Metabolomics (2016)12:131 DOI 10.1007/s11306-016-1067-9
ORIGINAL ARTICLE
Normal pregnancy-induced amino acid metabolic stress in a longitudinal cohort of pregnant women: novel insights generated from UPLC-QTOFMS-based urine metabolomic study Mu Wang1 • Qiande Liang2 • Han Li1 • Wei Xia1 • Jie Li2 • Yang Peng1 Yuanyuan Li1 • Zengchun Ma2 • Bing Xu1 • Yue Gao2 • Shunqing Xu1
•
Received: 27 February 2016 / Accepted: 21 June 2016 Springer Science+Business Media New York 2016
Abstract Introduction The maternal body often faces unique physiological challenges in amino acid metabolism due to the continuous requirement of nutrients and substrates for fetal development and additional energy stores for labor and lactation during pregnancy. Objective The aims of the present study is to find out the metabolites involved in amino acid metabolism in a large longitudinal healthy pregnant cohort and provide baseline data for future studies of pregnancy and disease from in utero environmental stress factors. Method We conducted a UPLC-QTOFMS based-urine metabolomics study to investigate the dynamic amino acid metabolic profiles and pathways of 232 healthy pregnant women in their first, second and third trimesters. After multivariate classification to select the metabolites with the strongest contributions to dynamic alterations in normal
Shunqing Xu is the primary corresponding author.
Electronic supplementary material The online version of this article (doi:10.1007/s11306-016-1067-9) contains supplementary material, which is available to authorized users. & Yue Gao
[email protected] & Shunqing Xu
[email protected] 1
Key Laboratory of Environment and Health, Ministry of Education & Ministry of Environmental Protection, and State Key Laboratory of Environmental Health (Incubation), School of Public Health, Tongji Medical College, Huazhong University of Science and Technology, Wuhan, Hubei, People’s Republic of China
2
Beijing Institute of Radiation Medicine, Beijing, People’s Republic of China
pregnancy, we applied the method of standard deviation step (SDSD) down for statistical significance analysis to enhance the value of metabolites in clinical practice. Results Kynurenic acid, an endogenous antagonist of Nmethyl-D-aspartate receptors, increased significantly in middle pregnancy. L-aspartyl-4-phosphate, a potential marker for lower tolerance against fatigue of human body, decreased significantly in the third trimester. Cysteinylglycine, a pyrolysis product of glutathione, significantly increased in late pregnancy. These findings presented a novel insight into normal pregnancy-related regulation of the generation of excitatory neurotransmitter receptor antagonists, maternal fatigue, oxidative stress and so on. Conclusion This normal pregnancy related amino acid metabolic profile as well as the pathways information might be valuable to explore the complex mechanisms of physiological metabolic challenge in amino acid metabolism with the potential capacity to generate a novel hypothesis, which in turn could provide an ideal start for a large-scale epidemiological study of women who subsequently develop diseases, e.g., gestational depression. Keywords Metabolomics UPLC-QTOFMS Pregnancy Amino acid metabolism Metabolic stress
1 Introduction Due to the continuous requirements of nutrients and substrates for fetal development and additional energy stores for labor and lactation, profound adaptive regulation of metabolism, including protein, carbohydrate, and fat metabolism, occurs during the course of normal pregnancy (Lowe and Karban 2014). Protein and nitrogen metabolism play an extremely important role in pregnancy and the
123
131
Page 2 of 11
development of the fetus (Luan et al. 2014). For conservation and accretion of nitrogen in the maternal and fetal bodies, the integral regulation of whole-body protein and nitrogen metabolism begin early in gestation (Duggleby and Jackson 2002). An anabolic condition occurs in pregnancy because the increased volume of many maternal tissues accounts for approximately 60 % of the increase in body weight (Di Cianni et al. 2003). Inhibition of the urea cycle and branched-chain amino acid metabolism have also been found in normal pregnancy and have been related to insulin resistance (Kalhan et al. 1998). Although there is a general consensus that placentally and non-placentally derived hormones during normal pregnancy are underlying drivers of many of the observed adaptations (Angueira et al. 2015), our knowledge remains incomplete regarding the exact mechanisms involved in normal pregnancy-induced amino acid metabolic stress. Metabolomics is the quantitative measurement of the multiparametric time-related metabolic responses of a complex (multicellular) system to a pathophysiological intervention or genetic modification (Nicholson et al. 1999, 2002). Mass spectrometry (MS) and nuclear magnetic resonance (NMR), which can provide multiparametric information about many classes of molecules, often at the same time, are the most common technological platforms for metabolomic study (Nicholson 2006). Such approaches have already been applied in amniotic fluid, maternal blood plasma and urine, umbilical cord blood, and placental cell cultures for prenatal health considerations (Fanos et al. 2013). In recent years, metabolomic studies have been conducted in normal pregnancy, and very meaningful research results have been reported (Luan et al. 2014; Diaz et al. 2013; Pinto et al. 2015; Lindsay et al. 2015). Of these four metabolomics studies, three of them, each of which had a cross-sectional study design by recruiting different subjects at each time point, reported maternal metabolomic changes in healthy pregnant women (Luan et al. 2014; Diaz et al. 2013; Pinto et al. 2015); thus, interference introduced by interindividual variation was inevitable. Further, one longitudinal cohort of healthy pregnant women was employed to investigate the blood profiles of amino acids and lipids during normal pregnancy (Lindsay et al. 2015). However, the metabolic trajectory of longitudinal cohorts of healthy pregnant women have only been evaluated using HPLC–MS-based blood amino acid profiles (Lindsay et al. 2015); results from urine samples obtained in a non-traumatic manner are still lacking. In this study, a UPLC-QTOFMS-based untargeted metabolomics approach was employed to investigate dynamic variations in the intermediates involved in amino acid metabolism in a cohort of 232 healthy pregnant women during the course of pregnancy. Partial leastsquares-discriminant analysis (PLS-DA), combined with
