Regular Articles
Gender- and Age-Associated Differences in Serum Metabolite Profiles among Japanese Populations
2016 年 39 巻 7 号 p. 1179-1186
詳細
2016 年 39 巻 7 号 p. 1179-1186
Serum metabolites can reflect the diffusion/export of biochemicals from various organs. They can serve as biomarkers related to diseases and therapeutic efficacy/toxicity. While studies in Caucasians suggested that subject gender and age can affect circulating metabolite profiles, the Japanese population has not been surveyed. Our objective was to delineate gender- and age-associated differences in serum metabolite profiles among Japanese populations. Using a mass spectrometry-based global metabolomics approach, 516 endogenous metabolites were detected in sera from Japanese individuals. The principal component analysis identified gender as the primary component, followed by age, suggesting that these two criteria were key contributors to variations in the dataset. Gender-associated differences were observed in 31 and 25% of metabolites in the young (age 25–35) and old (ages 55–65) populations, respectively, in redox homeostasis, and in steroid and purine nucleotide metabolism pathways. Age-associated differences were observed in 24 and 23% of metabolites in men and women, respectively. No pathway was commonly highlighted. Thus, gender and age impact on metabolite profiles in the Japanese population. Our results provide useful information to explore biomarkers for clinical applications in the Japanese population and to assess the applicability of known biomarkers identified in other populations to the Japanese population.
Circulating metabolites reflect contributions from various organs. This matrix can provide an overview of “organismal” biology, which would be expected to yield useful biomarkers for somatic responses (to disease or environment). In fact, recent studies demonstrated that several circulating metabolites reflect not only disease state, but can also inform on pharmaceutical efficacy and drug toxicity.1–4) To assess multiple aspects of organismal metabolism, these studies employed a global metabolomics approach, which allows the measurement of a wide range of metabolites simultaneously.5,6) High-throughput measurement of global metabolite levels can also ease the identification of altered metabolic pathways and the evaluation/identification of biomarkers of interest.
One key problem in the identification and/or validation of circulating metabolites as clinical biomarkers in humans is inter-individual variation. Thus, an understanding of classes of metabolites that can vary between individuals is beneficial to choosing clinical biomarkers for therapeutic efficacy as well as drug toxicity. Subject backgrounds such as gender and age have emerged as modulators of the global levels of circulating metabolites. For example, metabolomics approaches have demonstrated that levels of branched chain amino acids and long chain acyl carnitines differ between male and females in overweight Caucasian and African American cohorts.7) Other studies demonstrated that sphingomyelin levels were higher in females, whereas acyl carnitines and metabolites in the urea cycle are higher in German males.8) On the other hand, metabolites related to the tricarboxylic acid (TCA) cycle, lipid metabolism, and oxidative stress increased with age in a mixed population of Caucasian, African American, and Hispanics.9) In addition, another study demonstrated that specific fatty acids and sphingomyelins increased with age in the German population.10) Along with these observations, global metabolomics approaches clearly demonstrated gender- and age-associated differences in small molecule biochemicals as well as lipid metabolites in Caucasian cohorts.11,12) Metabolic phenotyping of the U.K. population also clearly demonstrated variations in biochemicals associated with gender and age.13) Generally, androgens and its derivatives are high in male and progesterone and its derivatives are high in female and those are also age-associated. Given these observations of differential changes in biochemicals with gender and age in Caucasian, African American, and/or Hispanic subjects, it is likely that the Japanese population might also show its own specific patterns of changes, though this has not been investigated. As the second largest market for pharmaceuticals, an assessment of metabolic shifts with gender and age in a Japanese cohort would be an important addition to these other populations.
