTaurine1.081.03–1.132.90E-043.20E-031.081.03–1.137.20E-040.011.081.03–1.131.10E-030.02
      SterolTotal-C1.091.06–1.132.00E-092.00E-071.071.03–1.117.20E-051.80E-031.061.03–1.091.70E-044.10E-03
      SterolLDL-C1.131.07–1.174.10E-066.90E-051.091.04–1.169.20E-040.011.091.03–1.151.30E-030.02
      LDL-C, low-density lipoprotein cholesterol; SAM, S-adenosylmethionine; TCA, tricarboxylic acid; Total-C, total cholesterol.
      Model 1: unadjusted. Model 2: adjusted for age. Model 3: adjusted for age, body mass index, systolic blood pressure, current smoking habits (yes or no), and drinking habits (yes or no).
      a Raw P values are shown.
      b False discovery rate P values are shown. p < 0.05 was considered significant.
      Of the six established lipid and glucose indices included in the analysis, TC and LDL-C levels exhibited significant differences in pre- versus post-, and pre- versus MT groups, respectively. Fig. 3 presents the adjusted average concentrations of 15 charged metabolites in addition to TC and LDL-C by menopausal status. The concentration of each metabolite gradually increased as groups transitioned from pre- to post-menopause. The trends observed for charged metabolites were similar to those for TC and LDL-C. The fold change values presented in Tables 2 and 3 also indicated that the percent differences in concentrations among groups were similar between charged metabolites and lipids.
      Fig. 3:
      Fig. 3Comparisons of plasma metabolite concentrations by menopausal status.
      The adjusted mean concentration of each metabolite was calculated using the fully adjusted model (covariates were age, BMI, systolic blood pressure, smoking status, and drinking status). Numbers in the middle of each bar reflect the adjusted mean values. Error bars reflect the 95% confident intervals. *: false discovery rate p < 0.05.
      Age stratification and other sensitivity analyses revealed similar results. Results of the multiple linear regression analysis of all 94 metabolites along with the six conventional markers associated with menopausal status (pre- versus post-menopausal groups) are shown in Supplementary Table 2.