123
M. Wang et al.
the standard deviation step down (SDSD) method, which placed greater focus on metabolite concentration information than classic false discovery rate (FDR) methods, was also introduced in the present study to reduce type II error rates and to improve the statistical efficiency (Wang et al. 2013).
2 Materials and methods 2.1 Materials and equipment L-tryptophan
d5 (Cambridge Isotope Laboratories, UK), formic acid (CNW Technologies GmbH, Germany), methanol (Sinopharm, China), acetonitriles (Fisher Scientific, USA), an Acquity TM ultra performance liquid chromatography system (Waters, USA), Synapt Mass Spectrometry (Waters, USA), Masslynx 4.1 (Waters, USA), an Acquity UPLC HSS T3 Column (2.1 9 100 mm, 1.8 lm, Waters, USA), and a water purifier (Millipore Simplicity, Germany) were used for this study. 2.2 Clinical samples A total of 232 healthy pregnant women, who were [18 (28.17 ± 3.24) years old with a singleton, intrauterine pregnancy and non-diabetic, from the Maternal and Child Health Hospital of Wuhan City, China, were recruited in their first trimesters and were prospectively followed until the end of their pregnancies. Women were invited to attend three prenatal visits for the study, once in each trimester, occurring at gestational ages of the first trimester (12.78 ± 1.18 weeks), second trimester (23.21 ± 3.00 weeks) and third trimester (34.45 ± 4.67 weeks), respectively. The demographic characteristics of these individuals are recorded in Table 1. A total of 696 urine samples from 232 healthy subjects were collected at each visit and were frozen at -80 C until analysis. A minimized freeze–thaw cycle was assured to reduce introduced interference as much as possible. The research protocol was approved by the Ethics Committees of the Tongji Medical College, Huazhong University of Science and Technology and by the study hospital. 2.3 Nontargeted metabolic profile acquisition and data analysis A total of 696 aliquots of 50 ll urine samples from all of the subjects were diluted with 100 ll of water. L-tryptophan d5 solution (50 ll, 0.25 mg ml-1) was used as an internal standard and was added to the diluted sample to acquire the semi-quantitative dataset of the metabolites (Dunn et al. 2011). The testing sample was fully mixed and centrifuged at 10,000 g for 10 min, and 5 ll were injected
Normal pregnancy-induced amino acid metabolic stress in a longitudinal cohort of pregnant… Table 1 Baseline characteristics of the subjects (N = 232)
Page 3 of 11
131
Characteristics Pre-pregnancy weight (kg)
54.0 ± 7.2a
Height (cm)
161.0 ± 4.0a
No smoking within 6 months before pregnancy (N %)
98.7 %
No smoking during pregnancy (N %)
100.0 %
Primiparous (N %)
87.9 %
a
Values are presented as mean ± standard deviation
into UPLC-QTOFMS for testing. Mobile phase A was 0.1 % formic acid in water (v/v), mobile phase B was 0.1 % formic acid in methanol (v/v), and the flow rate was 0.5 ml min-1. The gradient conditions of the mobile phase were as follows (Want et al. 2010): 0–1 min: 1 % B; 1–3 min: 1–15 % B; 3–6 min: 15–50 % B; 6–9 min: 50–95 % B; 9–10 min: 95 % B; 10–10.1 min: 95–1 % B; and 10.1–12 min: 1 % B. The temperatures of the column and autosampler were maintained at 40 and 4 C, respectively. The capillary voltage was 3.2 and 2.4 kV in positive and negative ionization mode, respectively. The desolvation temperature was 350 C, the sampling cone voltage was 40 V, the extraction cone voltage was 4.0 V, the source temperature was 120 C, the cone gas flow rate was 25 l h-1, and the desolution gas flow was 900 l h-1. The mass was corrected during acquisition with leucine-enkephalin, generating a reference ion at m/z 556.2771 Da [(M?H?)] in positive ion mode and of m/z 554.2615 Da [(M–H)-] in negative ion mode before metabolites profile acquisition. A single pooled QC sample was created with an equal volume of 10 ll of each tested urine sample. At the beginning and end of each batch for UPLC-MS analysis, five QC samples were injected, and then 1 QC sample was tested at a regular interval of every ten test samples. These QC samples were applied for conditioning of the analytical system, signal correction, and quality assurance, as previously described (Want et al. 2010). The variables were deleted if the coefficient of variation (CV) was greater than 20 %. PLS-DA was performed using Markerlynx XS software (Waters, USA) to analyze the differences among the three trimesters. The contributions of variables to classification were ranked by the values of variable importance projection (VIP). A decision tree algorithm was used to select the test with statistical significance for the variables screened out by PLS-DA (Goodpaster et al. 2010). SDSD testing was introduced in this study to reduce type II error rates and to improve the statistical efficiency (Wang et al. 2013). The variables that passed both the rule of VIP [1 and SDSD testing were considered potential markers for the next step in metabolite identification.