In the present study, using a global metabolomics approach, we determined the levels of 516 endogenous metabolites, including amino acids, peptides, carbohydrates, relatively hydrophilic lipids, nucleotides, and vitamins, in serum samples obtained from Japanese subjects categorized by either age or gender. Because serum is more resistant to effects of freeze-thaw cycles and 37°C incubation than plasma,11,14) we used serum as a matrix in the present study. To minimize unrelated variations, we controlled subjects’ age (young population, 25–35 years old; old population, 55–65 years old), range of body mass index (BMI) value (18–25.5) and collected samples following an overnight fast. Because circadian rhythms affect circulating metabolite levels,15) we also controlled the time of blood collection (10 a.m.). Our present study provides fundamental information useful for the identification of clinical biomarkers from circulating metabolites as well as for validating previously identified metabolomics biomarker candidates in the Japanese population.
After written informed consent was properly obtained from all participants, blood samples were collected at Yaesu Sakura-dori clinic (Tokyo, Japan) with the aid of NEUES Inc. (Tokyo, Japan). All participants were self-reported healthy volunteers without any medications at least for 1 week. Venous blood was collected from 60 healthy Japanese volunteers in the morning after fasting for 14 h. Participants were categorized into 4 groups as follows: young males (25–35 years old), old males (55–64 years old), young females (25–35 years old), and old females (55–65 years old). We selected these age ranges to overview age-associated differences in Japanese, because previous literature demonstrated that several metabolites gradually changed by age in adults.9,10) Each group included 15 individuals. Subject information is displayed in Table 1. We recruited subject with selected range of BMI value (18–25.5). The BMI value presented statistically significance between 1) young and old male, and 2) old male and female, but their median difference of the value are less than 2. Fresh blood from each individual was drawn into 4-mL Vacutainer Serum Separator Tubes with clot activators (Becton Dickinson, Franklin Lakes, NJ, U.S.A.). Samples were centrifuged and serum was separated within 2 h of blood collection and then immediately frozen. Upon receiving samples to the National Institute of Health Sciences (NIHS), all samples were once thawed on ice, divided into aliquots, and refrozen at −80°C until sample extraction. This study was approved by the Ethics Committees of the Yaesu Sakura-dori Clinic and NIHS, and was performed in accordance with the Declaration of Helsinki.
| Subject groups | Young male | Old male | Young female | Old female |
|---|---|---|---|---|
| Number of individuals | 15 | 15 | 15 | 15 |
| AGE (median) | 25–35 (32) | 55–64 (60) | 25–35 (32) | 55–65 (60) |
| BMI (median) | 18.53–23.52 (21.98) | 20.21–25.15 (23.11) | 18.64–24.98 (20.30) | 18.36–24.68 (21.27) |
| Ethnicity | Japanese | Japanese | Japanese | Japanese |
| Matrix | Serum | Serum | Serum | Serum |
Metabolomics and statistical analyses were conducted at Metabolon Inc.; a description of the Metabolon process can be found in reference16); a detailed description of the metabolon platform can be found in another reference.17) Briefly, data were acquired on a UHPLC MS/MS platform (Waters Aquity UPLC and Thermo Fisher Q-Exactive high resolution/accurate mass spectrometer interfaced with an HESI-II source and Orbitrap mass analyzer operating at 35000 mass resolution) and a GC/MS (Thermo-Finnigan Trace DSQ fast-scanning single-quadrupole mass spectrometer using electron impact ionization (EI) and operated at unit mass resolving power). Sera were subjected to methanol extraction then split into aliquots for analysis by ultrahigh performance liquid chromatography/mass spectrometry (UHPLC/MS) in the positive (LC/MS Pos) and negative ion mode (LC/MS Neg) separated with C18 column, negative ion mode separated with hydrophilic interaction chromatography (HILIC) column (LC/MS HILIC)16) and by gas chromatography/mass spectrometry (GC/MS). Metabolites were identified by automated comparison of ion features to a reference library of chemical standards followed by visual inspection for quality control (described in more detail in ref. 18). Some metabolites, which have not been confirmed based on a standard, were confirmed based on the ion features, including RT, m/z, preferred adducts, and in-source fragmentation. Briefly, three types of controls were analyzed in concert with the experimental samples: samples generated from a pool of human plasma (extensively characterized by Metabolon, Inc.) served as technical replicate throughout the data set, with extracted water samples serving as process blanks. Finally, a cocktail of standards spiked into every analyzed sample allowed instrument performance monitoring. Instrument variability was determined by calculating the median relative standard deviation (RSD) for the standards that were added to each sample prior to injection into the mass spectrometers (median RSD=4% for this study). Overall process variability was determined by calculating the median RSD for all endogenous metabolites (i.e., non-instrument standards) present in 100% of the pooled human plasma samples (median RSD=8% for this study). Experimental samples and controls were randomized across the platform run with QC samples spaced evenly among the injections.