      4. Discussion

      In this study, the associations with menopause on the human metabolome were assessed in more than 1000 middle-aged women from a large-scale population-based cohort study in Japan, specifically focusing on charged metabolites. By examining differences in metabolite concentrations among the premenopausal, MT, and postmenopausal groups, our study revealed various metabolites associated with menopausal status, including both previously reported and newly identified charged metabolites. These included substances related to the TCA cycle, urea cycle, and homocysteine metabolism, such as ornithine, taurine, glutamine, lysine, carnitine, hydroxyproline, trans- and cis-aconitate, betaine, isocitrate, succinate, cystine, proline betaine, arginine, and valine.
      The association between the menopausal transition and amino acid metabolites still remains unclear due to the lack of large studies. Wang et al. investigated the association of menopause and 74 metabolites measured via NMR in more than 1000 midlife women [
      • Wang Q
      • Ferreira DLS
      • Nelson SM
      • Sattar N
      • Ala-Korpela M
      • Lawlor DA
      Metabolic characterization of menopause: cross-sectional and longitudinal evidence.
      ]. They reported that transition from pre- to post-menopause induced metabolic changes, such as increased concentrations of very small very low-density lipoproteins, intermediate-density lipoproteins, and low-density lipoproteins. However, their coverage of charged metabolites was insufficient and only a weak effect was observed for branched-chain and aromatic amino acids. Auro et al. examined the impact of menopause status of 3204 women aged 40–55 on 135 metabolite measures including 9 amino acids and a few small molecules related to cell energy metabolism using NMR-based methods [
      • Auro K
      • Joensuu A
      • Fischer K
      • Kettunen J
      • Salo P
      • Mattsson H
      • Niironen M
      • Kaprio J
      • Eriksson JG
      • Lehtimäki T
      • Raitakari O
      • Jula A
      • Tiitinen A
      • Jauhiainen M
      • Soininen P
      • Kangas AJ
      • Kähönen M
      • Havulinna AS
      • Ala-Korpela M
      • Salomaa V
      • Metspalu A
      • Perola M
      A metabolic view on menopause and ageing.
      ]. They found significantly higher concentrations of glutamine and glycine in postmenopausal women, and suggestive association with tyrosine and valine. The authors suggested that menopause might regulate amino acid metabolism, thereby increasing the risk of CVD, but discussions on underlying mechanisms were limited. Amino acids have been identified as important biomarkers of diabetes and cardiovascular risk [
      • Wang TJ
      • Larson MG
      • Vasan RS
      • Cheng S
      • Rhee EP
      • McCabe E
      • Lewis GD
      • Fox CS
      • Jacques PF
      • Fernandez C
      • O’Donnell CJ
      • Carr SA
      • Mootha VK
      • Florez JC
      • Souza A
      • Melander O
      • Clish CB
      • Gerszten RE
      Metabolite profiles and the risk of developing diabetes.
      ,
      • Cheng S
      • Rhee EP
      • Larson MG
      • Lewis GD
      • McCabe EL
      • Shen D
      • Palma MJ
      • Roberts LD
      • Dejam A
      • Souza AL
      • Deik AA
      • Magnusson M
      • Fox CS
      • O’Donnell CJ
      • Vasan RS
      • Melander O
      • Clish CB
      • Gerszten RE
      • Wang TJ
      Metabolite profiling identifies pathways associated with metabolic risk in humans.
      ], and various platforms and separation methods have been used to measure them in epidemiological research. In the present study, absolute quantification of charged metabolites was possible with reliable accuracy in a large number of samples using CE-MS [
      • Harada S
      • Hirayama A
      • Chan Q
      • Kurihara A
      • Fukai K
      • Iida M
      • Kato S
      • Sugiyama D
      • Kuwabara K
      • Takeuchi A
      • Akiyama M
      • Okamura T
      • Ebbels TMD
      • Elliott P
      • Tomita M
      • Sato A
      • Suzuki C
      • Sugimoto M
      • Soga T
      • Takebayashi T
      Reliability of plasma polar metabolite concentrations in a large-scale cohort study using capillary electrophoresis-mass spectrometry.
      ].
      The levels of the TCA cycle-related metabolites, succinate, cis- and trans-aconitate, and isocitrate, were all higher in the postmenopausal group. The TCA cycle is a major metabolic pathway of the aerobic process that catabolizes organic molecules such as carbohydrates, amino acids, and fatty acids to produce energy. An animal experiment discovered that succinate levels in blood were 4-fold higher in hypertensive rats [