2.4 Metabolite identification There are several ion adducts, including [M-H]-, [M?HCOO]-, [2 M-H]-, [M?H]?, [M?NH4]?, [M?Na]?, [M?K]?, [2 M?H]?, [2 M?H?K]2? and [2 M?H?Na]2?, in high-resolution electrospray ionization in negative and positive ion modes, causing many difficulties for metabolite identification because accurate calculation and comparison for each type of adduct are very time consuming. In this study, software for automatic putative batch identification, by matching the measured m/z data list with a reference m/z data list derived from the HMDB database, was developed for rapid metabolite identification (Liang et al. 2015). The identifications of metabolites were further checked whether they were consistent with the chromatographic behavior and the fragmentation patterns. 2.5 Data visualization and biomarker network reconstruction The relative change trend in the corresponding metabolites was visualized by heat maps in R software (R Core Team 2015). Metabolites were further retrieved online (https:// pubchem.ncbi.nlm.nih.gov/edit/ and http://www.kegg.jp/) to obtain detailed chemical and biological information, and the metabolic networks were reconstructed using MetaMapp network software (http://metamapp.fiehnlab.ucdavis. edu/homePage) and cytoscape (Shannon et al. 2003).
3 Results 3.1 Data pretreatment and PLS-DA analysis Urine peak data from three trimesters collected from UPLC-QTOFMS were pre-processed in several steps: data alignment, normalization, missing value correction, scaling and transformation. In this study, the data was normalized with the maximum peak. A mask effect, which refers to low abundance metabolites often masked by broad variations in abundant metabolites, was a common challenge in
123
131
Page 4 of 11
M. Wang et al.
metabolomics data analysis. In this study, unit variance (UV) scaling was applied to avoid the mask effect mentioned above. The validity of the PLS-DA model was assessed by values of R2Y and Q2, the former indicating the goodness of fit and the latter indicating the goodness of prediction. According to PLS-DA scores plots, there was a clear separation between the second and first trimesters, the third and second trimesters, respectively, indicating significant differences in urine metabolite profiles in the different trimesters during pregnancy (Supplementary Figs. 1, 2). 3.2 Maternal urine metabolites involved in amino acid metabolism After PLS-DA, SDSD and creatinine correction, the six most changed metabolites in the second trimester and the five most changed metabolites in the third trimester involved in amino acid metabolism were screened out. Supplementary Table 3 showed the identified information on the normal pregnancy related markers involved in amino acid metabolism in the urine of healthy pregnant women. Tables 2 and 3 show the most changed metabolites involved in amino acid metabolism, including the concrete names, p values, VIP values and fold changes. Although all of the relative concentration data were scaled by UV, the mask effect of low abundance metabolites being covered by high abundance metabolites remained, especially when the heat map was used for data visualization, so the relative concentration data were further zoomed to between 0 and 1 in this study. Figures 1 and 2 display heat maps of relative concentrations of the most changed metabolites involved in amino acid metabolism. Compared with the first trimester,
hydantoin-5-propionic acid and 5-aminoimidazole-4-carboxamide ribonucleotide (AICAR) were significantly decreased, while kynurenic acid, 3-indoleacetonitrile, tyrosol and imidazole-4-acetaldehyde were significantly increased in the second trimester. Compared with the second trimester, L-aspartyl-4-phosphate were significantly decreased, while cysteinylglycine, indole, 3-indoleacetonitrile and indole-5,6-quinone were significantly increased in the third trimester. 3.3 Metabolic network of amino acid intermediates during the course of pregnancy Significantly changed intermediates in amino acid metabolism, identified from UPLC-QTOFMS data, were mapped onto metabolite network graphs, which could integration both biochemical and chemical similarities well (Fig. 3) (Meissen et al. 2012). As shown in Fig. 3a, three metabolites—hydantoin-5-propionic acid, imidazole-4-acetaldehyde and AICAR—were involved in histidine metabolism; the single metabolite tyrosol was involved in tyrosine metabolism; two metabolites—kynurenic acid and 3-indoleacetonitrile—were involved in tryptophan metabolism. In Fig. 3b, the single metabolite L-aspartyl-4phosphate was involved in cysteine and methionine metabolism; the single metabolite of cysteinylglycine was involved in glutathione metabolism; two metabolites—indole and 3-indoleacetonitrile—were involved in tryptophan metabolism; the single metabolite indole-5,6-quinone was involved in tyrosine metabolism. The single metabolite 3-indoleacetonitrile appeared in a tandem increase in the comparison between the second and first trimesters and between the third and second trimesters.