For statistical analyses and data display, any missing values were assumed to be below the lower limits of detection; these values were imputed with the compound minimum (minimum value imputation) instead of zero. Data were median scaled and imputed prior to statistical analysis, an identical volume of serum was extracted per sample and run across the platform. To determine statistical significance, Welch’s two-factor t-tests (comparing young and old or male and female populations) were performed in ArrayStudio (Omicsoft). p<0.05 was considered significant. An estimate of the false discovery rate (Q-value) was calculated to take into account the multiple comparisons that normally occur in metabolomics-based studies, with Q<0.05 used as an indication of high confidence in a result. The average values and standard deviation obtained from normalized levels of each metabolite as well as the categories and pathways of each metabolite are displayed in Supplementary Table S1.
Principal Component Analysis (PCA)Metabolite data, following run-day normalization which minimize the effect of day-to-day variation of mass spectrometry on metabolite levels, and minimum value imputation, were loaded into SIMCA-P+12 (Umetrics, Umea, Sweden), parate-scaled, and analyzed using PCA-X to visualize the variance among the groups evaluated in this study. The PCA-X results were provided as score plots to represent the similarity of overall metabolic profiles.
Pathway Occupancy AnalysisTo construct pathway occupancy maps, pathways represented by more than 5 metabolites were selected and scored with statistically different metabolites within specific pathways (p<0.05, scored as 1; p>0.05, scored as 0). The scored values were divided by the number of metabolites within specific pathways, resulting in the ratio of occupied metabolites that reached statistical significance within a pathway.
Global metabolite profiling resulted in the identification of 516 metabolites in the serum of Japanese subjects (Table 2). These metabolites could be divided into several superpathways, including amino acids (154 metabolites), peptides (25 metabolites), carbohydrates (23 metabolites), energy metabolites (8 metabolites), lipids (251 metabolites), nucleotides (29 metabolites), and cofactors and vitamins (26 metabolites). To assess the effects of grouping variables (gender and age) on organismal metabolism, we first used a PCA to generate a high-level overview of the dataset. As shown in Fig. 1, the male and female sample groups clearly separated into 2 distinct groups, separated mainly by component 1 (R2X=0.112). In addition, young and old sample groups formed partially overlapping groups separated mainly by component 2 (R2X=0.083). The loading plot was also shown in Fig. 1. The relatively low rate of contribution rate of components 1 and 2 indicating that specific metabolites, such as myristoyl carnitine, 5alpha-androstan-3alpha, 17beta-diol disulfate, 5alpha-pregnan-3beta, 20alpha-diol monosulfate, and pregnanolone/allopregnanolone sulfate, contribute to the separation among genders and ages.