      • Sadagopan N
      • Li W
      • Roberds SL
      • Major T
      • Preston GM
      • Yu Y
      • Tones MA
      Circulating succinate is elevated in rodent models of hypertension and metabolic disease.
      ], and aconitate was also reported to be associated with hypertension [
      • Toyohara T
      • Suzuki T
      • Morimoto R
      • Akiyama Y
      • Souma T
      • Shiwaku HO
      • Takeuchi Y
      • Mishima E
      • Abe M
      • Tanemoto M
      • Masuda S
      • Kawano H
      • Maemura K
      • Nakayama M
      • Sato H
      • Mikkaichi T
      • Yamaguchi H
      • Fukui S
      • Fukumoto Y
      • Shimokawa H
      • Inui K
      • Terasaki T
      • Goto J
      • Ito S
      • Hishinuma T
      • Rubera I
      • Tauc M
      • Fujii-Kuriyama Y
      • Yabuuchi H
      • Moriyama Y
      • Soga T
      • Abe T
      SLCO4C1 transporter eliminates uremic toxins and attenuates hypertension and renal inflammation.
      ]. Although not significant, the levels of other metabolites related to TCA cycle, such as citrate, alpha-ketoglutaric acid, fumarate, and malate, tended to be higher in our postmenopausal women. These results suggest an enhancement of the TCA cycle, leading to increased production of reactive oxygen species via the electron transport system and the progression of atherosclerosis, as observed in postmenopausal CVD and diabetes [
      • Cervantes Gracia K
      • Llanas-Cornejo D
      • Husi H
      CVD and oxidative stress.
      ].
      Furthermore, it has been suggested that promotion of the TCA cycle leads to the increased use of amino acids, and therefore, enhanced activity of the urea cycle to convert toxic ammonia into urea [
      • Shambaugh GE
      Urea biosynthesis I. The urea cycle and relationships to the citric acid cycle.
      ]. In our study, metabolites related to the urea cycle were increased in the postmenopausal group. The urea cycle is the final pathway to remove excess nitrogen from the body and the major pathway for ammonia detoxification in humans [
      • Walker V
      Ammonia toxicity and its prevention in inherited defects of the urea cycle.
      ]. In a metabolomic profiling study of bone mineral density [
      • Miyamoto T
      • Hirayama A
      • Sato Y
      • Koboyashi T
      • Katsuyama E
      • Kanagawa H
      • Fujie A
      • Morita M
      • Watanabe R
      • Tando T
      • Miyamoto K
      • Tsuji T
      • Funayama A
      • Soga T
      • Tomita M
      • Nakamura M
      • Matsumoto M
      Metabolomics-based profiles predictive of low bone mass in menopausal women.
      ], levels of metabolites associated with the urea cycle were increased in low estrogen groups. Increased plasma hydroxyproline levels were reported after surgery in ovariectomized rats [
      • Ma B
      • Li X
      • Zhang Q
      • Wu D
      • Wang G
      • A J
      • Sun J
      • Li J
      • Liu Y
      • Wang Y
      • Ying H
      Metabonomic profiling in studying anti-osteoporosis effects of strontium fructose 1,6-diphosphate on estrogen deficiency-induced osteoporosis in rats by GC/TOF-MS.
      ] and in humans with low bone density [
      • Qi H
      • Bao J
      • An G
      • Ouyang G
      • Zhang P
      • Wang C
      • Ying H
      • Ouyang P
      • Ma B
      • Zhang Q
      Association between the metabolome and bone mineral density in pre- and post-menopausal Chinese women using GC–MS.
      ]. The elevation of hydroxyproline levels in the postmenopausal group of this study may reflect the risk of osteoporosis after menopause from a metabolic aspect. Glutamine, which acts as an ammonium ion donor in the urea cycle, were also significantly increased in the postmenopausal group. High blood levels of glutamine increase the risk of diabetes [
      • Xu FG
      • Tavintharan S
      • Sum CF
      • Woon K
      • Lim SC
      • Ong CN
      Metabolic signature shift in Type 2 Diabetes Mellitus revealed by mass spectrometry-based metabolomics.
      ] and the intima-media thickness of the carotid artery [
      • Wurtz P
      • Raiko JR
      • Magnussen CG
      • Soininen P
      • Kangas AJ
      • Tynkkynen T
      • Thomson R
      • Laatikainen R
      • Savolainen MJ
      • Laurikka J
      • Kuukasjärvi P
      • Tarkka M
      • Karhunen PJ
      • Jula A
      • Viikari JS
      • Kähönen M