Table 2 Most changed metabolites involved in amino acid metabolism (T2 vs. T1) Metabolites
No correction with creatinine a
FC
Correction with creatinine
b
VIP
p valuea
FCb
VIP
HMDB ID
Metabolite name
p value
HMDB01180
N2-Succinyl-L-glutamic acid 5-semialdehyde
\1.68 9 10-04
0.66
1.53
[9.00 9 10-06
-05
0.94
1.01
HMDB12134
1,2-Dihydroxy-3-keto-5-methylthiopentene
\8.85 9 10
0.66
1.90
[3.67 9 10-06
0.85
1.14
HMDB00715
Kynurenic acid
\5.76 9 10-06
2.32
2.79
\5.17 9 10-06
3.96
2.10
-06
-06
HMDB06524
3-Indoleacetonitrile
\3.84 9 10
1.37
1.40
\9.69 9 10
2.12
1.60
HMDB04284
Tyrosol
\8.78 9 10-06
1.49
1.44
\5.45 9 10-05
1.93
2.23
HMDB01212
Hydantoin-5-propionic acid
\1.03 9 10-05
0.47
2.17
\5.86 9 10-06
0.60
1.47
HMDB03905 HMDB01517
Imidazole-4-acetaldehyde AICAR
\7.76 9 10-06 \3.13 9 10-03
1.50 0.48
1.91 2.57
\8.38 9 10-06 \1.79 9 10-05
2.29 0.60
2.56 1.58
a
SDSD-adjusted p value
b
Fold change (FC) was calculated from the ratio of the arithmetic mean values of the relative concentration of each group generated from reference to the signal of L-tryptophan d5. FC with a value more than 1.00 indicated that the concentration of certain metabolite was up-regulated in the relative group, while FC with a value less than 1.00 indicated that the concentration of certain metabolite was down-regulated in the relative group. T1 and T2 denote the first trimester and the second trimester, respectively
123
Normal pregnancy-induced amino acid metabolic stress in a longitudinal cohort of pregnant…
Page 5 of 11
131
Table 3 Most changed metabolites involved in amino acid metabolism (T3 vs. T2) Metabolites
No correction with creatinine a
FC
VIP
HMDB ID
Metabolite name
p value
HMDB00562
Creatinine
\1.03 9 10-05
0.78
HMDB01180
N2-Succinyl-L-glutamic acid 5-semialdehyde
\7.39 9 10-06
0.59
HMDB12250
L-Aspartyl-4-phosphate
\1.60 9 10-05
0.40
-06
Correction with creatinine
b
p valuea
FCb
VIP
1.81
–
–
–
1.78
[2.07 9 10-04
0.68
1.36
2.12
\5.94 9 10-05
0.51
1.98
-06
HMDB00078
Cysteinylglycine
\3.16 9 10
1.63
2.24
\3.65 9 10
1.86
1.93
HMDB00738
Indole
\1.80 9 10-05
1.91
2.14
\5.22 9 10-06
2.03
1.80
HMDB06524
3-Indoleacetonitrile
\9.79 9 10-06
1.49
1.69
\9.51 9 10-06
1.67
1.46
HMDB06779 HMDB03905
Indole-5,6-quinone Imidazole-4-acetaldehyde
\3.23 9 10-06 \6.42 9 10-06
1.62 0.82
1.94 1.24
\6.76 9 10-06 [1.01 9 10-05
1.83 1.15
1.73 1.08
a
SDSD-adjusted p value
b
Fold change (FC) was calculated from the ratio of the arithmetic mean values of the relative concentration of each group generated from reference to the signal of L-tryptophan d5. FC with a value more than 1.00 indicated that the concentration of certain metabolite was up-regulated in the relative group, while FC with a value less than 1.00 indicated that the concentration of certain metabolite was down-regulated in the relative group. T2 and T3 denote the second trimester and the third trimester, respectively
4 Discussion An important advantage of urine metabolites is that the metabolites in the urine represent the final state of metabolism, while the metabolites in blood most likely continue to participate in metabolism. So far, large-scale data on the urine metabolomics in a cohort of healthy pregnant women remain lacking. In this study, a UPLC-QTOFMS-based untargeted metabolomic approach was employed to investigate dynamic variations in the urine metabolite profiles and metabolic pathways of a cohort of 232 healthy pregnant women during the course of pregnancy. Further, eleven metabolites involved in amino acid metabolism were screened out, and classic amino acid metabolic pathways, including cysteine and methionine metabolism, tryptophan metabolism, tyrosine metabolism, histidine metabolism and glutathione metabolism, were significantly changed during normal pregnancy. Tryptophan is an essential amino acid and many important bioactivators including serotonin, tryptamine, indolepyruvic acid and kynurenic acid are generated from tryptophan metabolism (Badawy 2015). Tryptophan metabolism is extremely important for maternal health and fetal development. During pregnancy, there is an obvious decrease in human maternal plasma total tryptophan due to the high expression on indoleamine 2,3-dioxygenase by human syncytiotrophoblast cells in placenta (Badawy 2015). Kynurenic acid (KYN), indole and 3-indoleacetonitrile are important intermediates generated from tryptophan metabolism. KYN, generated from kynurenine through transamination, is a well-known endogenous antagonist of N-methyl-D-aspartate (NMDA) receptors, and it is an important risk factor for depressive disorder (Pershing et al. 2015; Sublette et al. 2011; Miller et al. 2008).