| Categories | Pathways | Number of metabolites |
|---|---|---|
| Amino acid | Glycine, serine and threonine metabolism | 8 |
| Alanine and aspartate metabolism | 5 | |
| Glutamate metabolism | 6 | |
| Histidine metabolism | 11 | |
| Lysine metabolism | 10 | |
| Phenylalanine and tyrosine metabolism | 29 | |
| Tryptophan metabolism | 16 | |
| Leucine, isoleucine and valine metabolism | 28 | |
| Methionine, cysteine, sam and taurine metabolism | 15 | |
| Urea cycle, arginine and proline metabolism | 14 | |
| Creatine metabolism | 3 | |
| Polyamine metabolism | 4 | |
| Guanidino and acetamido metabolism | 2 | |
| Glutathione metabolism | 3 | |
| Peptide | Gamma-glutamyl amino acid | 13 |
| Dipeptide derivative | 1 | |
| Dipeptide | 7 | |
| Polypeptide | 2 | |
| Fibrinogen cleavage peptide | 2 | |
| Carbohydrate | Glycolysis, gluconeogenesis, and pyruvate metabolism | 5 |
| Pentose metabolism | 9 | |
| Disaccharides and oligosaccharides | 1 | |
| Fructose, mannose and galactose metabolism | 5 | |
| Aminosugar metabolism | 3 | |
| Energy | TCA Cycle | 7 |
| Oxidative phosphorylation | 1 | |
| Lipid | Short chain fatty acid | 1 |
| Medium chain fatty acid | 7 | |
| Long chain fatty acid | 15 | |
| Polyunsaturated fatty acid (n3 and n6) | 14 | |
| Fatty acid, branched | 3 | |
| Fatty acid, dicarboxylate | 12 | |
| Fatty acid, amide | 3 | |
| Fatty acid, amino | 2 | |
| Fatty acid, keto | 1 | |
| Fatty acid synthesis | 3 | |
| Fatty acid metabolism (also BCAA metabolism) | 4 | |
| Fatty acid metabolism (Acyl glycine) | 3 | |
| Fatty acid metabolism (Acyl carnitine) | 13 | |
| Carnitine metabolism | 2 | |
| Ketone bodies | 2 | |
| Fatty acid, monohydroxy | 13 | |
| Fatty acid, dihydroxy | 2 | |
| Eicosanoid | 1 | |
| Endocannabinoid | 5 | |
| Inositol metabolism | 4 | |
| Phospholipid metabolism | 6 | |
| Lysolipid | 28 | |
| Phosphatidylcholine (PC) | 13 | |
| Phosphatidylethanolamine (PE) | 2 | |
| Phosphatidylinositol (PI) | 3 | |
| Glycerolipid metabolism | 2 | |
| Monoacylglycerol | 13 | |
| Sphingolipid metabolism | 10 | |
| Mevalonate metabolism | 1 | |
| Sterol | 5 | |
| Steroid | 37 | |
| Primary bile acid metabolism | 7 | |
| Secondary bile acid metabolism | 14 | |
| Nucleotide | Purine metabolism, (hypo)xanthine/inosine containing | 6 |
| Purine metabolism, adenine containing | 6 | |
| Purine metabolism, guanine containing | 5 | |
| Pyrimidine metabolism, orotate containing | 1 | |
| Pyrimidine metabolism, uracil containing | 9 | |
| Pyrimidine metabolism, cytidine containing | 1 | |
| Pyrimidine metabolism, thymine containing | 1 | |
| Cofactors and vitamins | Nicotinate and nicotinamide metabolism | 5 |
| Riboflavin metabolism | 2 | |
| Pantothenate and CoA metabolism | 1 | |
| Ascorbate and aldarate metabolism | 5 | |
| Tocopherol metabolism | 8 | |
| Hemoglobin and porphyrin metabolism | 4 | |
| Vitamin B6 metabolism | 1 | |
| Total | 516 |
Data obtained from Japanese human serum samples of young males (blue), old males (green), young females (red), and old females (orange) were analyzed. Red circle in loading plots indicates representative metabolites: a, 5alpha-androstan-3alpha, 17beta-diol disulfate; b, myristoyl carnitine; c, 5alpha-pregnan-3beta, 20alpha-diol monosulfate; d, pregnanolone/allopregnanolone sulfate.