      • Lehtimäki T
      • Juonala M
      • Ala-Korpela M
      • Raitakari OT
      High-throughput quantification of circulating metabolites improves prediction of subclinical atherosclerosis.
      ].
      Significant elevations were observed in the postmenopausal group with substances related to the metabolism of homocysteine, a risk factor for arteriosclerosis. Carnitine and its precursor, lysine, are linked with increased risk of CVD and diabetes [
      • Koeth RA
      • Wang Z
      • Levison BS
      • Buffa JA
      • Org E
      • Sheehy BT
      • Britt EB
      • Fu X
      • Wu Y
      • Li L
      • Smith JD
      • DiDonato JA
      • Chen J
      • Li H
      • Wu GD
      • Lewis JD
      • Warrier M
      • Brown JM
      • Krauss RM
      • Tang WH
      • Bushman FD
      • Lusis AJ
      • Hazen SL
      Intestinal microbiota metabolism of L-carnitine, a nutrient in red meat, promotes atherosclerosis.
      ,
      • Razquin C
      • Ruiz-Canela M
      • Clish CB
      • Li J
      • Toledo E
      • Dennis C
      • Liang L
      • Salas-Huetos A
      • Pierce KA
      • Guasch-Ferré M
      • Corella D
      • Ros E
      • Estruch R
      • Gómez-Gracia E
      • Fitó M
      • Lapetra J
      • Romaguera D
      • Alonso-Gómez A
      • Serra-Majem L
      • Salas-Salvadó J
      • Hu FB
      • Martínez-González MA
      Lysine pathway metabolites and the risk of type 2 diabetes and cardiovascular disease in the PREDIMED study: results from two case-cohort studies.
      ]. The significant differences observed in these homocysteine-related metabolites suggest that the risk of CVD is affected by menopause through these metabolic alterations. The possible mechanisms are summarized in Supplementary Figure 2.
      One strength of this study was the characteristics of our study population. Obesity is closely related to menopause as well as alterations in amino acid metabolism, but average BMI of our study population was 22 kg/m2 and still exhibited sharp increases in TC and LDL-C around the age of 50 and during MT, in line with previous studies. Our findings of charged metabolite alterations similar to conventional lipid markers suggest that menopause itself induces changes to both amino acid and lipid metabolism without the influence of body size. Exclusion of external factors that affect the metabolome, such as smoking, drinking, supplement use, or coincident dyslipidemia and diabetes, resulted in similar observations. The results were also comparable when participant age was limited to 45–55 years to focus on the menopause phenomenon. Another study strength was the measurement of E2 and FSH levels to evaluate the appropriateness of menopausal status assessment using questionnaires. Very few studies on menopause and metabolomics have validated self-reported reproductive status using both hormones. We confirmed that participants were properly classified by applying these objective measurements.
      This study had several limitations. As a cross-sectional study, identifying causal relationships was difficult when interpreting the metabolic profiling data. Our findings require careful interpretation when applied to different race/ethnic populations since the study was performed in a single-race population. The number of women in the MT group was much smaller than that in the pre- or post-menopausal groups. This may have affected the significance of the results. The effect of dietary habits was not taken into consideration, which might influence the metabolome differently by menopausal status. It is also possible that confounding from unmeasured variables is present. Accumulation of longitudinal data is expected to permit investigations of whether the changes in metabolite levels occur depending on the menopausal status. Furthermore, although our study assessed plasma metabolite levels, the uncertainty of the relationship between plasma and intracellular metabolite levels is another limitation. Replication analyses in a different population is needed for validation.
      In conclusion, increased blood levels of multiple metabolites were identified by comprehensively examining differences in the levels of charged metabolites associated with menopause among middle-aged women in Japan. These findings might reflect the early changes behind the increased risk of dyslipidemia, diabetes, CVD, and osteoporosis after menopause. Further investigations are needed to develop biomarkers for the prevention and early detection of these diseases.