Previous study shown that 9.1 % of pregnant women met criteria for a major depressive episode in America (Hoertel et al. 2015). Latest research shows KYN is associated with hippocampal activity during autobiographical memory recall in patients with depression (Young et al. 2016). In this study, KYN presented a profound increase in the second trimester (FC = 3.96, compared with the first trimester), and the same change trend was also found in a previous plasma metabolomic study (Luan et al. 2014). The significant increase of KYN in the second trimester indicated that KYN may play an important role in pregnancy induced emotional reactions as a negative neuromodulator. Indole is mainly produced by bacteria as a degradation product of tryptophan. In the third trimester, secretion of indole increased by 2.03 times, indicating a significant increase in the degradation of tryptophan in late pregnancy. 3-indoleacetonitrile is a form of dietary indole common in cruciferous vegetables, such as cabbage, cauliflower, broccoli, and Brussels sprouts, and it has shown the ability to prevent tumors in various animal models and human populations (Ciska and Pathak 2004; Smith et al. 2005; Michnovicz and Bradlow 1990). 3-indoleacetonitrile was increased significantly in the second trimester (FC = 2.12, compared with the first trimester) and in the third trimester (FC = 1.67, compared with the second trimester), which might further have confirmed the enhancement of tryptophan metabolism in pregnant women throughout pregnancy. Tyrosine is an important raw material for the synthesis of thyroid hormone and adrenomedullin (AM) in the body. Hormonal changes in human chorionic gonadotrophin or metabolic demands during pregnancy result in profound alterations in the biochemical parameters of thyroid function (Glinoer 1999; Gru¨n et al. 1997). Moreover, maternal
123
131
Page 6 of 11
Fig. 1 Heat map of relative concentrations of significantly changed metabolites involved in amino acid metabolism for the second trimester compared with first the trimester (the data was corrected by creatinine). T1 and T2 denote the first trimester and the second trimester, respectively. Red indicates low abundance metabolites, green represents high abundance metabolites, and black represents moderate abundance of metabolites
123
M. Wang et al.
Normal pregnancy-induced amino acid metabolic stress in a longitudinal cohort of pregnant…
Page 7 of 11
131
Fig. 2 Heat map of relative concentrations of significantly changed metabolites involved in amino acid metabolism for the third trimester, compared with the second trimester (the data was corrected by creatinine). T2 and T3 denote the second trimester and the third trimester, respectively. Red indicates low abundance metabolites, green represents high abundance metabolites, and black represents moderate abundance of metabolites
123
131
Page 8 of 11
M. Wang et al.
Fig. 3 Amino acid metabolic networks generated from the significantly changed metabolites in different trimesters. a The second trimester compared with the first trimester, b the third trimester compared with the second trimester. The red nodes denote significantly elevated metabolites, the green nodes denote significantly
reduced metabolites, and the grey nodes denote metabolites whose VIP values were more than 1.00 but the raw p value hadn’t pass through the SDSD correction. The size of the nodes is coded with the corresponding FC value. The solid edges denote the relationship generated by KEGG among the different metabolites
plasma AM concentrations increased throughout pregnancy and increased as gestational age progressed (Senna et al. 2008). Tyrosol and indole-5,6-quinone are intermediates in tyrosine metabolism. In our work, tyrosol increased significantly in the second trimester (FC = 1.93, compared with the first trimester), and indole-5,6-quinone increased significantly in the third trimester (FC = 1.83, compared with the second trimester), which were perhaps related to changes in thyroid hormone and AM during normal pregnancy; however, further evidence is needed. Many pregnant women experience maternal fatigue which has been found to be related to not only prenatal depression, anxiety, and a fear of childbirth during pregnancy and postpartum (Cheng et al. 2014). In late pregnancy, mothers are more likely to suffer fatigue because they have to pay more physical strength to cope with the rapid increase in the body weight of the fetus. Previous study considered the reduction of L-aspartyl-4-phosphate as a potential marker for lower tolerance against fatigue of human body (Wang et al. 2015). In our work, L-aspartyl-4-
phosphate, an intermediate involved in cysteine and methionine metabolism, decreased significantly in the third trimester (FC = 0.51, compared with the second trimester), which might be related to maternal fatigue in late pregnancy. Pregnancy is a physiological state associated with enhanced oxidative stress related to high metabolic turnover and elevated tissue oxygen requirements (Shoji and Koletzko 2007; Little and Gladen 1999; Carone et al. 1993). Cysteinylglycine forming from glutathione, homocysteine and cysteine in plasma interacts via redox and disulfide exchange reactions, and reduced, free-oxidized and protein-bound forms of these aminothiol species compromise a dynamic system referred to as redox thiol status (Ozkan et al. 2012; Ueland et al. 1996). Previous studies have shown that the concentration of cysteinylglycine in the third trimester was significantly higher than that in the first and second trimester (Ozkan et al. 2012). In this study, we found a similar change trend of cysteinylglycine in the third trimester (FC = 1.86, compared with