Next, we analyzed the differences in the levels of individual metabolites between genders. The number of metabolites with statistically significant differences (p<0.05) between male and female subjects were 158 (31%) and 130 (25%) out of 516 metabolites in the young and old populations, respectively (Supplementary Table S2). Specifically, 138 and 113 metabolites in young (25–35 years old) and old (55–65 years old) groups, respectively, were dominant in male, with 65 and 53 biochemicals, respectively, showing more than a 50% difference in their levels (Fig. 2A). On the other hand, only 20 and 17 metabolites in young and old populations, respectively, were profoundly increased in females, with 10 and 5 biochemicals showing more than a 50% difference (Fig. 2A). Young and old populations shared 67 and 3 metabolites of the male- and female-enriched metabolites, respectively. Examples of these common male-enriched metabolites were gamma-glutamyl amino acids such as gamma-glutamyl-alanine, gamma-glutamyl-leucine, and gamma-glutamyl-methionine and androgens (4-androsten-3beta,17beta-diol disulfate, 5alpha-androstan-3alpha,17alpha-diol monosulfate, and 5alpha-androstan-3beta,17beta-diol disulfate). Female-enriched metabolites were creatine, choline phosphate, and beta-sitosterol, a diet-derived phytosterol.
To further understand the differences in the metabolic profiles between male and female subjects, we next determined the pathway occupancy of metabolites with significantly different levels between male and female groups (Fig. 2B). An overview of pathway occupancy suggested that several pathways share common trends among young and old populations. In agreement with individual metabolites, gamma-glutamyl amino acids and steroids, which include androgens, scored more than 60% of pathway occupancy in both young and old populations, with the majority being male-enriched metabolites. Purine metabolism (containing guanine) also scored more than 60% of pathway occupancy. In contrast, urea cycle and fatty acid metabolism (Acyl carnitine) scored more than 60% of pathway occupancy only in the young population. The score for these pathways in older individuals was less than 20%, suggesting that gender-related differences in urea cycle and fatty acid metabolism (Acyl carnitine) are only found in the young population.
A. Number of metabolites with statistically significant differences between male and female subjects. Values within boxes indicate the number of metabolites. FC: fold change. B. Pathway occupancy rates of statistically different metabolites between male and female populations. Light red, the ratio of metabolites higher in males than in females; light green, the ratio of metabolites higher in females than in males.
Subsequently, we also analyzed the differences in the levels of individual metabolites between the two age groups. The number of metabolites with statistically significant differences (p<0.05) between young and old subjects were 119 (23%) and 122 (24%) out of 516 metabolites in male and female, respectively (Supplementary Table S3). Specifically, 35 and 41 metabolites in male and female, respectively, were higher in the young than in the old population, with 21 and 26 metabolites showing more than a 50% difference in their levels (Fig. 3A). On the other hand, 84 and 81 metabolites in male and female, respectively, were higher in the old than in the young population, with 32 and 21 metabolites showing more than a 50% level difference (Fig. 3A). Both male and female groups shared 21 and 32 metabolites of the young- and old-enriched metabolites. Examples of metabolites enriched in male and female young individuals were progestogens such as pregnenolone sulfate, 21-hyroxypregnenolone disulfate, and pregnanediol-glucuronide, with higher fold changes in females (1.55–2.17 folds in males vs. 1.75–10.80 folds in females). On the other hand, metabolites enriched in older individuals were several polyunsaturated fatty acids such as eicosapetaenoate, docosapentaenoate, and docosahexaenoate.
To further understand the differences in the metabolic profiles between young and old subjects, we next determined the pathway occupancy of metabolites with significantly different levels between young and old groups (Fig. 3B). Unlike the comparison between young and old populations, a survey of pathway occupancy identified a substantially different trend between males and females. No pathways scored more than 60% pathway occupancy in both males and females. Pathways involved in glycolysis, gluconeogenesis, pyruvate metabolism, TCA cycle, monoacylglycerol, sphingolipid metabolism, and sterol scored more than 60% of pathway occupancy in males, while those of gamma-glutamyl amino acids and steroid scored more than 60% of pathway occupancy in females.
A. Number of metabolites with statistically significant differences between young and old subjects. Values within boxes indicate the number of metabolites. FC: fold change. B. Pathway occupancy rates of statistically different metabolites between young and old populations. Light orange, the ratio of metabolites higher in young than in old subjects; light blue, the ratio of metabolites higher in old than in young subjects.