      5. Contributors

      Keiko Watanabe conceived the study concept and design, conducted the statistical analyses, interpreted the data, and drafted the manuscript.
      Miho Iida participated in refining the study design, data analysis and interpretation, and critical revision of the manuscript for important intellectual content.
      Sei Harada participated in data collection and provided critical review of the content.
      Suzuka Kato participated in data collection and provided critical review of the content.
      Kazuyo Kuwabara participated in data collection and provided critical review of the content.
      Ayako Kurihara participated in data collection and provided critical review of the content.
      Ayano Takeuchi participated in data analysis and interpretation, and critical revision of the manuscript for important intellectual content.
      Daisuke Sugiyama participated in data analysis and interpretation, and critical revision of the manuscript for important intellectual content.
      Tomonori Okamura participated in data analysis and interpretation, and critical revision of the manuscript for important intellectual content.
      Asako Suzuki participated in metabolomics data analysis and critical revision of the manuscript for important intellectual content.
      Kaori Amano participated in metabolomics data analysis and critical revision of the manuscript for important intellectual content.
      Akiyoshi Hirayama participated in metabolomics data analysis and critical revision of the manuscript for important intellectual content.
      Masahiro Sugimoto participated in metabolomics data analysis and critical revision of the manuscript for important intellectual content.
      Tomoyoshi Soga participated in metabolomics data analysis and critical revision of the manuscript for important intellectual content.
      Masaru Tomita participated in metabolomics data analysis and critical revision of the manuscript for important intellectual content.
      Yusuke Kobayashi participated in data interpretation and critical review of the content.
      Kouji Banno participated in data interpretation and critical review of the content.
      Daisuke Aoki participated in data interpretation and critical review of the content.
      Toru Takebayashi participated in refining the study design, took full responsibility for the integrity of the data and the accuracy of the data analysis, and participated in the manuscript reviews and edits.
      All authors approved this research article after careful reading.