123
Normal pregnancy-induced amino acid metabolic stress in a longitudinal cohort of pregnant…
the second trimester) using a UPLC-QTOFMS based metabolomics study. However, the mean concentration of cysteinylglycine was significantly lower in all trimesters of pregnancy compared with non-pregnant controls (Ozkan et al. 2012). The significant increase of cysteinylglycine in the third trimester of pregnancy could be explained by the adaptive regulation on redox system in the pregnant women body. Histamine has been assumed to be beneficial to embryouterine interactions because of its vasoactive, differentiation and growth-promoting properties (Maintz et al. 2008). Large amounts of diamine oxidase (DAO) produced by the placenta are believed to act as a metabolic barrier against histamine to maintain a balance of bioactive histamine in the maternal or fetal circulation (Maintz et al. 2008). Histidine is the precursor of histamine. In this study, hydantoin-5-propionic acid, imidazole-4-acetaldehyde and AICAR were involved in histidine metabolism. Imidazole4-acetaldehyde was a product of histamine under the catalysis of DAO. The concentration of imidazole-4-acetaldehyde was increased significantly in the second trimester (FC = 2.29, compared with the first trimester), which may be related to histamine depletion in the mid pregnancy. Hydantoin-5-propionic acid is a raw material for glutamate synthesis, and its concentration decreased significantly in the second trimester (FC = 0.60, compared with the first trimester), indicating regulation of alanine, aspartate and glutamate metabolism by histidine metabolism being significantly affected by normal pregnancy. AICAR was an intermediate between histidine metabolism and purine metabolism, and its concentration decreased significantly in the second trimester (FC = 0.60, compared with the first trimester), indicating regulation of purine metabolism by histidine metabolism being significantly affected by normal pregnancy. The dynamic alteration of these biomarkers provides important baseline data of amino acid metabolism of healthy pregnant women in different trimesters, which may be of great potential in clinical diagnosis: pregnant women could collect their urine samples and make a fast detection of these metabolites based on the protocol introduced in the present study, and then they could gauge the progress of the amino acid metabolism at home during pregnancy with the reference to the results presented in this study. However, there is still a long way to go before the real clinical application, because validation from studies of a much larger scale population including enough sample size of the related patients (subjects with pregnancy depression, pregnancy fatigue, as well as some inborn errors, etc.) is essential. And future studies might benefit from metabolomics study of baseline samples post-delivery and therein we can estimate when and how the metabolism of the women experienced delivery returns to normal.
Page 9 of 11
131
5 Concluding remarks This work employed a UPLC-QTOFMS-based metabolomics approach to investigate urine metabolite variations in a large cohort of healthy pregnant women in different trimesters of pregnancy. The results obtained enabled eleven metabolites related to amino acid metabolism to be identified and their variations throughout pregnancy to be followed, thus displaying a detailed signature of amino acid metabolism in normal pregnancy. Within the variations discovered, metabolites in metabolic networks of cysteine and methionine metabolism, tryptophan metabolism, tyrosine metabolism, histidine metabolism and glutathione metabolism were observed here from the most comprehensive perspective in connection with normal pregnancy. Metabolic network-based metabolomic analysis provided an extraordinarily novel view of normal pregnancy progression. Herein, we illustrated the key pathways regulating neurotransmitter generation, oxidative stress so on to explore the mechanisms of normal pregnancy-induced adaptation of amino acid metabolism, providing sufficient potential capacity to generate novel hypotheses of the mechanisms of physiologically metabolic adaptation in healthy pregnant women.
Funding This work was supported by the National Natural Science Foundation of China (21437002, 81372959, 81402649), the R&D Special Fund for Public Welfare Industry (Environment) (201309048), and the Fundamental Research Funds for the Central Universities, HUST (2016YXZD043). Compliance with ethical standards Conflict of interest The authors declare no conflict of interest. Ethical approval All procedures performed in this study were in accordance with the ethical standards of the Ethics Committees of the Tongji Medical College, Huazhong University of Science and Technology, and the Study Hospital of the Maternal and Child Health Hospital of Wuhan City in China, as well as with the 1964 Helsinki declaration and its later amendments or comparable ethical standards. Informed consent Written informed consent was obtained from all individual participants included in the study.