In the present study, we demonstrated that age and gender have substantial effects on global serum metabolite profiles in a Japanese cohort. In the PCA, samples separated by both age and gender. More than 20% of metabolites consistently presented significantly different levels among either ages or genders. In addition, we also identified pathways associated with gender and age in the Japanese population. Gamma-glutamyl amino acid, steroid, and purine metabolism (containing guanine) are highlighted in gender-difference commonly observed in young and old populations, while the urea cycle and fatty acid metabolism (Acyl carnitine) are specific gender-associated pathways in the young cohort. On the other hand, several highlighted pathways presenting age-associated differences were specific for a single gender: Glycolysis, gluconeogenesis, pyruvate metabolism, TCA cycle, monoacylglycerol, sphingolipid metabolism, and sterol are highlighted in males, while gamma-glutamyl amino acids and steroids are highlighted in females.
Age- and gender-associated differences of circulating metabolites have been extensively characterized in non-Japanese populations. For example, androgens are male-enriched in both young and old populations, with progestogens declining in postmenopausal Caucasian females.11) Metabolites related to the TCA cycle gradually increase with age10) and metabolites related to the urea cycle and acyl carnitine are male-enriched metabolites in a German population.8) Moreover, urinary N-methylated guanine and guanosine are present in higher levels in Polish males than in females.19) In agreement with previous observations, the age- and/or gender-associated differences in androgens, progestogens, acylcarnitines, and metabolites related to guanine, the TCA cycle, and the urea cycle, are also highlighted in the Japanese population. However, precise analysis among genders/ages in the Japanese population indicated that the pathway involved in the TCA cycle is only highlighted in males and those of the urea cycle and fatty acid metabolism (Acyl carnitine) are only highlighted in the young population. Thus, it is possible that age-associated alteration of TCA cycle-related metabolites is specific for males and that gender-associated differences in metabolites related to urea cycle and acyl carnitines are specific to the young population.
In the present study, gamma-glutamyl amino acids were the most highlighted pathway regarding gender- and age-associated differences in the Japanese population. Changes in gamma-glutamyl amino acids can reflect differences in redox homeostasis and oxidative stress. On the other hand, it has been reported that only a few gamma-glutamyl amino acids are gender- or age-associated in Caucasian and mixed populations.9,11) Taken together, these results and the present study indicate that gender- and age-associated differences in gamma-glutamyl amino acids might be specific to the Japanese population. In addition, gamma-glutamyl amino acids such as gamma-glutamyl alanine and gamma-glutamyl leucine have been proposed as blood biomarkers for liver diseases, including drug-induced liver injury, hepatitis C infection, and hepatocellular carcinoma.20) This report also proposed that the combinatorial use of several gamma-glutamyl amino acids would discriminate among different forms of hepatic disease with multiple logistic regressions. When using gamma-glutamyl amino acids as biomarkers for hepatic diseases, age- and gender-related differences should be considered as a confounding factor for their level changes.
Along with gamma-glutamyl amino acids, we demonstrated that effects of gender and age on several metabolites were different compared to previous observations with non-Japanese populations. Such metabolites mainly clustered into lipids, including sphingolipids, monoacylglycerols, and long chain fatty acids. For example, many sphingolipids are female-enriched in Caucasian populations.8,12) In the Japanese population, only a limited number of sphingolipids were female-enriched, which was only observed in the older population. In addition, the current study demonstrated that age-enriched levels of monoacylglycerols were highlighted in Japanese males, while our recent work showed that young-enriched levels of long chain fatty acids were highlighted in Caucasian females.11) No such age-associated differences of metabolites were observed in the other populations. Therefore, circulating levels of several metabolites, especially lipids, might be regulated by different mechanisms among Japanese and non-Japanese populations. Several reports demonstrated that genetic polymorphism affects biochemical levels in serum,21,22) suggesting that genetic variations may play a role in these racial differences in the gender- and/or age-associated variations on circulating levels of metabolites. However, which factors other than genetic variations are responsible for these ethnic differences remains unclear. It has been recently reported that blood levels of sphingolipids are different between traditional lifestyle and non-traditional lifestyle in the Swedish population.23) In addition, sphingolipid levels are strongly associated with fruit and vegetable intake with significant genetic contributions.24) Thus, along with genetic contribution, food preference might be one of such factors by which several metabolites present racially unique age- and gender-associated differences.