      6. Funding

      This study was supported in part by research funds from the Yamagata Prefectural Government (http://www.pref.yamagata.jp/) and the city of Tsuruoka (https://www.city.tsuruoka.lg.jp/), and a Grant-in-Aid for Scientific Research (B) (Grant number 24390168), Grant-in-Aid for Challenging Exploratory Research (Grant number 25670303), and Grant-in-Aid for Young Scientists (Grant number 19K19416) from the Japan Society for the Promotion of Science (http://www.jsps.go.jp/).  The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

      7. Ethical approval

      The study was approved by the Medical Ethics Committee of the Keio University School of Medicine, Tokyo, Japan (approval no. 20110264), and all participants provided written informed consent.

      8. Provenance and peer review

      This article was not commissioned and was externally peer reviewed.

      9. Research data (data sharing and collaboration)

      There are no linked research data sets for this paper. The ethics approval for the study reported in the paper did not contain provisions for data sharing.

      Acknowledgments

      We would like to thank the residents of Tsuruoka City for cooperating in our study and TMCS team members for putting their effort into the project. The authors would like to thank Takako Hishiki (Division of Translational Research, Keio University Hospital Clinical and Translational Research Center) for her sincere advice.

      Appendix. Supplementary materials

      References

        • Yu B
        • Zanetti KA
        • Temprosa M
        • Albanes D
        • Appel N
        • Barrera CB
        • Ben-Shlomo Y
        • Boerwinkle E
        • Casas JP
        • Clish C
        • Dale C
        • Dehghan A
        • Derkach A
        • Eliassen AH
        • Elliott P
        • Fahy E
        • Gieger C
        • Gunter MJ
        • Harada S
        • Harris T
        • Herr DR
        • Herrington D
        • Hirschhorn JN
        • Hoover E
        • Hsing AW
        • Johansson M
        • Kelly RS
        • Khoo CM
        • Kivimäki M
        • Kristal BS
        • Langenberg C
        • Lasky-Su J
        • Lawlor DA
        • Lotta LA
        • Mangino M
        • Le Marchand L
        • Mathé E
        • Matthews CE
        • Menni C
        • Mucci LA
        • Murphy R
        • Oresic M
        • Orwoll E
        • Ose J
        • Pereira AC
        • Playdon MC
        • Poston L
        • Price J
        • Qi Q
        • Rexrode K
        • Risch A
        • Sampson J
        • Seow WJ
        • Sesso HD
        • Shah SH
        • Shu XO
        • Smith GCS
        • Sovio U
        • Stevens VL
        • Stolzenberg-Solomon R
        • Takebayashi T
        • Tillin T
        • Travis R
        • Tzoulaki I
        • Ulrich CM
        • Vasan RS
        • Verma M
        • Wang Y
        • Wareham NJ
        • Wong A
        • Younes N
        • Zhao H
        • Zheng W
        • Moore SC
        The Consortium of Metabolomics Studies (COMETS): metabolomics in 47 Prospective Cohort Studies.
        Am J Epidemiol. 2019; 188: 991-1012https://doi.org/10.1093/aje/kwz028
        • Darst BF
        • Koscik RL
        • Hogan KJ
        • Johnson SC
        • Engelman CD
        Longitudinal plasma metabolomics of aging and sex.
        Aging. 2019; 11: 1262-1282https://doi.org/10.18632/aging.101837
        • Ke C
        • Hou Y
        • Zhang H
        • Yang K
        • Wang J
        • Guo B
        • Zhang F
        • Li H
        • Zhou X
        • Li Y
        • Li K
        Plasma metabolic profiles in women are menopause dependent.
        PLoS One. 2015; 10e0141743https://doi.org/10.1371/journal.pone.0141743
        • Auro K
        • Joensuu A
        • Fischer K
        • Kettunen J
        • Salo P
        • Mattsson H
        • Niironen M
        • Kaprio J
        • Eriksson JG
        • Lehtimäki T
        • Raitakari O
        • Jula A
        • Tiitinen A
        • Jauhiainen M
        • Soininen P
        • Kangas AJ
        • Kähönen M
        • Havulinna AS
        • Ala-Korpela M
        • Salomaa V
        • Metspalu A
        • Perola M
        A metabolic view on menopause and ageing.
        Nat Commun. 2014; 5: 4708https://doi.org/10.1038/ncomms5708
        • Rist MJ
        • Roth A
        • Frommherz L
        • Weinert CH
        • Krüger R
        • Merz B
        • Bunzel D
        • Mack C
        • Egert B
        • Bub A
        • Görling B
        • Tzvetkova P
        • Luy B
        • Hoffmann I
        • Kulling SE
        • Watzl B
        Metabolite patterns predicting sex and age in participants of the Karlsruhe Metabolomics and Nutrition (KarMeN) study.
        PLoS One. 2017; 12e0183228https://doi.org/10.1371/journal.pone.0183228
        • Wang TJ
        • Larson MG
        • Vasan RS
        • Cheng S
        • Rhee EP
        • McCabe E
        • Lewis GD
        • Fox CS