References Angueira, A. R., Ludvik, A. E., Reddy, T. E., Wicksteed, B., Lowe, W. L., Jr., & Layden, B. T. (2015). New insights into gestational glucose metabolism: Lessons learned from 21st century approaches. Diabetes, 64(2), 327–334. Badawy, A. A. (2015). Tryptophan metabolism, disposition and utilization in pregnancy. Bioscience Reports, 35(5), e00261. Cheng, C. Y., Chou, Y. H., Wang, P., Tsai, J. M., & Liou, S. R. (2014). Survey of trend and factors in perinatal maternal fatigue. Nursing & Health Sciences,. doi:10.1111/nhs.12149.
123
131
Page 10 of 11
Carone, D., Loverro, G., Greco, P., Capuano, F., & Selvaggi, L. (1993). Lipid peroxidation products and antioxidant enzymes in red blood cells during normal and diabetic pregnancy. European Journal of Obstetrics, Gynecology, and Reproductive Biology, 51(2), 103–109. Ciska, E., & Pathak, D. R. (2004). Glucosinolate derivatives in stored fermented cabbage. Journal of Agriculture and Food Chemistry, 52(26), 7938–7943. Di Cianni, G., Miccoli, R., Volpe, L., Lencioni, C., & Del Prato, S. (2003). Intermediate metabolism in normal pregnancy and in gestational diabetes. Diabetes Metabolism Research and Reviews, 19(4), 259–270. Diaz, S. O., Barros, A. S., Goodfellow, B. J., Duarte, I. F., Carreira, I. M., Galhano, E., et al. (2013). Following healthy pregnancy by nuclear magnetic resonance (NMR) metabolic profiling of human urine. Journal of Proteome Research, 12(2), 969–979. Duggleby, S. L., & Jackson, A. A. (2002). Protein, amino acid and nitrogen metabolism during pregnancy: How might the mother meet the needs of her fetus? Current Opinion in Clinical Nutrition and Metabolic Care, 5(5), 503–509. Dunn, W. B., Broadhurst, D., Begley, P., Zelena, E., FrancisMcIntyre, S., Anderson, N., et al. (2011). Procedures for largescale metabolic profiling of serum and plasma using gas chromatography and liquid chromatography coupled to mass spectrometry. Nature Protocols, 6(7), 1060–1083. Fanos, V., Atzori, L., Makarenko, K., Melis, G. B., & Ferrazzi, E. (2013). Metabolomics application in maternal-fetal medicine. BioMed Research International, 2013, 720514. Goodpaster, A. M., Romick-Rosendale, L. E., & Kennedy, M. A. (2010). Statistical significance analysis of nuclear magnetic resonance-based metabonomics data. Analytical Biochemistry, 401(1), 134–143. Glinoer, D. (1999). What happens to the normal thyroid during pregnancy? Thyroid, 9(7), 631–635. Gru¨n, J. P., Meuris, S., De Nayer, P., & Glinoer, D. (1997). The thyrotrophic role of human chorionic gonadotrophin (hCG) in the early stages of twin (versus single) pregnancies. Clinical Endocrinology (Oxf), 46(6), 719–725. Hoertel, N., Lo´pez, S., Peyre, H., Wall, M. M., Gonza´lez-Pinto, A., Limosin, F., et al. (2015). Are symptom features of depression during pregnancy, the postpartum period and outside the peripartum period distinct? Results from a nationally representative sample using item response theory (IRT). Depression and Anxiety, 32(2), 129–140. Kalhan, S. C., Rossi, K. Q., Gruca, L. L., Super, D. M., & Savin, S. M. (1998). Relation between transamination of branched-chain amino acids and urea synthesis: evidence from human pregnancy. American Journal of Physiology, 275(3 Pt 1), E423–E431. Luan, H., Meng, N., Liu, P., Feng, Q., Lin, S., Fu, J., et al. (2014). Pregnancy-induced metabolic phenotype variations in maternal plasma. Journal of Proteome Research, 13(3), 1527–1536. Lindsay, K. L., Hellmuth, C., Uhl, O., Buss, C., Wadhwa, P. D., Koletzko, B., et al. (2015). Longitudinal metabolomic profiling of amino acids and lipids across healthy pregnancy. PLoS One, 10(12), e0145794. Liang, Q., Xu, W., Hong, Q., Xiao, C., Yang, L., Ma, Z., et al. (2015). Rapid comparison of metabolites in humans and rats of different sexes using untargeted UPLC-TOFMS and an in-house software platform. European Journal of Mass Spectrometry (Chichester, Eng), 21(6), 801–821. Little, R. E., & Gladen, B. C. (1999). Levels of lipid peroxides in uncomplicated pregnancy: A review of the literature. Reproductive Toxicology, 13(5), 347–352. Lowe, W. L., Jr., & Karban, J. (2014). Genetics, genomics and metabolomics: New insights into maternal metabolism during pregnancy. Diabetic Medicine, 31(3), 254–262.