There were several important limitations. First, although we selected subject with controlled range of BMI values (18–25.5), the values are statistically significant between 1) young and old male and 2) old male and female. It has been delineated that levels of metabolites, such as lipids and amino acids, associated with BMI values by comparing lean/healthy and obese group.25–27) The difference associated with BMI value might mimics the age- and gender-associated differences in metabolite levels. However, our controlled range usually discriminated as lean/healthy group. In addition, the difference in the median BMI value in the present study (less than 2) are very limited when compared with previous studies (more than 5).25–27) Thus, the impact associated with BMI values would be very limited in the present study. Second, although present study clearly demonstrated the gender- and age-associated differences, the sample numbers were relatively small. However, the minimum sample size reported the differences in metabolite levels were around 15 subjects in each group,27) suggesting 15 subjects are probably enough to overview substantial differences in metabolite levels. In addition, controlled ranges of BMI, ages and sexes, which were not controlled in other studies, would further enhance to reveal the true difference of our interest. Therefore, we considered 15 subjects in each group would be enough for our purpose. Third, although we collected self-reported healthy subjects without any medication at least 1 week, unaware of disease is inevitable. Forth, in the present study, we did not control subject’s alcohol/food consumption as well as smoking that might affect levels of metabolites. Fifth, although we control the rage of sample preparation time up to 2 h, there might be variables on serum preparation time among subject groups and some unstable metabolite may degrade during these time. This factor possibly affected the levels of several unstable metabolites. Moreover, sample numbers used in the present study might not be enough to fully reflect all Japanese population. Therefore, to fulfill these limitations and to determine general range of each metabolite level in Japanese populations, large-scaled study would be required in future study.
In conclusion, using global metabolomics profiling, we surveyed the gender- and age-associated differences in circulating metabolite profiles in the Japanese population. Our results demonstrated that profiles of circulating metabolites are substantially different among genders and ages in the Japanese population. In addition, our results showed that some metabolites such as steroids (androgens/progestogens) and acyl carnitines presented gender- and/or age-associated differences in the Japanese population, similar to those in non-Japanese populations. On the other hand, other metabolites such as sphingolipids and monoacylglycerols presented different trends regarding the effects of gender and age in the Japanese population compared to the Caucasian population. Therefore, our study propose that clinical application of metabolites as biomarkers should be taken into consideration of gender- and age-associated differences in Japanese as well as the racial differences in gender- and age-associated variations in metabolite levels. Taken all together, our results provide useful and fundamental information for exploring and validating biomarkers using circulating metabolites in future clinical studies with the Japanese population and may also help establishing the regulatory guidelines for such studies.
The authors thank Ms. Masayo Urata for experimental support and Ms. Chie Sudo for secretary assistance. This work was supported by the Research on Regulatory Harmonization and Evaluation of Pharmaceuticals, Medical Devices, Regenerative and Cellular Therapy Products, Gene Therapy Products, and Cosmetics (16mk0101045j0102) and the Platform Project for Supporting Drug Discovery and Life Science Research Grants (16ak0101029j0003, 16ak0101043j0602) from the Japan Agency for Medical Research and Development.
The authors declare no conflict of interest.
The online version of this article contains supplementary materials. Supplementary Table S1. Determined metabolites and their relative levels in each groups. Supplementary Table S2. Statistically significant gender-differences in metabolite levels with fold changes. Supplementary Table S3. Statistically significant age-differences in metabolite levels with fold changes.