        • Jacques PF
        • Fernandez C
        • O’Donnell CJ
        • Carr SA
        • Mootha VK
        • Florez JC
        • Souza A
        • Melander O
        • Clish CB
        • Gerszten RE
        Metabolite profiles and the risk of developing diabetes.
        Nat. Med. 2011; 17: 448-453https://doi.org/10.1038/nm.2307
        • Dal Canto E
        • Ceriello A
        • Rydén L
        • Ferrini M
        • Hansen TB
        • Schnell O
        • Standl E
        • Beulens JW
        Diabetes as a cardiovascular risk factor: an overview of global trends of macro and micro vascular complications.
        Eur. J. Prev. Cardiol. 2019; 26: 25-32https://doi.org/10.1177/2047487319878371
        • Soriguer F
        • Morcillo S
        • Hernando V
        • Valdés S
        • Ruiz de Adana MS
        • Olveira G
        • Fuentes EG
        • González I
        • Tapia MJ
        • Esteva I
        • Rojo-Martínez G
        Type 2 diabetes mellitus and other cardiovascular risk factors are no more common during menopause: longitudinal study.
        Menopause. 2009; 16: 817-821
        • Brand JS
        • van der Schouw YT
        • Onland-Moret NC
        • et al.
        Age at menopause, reproductive life span, and type 2 diabetes risk: results from the EPIC-InterAct study.
        Diabetes Care. 2013; 36: 1012-1019https://doi.org/10.2337/dc12-1020
        • Harada S
        • Takebayashi T
        • Kurihara A
        • Akiyama M
        • Suzuki A
        • Hatakeyama Y
        • Sugiyama D
        • Kuwabara K
        • Takeuchi A
        • Okamura T
        • Nishiwaki Y
        • Tanaka T
        • Hirayama A
        • Sugimoto M
        • Soga T
        • Tomita M
        Metabolomic profiling reveals novel biomarkers of alcohol intake and alcohol-induced liver injury in community-dwelling men.
        Environ. Health Prev. Med. 2016; 21: 18-26https://doi.org/10.1007/s12199-015-0494-y
        • Iida M
        • Harada S
        • Kurihara A
        • Fukai K
        • Kuwabara K
        • Sugiyama D
        • Takeuchi A
        • Okamura T
        • Akiyama M
        • Nishiwaki Y
        • Suzuki A
        • Hirayama A
        • Sugimoto M
        • Soga T
        • Tomita M
        • Banno K
        • Aoki D
        • Takebayashi T
        Profiling of plasma metabolites in postmenopausal women with metabolic syndrome.
        Menopause. 2016; 23: 749-758https://doi.org/10.1097/GME.0000000000000630
        • Harada S
        • Hirayama A
        • Chan Q
        • Kurihara A
        • Fukai K
        • Iida M
        • Kato S
        • Sugiyama D
        • Kuwabara K
        • Takeuchi A
        • Akiyama M
        • Okamura T
        • Ebbels TMD
        • Elliott P
        • Tomita M
        • Sato A
        • Suzuki C
        • Sugimoto M
        • Soga T
        • Takebayashi T
        Reliability of plasma polar metabolite concentrations in a large-scale cohort study using capillary electrophoresis-mass spectrometry.
        PLoS One. 2018; 13e0191230https://doi.org/10.1371/journal.pone.0191230
        • Sugimoto M
        • Wong DT
        • Hirayama A
        • Soga T
        • Tomita M
        Capillary electrophoresis mass spectrometry-based saliva metabolomics identified oral, breast and pancreatic cancer-specific profiles.
        Metabolomics. 2010; 6: 78-95https://doi.org/10.1007/s11306-009-0178-y
        • Benjamini Y
        • Yekutieli D
        The control of the false discovery rate in multiple testing under dependency.
        Ann. Stat. 2001; 29: 1165-1188
        • Narum SR
        Beyond Bonferroni: Less conservative analyses for conservation genetics.
        Conserv. Genet. 2006; 7: 783-787https://doi.org/10.1007/s10592-005-9056-y
        • Hirota K
        • Honjo H
        • Shintani M
        Factors affecting menopause.
        Adv. Obstet. Gynecol. 1995; 47: 389-392
        • Wang Q
        • Ferreira DLS
        • Nelson SM
        • Sattar N
        • Ala-Korpela M
        • Lawlor DA
        Metabolic characterization of menopause: cross-sectional and longitudinal evidence.
        BMC Med. 2018; 16: 17https://doi.org/10.1186/s12916-018-1008-8
        • Cheng S
        • Rhee EP
        • Larson MG
        • Lewis GD
        • McCabe EL
        • Shen D
        • Palma MJ
        • Roberts LD
        • Dejam A
        • Souza AL
        • Deik AA
        • Magnusson M
        • Fox CS
        • O’Donnell CJ
        • Vasan RS
        • Melander O
        • Clish CB
        • Gerszten RE
        • Wang TJ
        Metabolite profiling identifies pathways associated with metabolic risk in humans.
        Circulation. 2012; 125: 2222-2231https://doi.org/10.1161/CIRCULATIONAHA.111.067827
        • Sadagopan N
        • Li W
        • Roberds SL
        • Major T
        • Preston GM
        • Yu Y
        • Tones MA
        Circulating succinate is elevated in rodent models of hypertension and metabolic disease.
        Am. J. Hypertens. 2007; 20: 1209-1215https://doi.org/10.1016/j.amjhyper.2007.05.010
        • Toyohara T
        • Suzuki T
        • Morimoto R
        • Akiyama Y
        • Souma T
        • Shiwaku HO
        • Takeuchi Y
        • Mishima E
        • Abe M
        • Tanemoto M
        • Masuda S
        • Kawano H
        • Maemura K
        • Nakayama M
        • Sato H
        • Mikkaichi T
        • Yamaguchi H
        • Fukui S
        • Fukumoto Y
        • Shimokawa H
        • Inui K
        • Terasaki T
        • Goto J
        • Ito S
        • Hishinuma T
        • Rubera I
        • Tauc M
        • Fujii-Kuriyama Y
        • Yabuuchi H
        • Moriyama Y
        • Soga T
        • Abe T
        SLCO4C1 transporter eliminates uremic toxins and attenuates hypertension and renal inflammation.
        J. Am. Soc. Nephrol. 2009; 20: 2546-2555https://doi.org/10.1681/ASN.2009070696
        • Cervantes Gracia K
        • Llanas-Cornejo D
        • Husi H
        CVD and oxidative stress.
        J. Clin. Med. 2017; 6: 22https://doi.org/10.3390/jcm6020022
        • Shambaugh GE
        Urea biosynthesis I. The urea cycle and relationships to the citric acid cycle.
        Am. J. Clin. Nutr. 1977; 30: 2083-2087
        • Walker V
        Ammonia toxicity and its prevention in inherited defects of the urea cycle.
        Diabetes Obes. Metab. 2009; 11: 823-835
        • Miyamoto T
        • Hirayama A
        • Sato Y
        • Koboyashi T
        • Katsuyama E
        • Kanagawa H
        • Fujie A
        • Morita M
        • Watanabe R
        • Tando T
        • Miyamoto K
        • Tsuji T
        • Funayama A
        • Soga T
        • Tomita M
        • Nakamura M
        • Matsumoto M
        Metabolomics-based profiles predictive of low bone mass in menopausal women.
        Bone. Rep. 2018; 9: 11-18https://doi.org/10.1016/j.bonr.2018.06.004
        • Ma B
        • Li X
        • Zhang Q
        • Wu D
        • Wang G
        • A J
        • Sun J
        • Li J
        • Liu Y
        • Wang Y
        • Ying H
        Metabonomic profiling in studying anti-osteoporosis effects of strontium fructose 1,6-diphosphate on estrogen deficiency-induced osteoporosis in rats by GC/TOF-MS.
        Eur. J. Pharmacol. 2013; 718: 524-532https://doi.org/10.1016/j.ejphar.2013.06.030
        • Qi H
        • Bao J
        • An G
        • Ouyang G
        • Zhang P
        • Wang C
        • Ying H
        • Ouyang P
        • Ma B
        • Zhang Q
        Association between the metabolome and bone mineral density in pre- and post-menopausal Chinese women using GC–MS.
        Mol. Biosyst. 2016; 12: 2265-2275https://doi.org/10.1039/c6mb00181e
        • Xu FG
        • Tavintharan S
        • Sum CF
        • Woon K
        • Lim SC
        • Ong CN
        Metabolic signature shift in Type 2 Diabetes Mellitus revealed by mass spectrometry-based metabolomics.
        J. Clin. Endocr. Metab. 2013; 98: E1060-E10E5https://doi.org/10.1210/jc.2012-4132
        • Wurtz P
        • Raiko JR
        • Magnussen CG
        • Soininen P
        • Kangas AJ
        • Tynkkynen T
        • Thomson R
        • Laatikainen R
        • Savolainen MJ
        • Laurikka J
        • Kuukasjärvi P
        • Tarkka M
        • Karhunen PJ
        • Jula A
        • Viikari JS
        • Kähönen M
        • Lehtimäki T
        • Juonala M
        • Ala-Korpela M
        • Raitakari OT
        High-throughput quantification of circulating metabolites improves prediction of subclinical atherosclerosis.
        Eur. Heart J. 2012; 33: 2307-2316https://doi.org/10.1093/eurheartj/ehs020
        • Koeth RA
        • Wang Z
        • Levison BS
        • Buffa JA
        • Org E
        • Sheehy BT
        • Britt EB
        • Fu X
        • Wu Y
        • Li L
        • Smith JD
        • DiDonato JA
        • Chen J
        • Li H
        • Wu GD
        • Lewis JD
        • Warrier M
        • Brown JM
        • Krauss RM
        • Tang WH
        • Bushman FD
        • Lusis AJ
        • Hazen SL
        Intestinal microbiota metabolism of L-carnitine, a nutrient in red meat, promotes atherosclerosis.
        Nat. Med. 2013; 19: 576-585https://doi.org/10.1038/nm.3145
        • Razquin C
        • Ruiz-Canela M
        • Clish CB
        • Li J
        • Toledo E
        • Dennis C
        • Liang L
        • Salas-Huetos A
        • Pierce KA
        • Guasch-Ferré M
        • Corella D
        • Ros E
        • Estruch R
        • Gómez-Gracia E
        • Fitó M
        • Lapetra J
        • Romaguera D
        • Alonso-Gómez A
        • Serra-Majem L
        • Salas-Salvadó J
        • Hu FB
        • Martínez-González MA
        Lysine pathway metabolites and the risk of type 2 diabetes and cardiovascular disease in the PREDIMED study: results from two case-cohort studies.
        Cardiovasc. Diabetol. 2019; 18: 151https://doi.org/10.1186/s12933-019-0958-2