123
M. Wang et al. Miller, A. L. (2008). The methylation, neurotransmitter, and antioxidant connections between folate and depression. Alternative Medicine Review, 13(3), 216–226. Meissen, J. K., Yuen, B. T., Kind, T., Riggs, J. W., Barupal, D. K., et al. (2012). Induced pluripotent stem cells show metabolomic differences to embryonic stem cells in polyunsaturated phosphatidylcholines and primary metabolism. PLoS One, 7(10), e46770. Michnovicz, J. J., & Bradlow, H. L. (1990). Induction of estradiol metabolism by dietary indole-3-carbinol in humans. Journal of the National Cancer Institute, 82(11), 947–949. Maintz, L., Schwarzer, V., Bieber, T., van der Ven, K., & Novak, N. (2008). Effects of histamine and diamine oxidase activities on pregnancy: A critical review. Human Reproduction Update, 14(5), 485–495. Nicholson, J. K. (2006). Global systems biology, personalized medicine and molecular epidemiology. Molecular Systems Biology, 2, 52. Nicholson, J. K., Connelly, J., Lindon, J. C., & Holmes, E. (2002). Metabonomics: A platform for studying drug toxicity and gene function. Nature Reviews Drug Discovery, 1(2), 153–161. Nicholson, J. K., Lindon, J. C., & Holmes, E. (1999). Metabonomics: Understanding the metabolic responses of living systems to pathophysiological stimuli via multivariate statistical analysis of biological NMR spectroscopic data. Xenobiotica, 29(11), 1181–1189. Ozkan, Y., Yardim-Akaydin, S., Erdem, A., & Sims¸ ek, B. (2012). Variability of total thiol compounds, oxidative and nitrosative stress in uncomplicated pregnant women and nonpregnant women. Archives of Gynecology and Obstetrics, 285(5), 1319–1324. Pinto, J., Barros, A. S., Domingues, M. R., Goodfellow, B. J., Galhano, E., Pita, C., et al. (2015). Following healthy pregnancy by NMR metabolomics of plasma and correlation to urine. Journal of Proteome Research, 14(2), 1263–1274. Pershing, M. L., Bortz, D. M., Pocivavsek, A., Fredericks, P. J., Jørgensen, C. V., Vunck, S. A., et al. (2015). Elevated levels of kynurenic acid during gestation produce neurochemical, morphological, and cognitive deficits in adulthood: Implications for schizophrenia. Neuropharmacology, 90, 33–41. R Core Team. (2015). R: A language and environment for statistical computing. R foundation for statistical computing. Vienna, Austria. https://www.R-project.org/. Senna, A. A., Zedan, M., el-Salam, G. E., & el-Mashad, A. I. (2008). Study of plasma adrenomedullin level in normal pregnancy and preclampsia. Medscape Journal of Medicine, 10(2), 29. Shoji, H., & Koletzko, B. (2007). Oxidative stress and antioxidant protection in the perinatal period. Current Opinion in Clinical Nutrition and Metabolic Care, 10(3), 324–328. Sublette, M. E., Galfalvy, H. C., Fuchs, D., Lapidus, M., Grunebaum, M. F., Oquendo, M. A., et al. (2011). Plasma kynurenine levels are elevated in suicide attempters with major depressive disorder. Brain, Behavior, and Immunity, 25(6), 1272–1278. Shannon, P., Markiel, A., Ozier, O., Baliga, N. S., Wang, J. T., Ramage, D., et al. (2003). Cytoscape: A software environment for integrated models of biomolecular interaction networks. Genome Research, 13(11), 2498–2504. Smith, T. K., Lund, E. K., Clarke, R. G., Bennett, R. N., & Johnson, I. T. (2005). Effects of Brussels sprout juice on the cell cycle and adhesion of human colorectal carcinoma cells (HT29) in vitro. Journal of Agriculture and Food Chemistry, 53(10), 3895–3901. Ueland, P. M., Mansoor, M. A., Guttormsen, A. B., Mu¨ller, F., Aukrust, P., Refsum, H., et al. (1996). Reduced, oxidized and protein-bound forms of homocysteine and other aminothiols in plasma comprise the redox thiol status—A possible element of the extracellular antioxidant defense system. Journal of Nutrition, 126(4 Suppl), 1281S–1284S.
Normal pregnancy-induced amino acid metabolic stress in a longitudinal cohort of pregnant… Wang, B., Shi, Z., Weber, G. F., & Kennedy, M. A. (2013). Introduction of a new critical p value correction method for statistical significance analysis of metabonomics data. Analytical and Bioanalytical Chemistry, 405(26), 8419–8429. Want, E. J., Wilson, I. D., Gika, H., Theodoridis, G., Plumb, R. S., Shockcor, J., et al. (2010). Global metabolic profiling procedures for urine using UPLC-MS. Nature Protocols, 5(6), 1005–1018. Wang, X., Xie, G., Wang, X., Zhou, M., Yu, H., Lin, Y., et al. (2015). Urinary metabolite profiling offers potential for differentiation of
Page 11 of 11
131
liver-kidney yin deficiency and dampness-heat internal smoldering syndromes in posthepatitis B cirrhosis patients. Evidence Based Complementary Alternative Medicine, 2015, 464969. Young, K. D., Drevets, W. C., Dantzer, R., Teague, T. K., Bodurka, J., & Savitz, J. (2016). Kynurenine pathway metabolites are associated with hippocampal activity during autobiographical memory recall in patients with depression. Brain, Behavior, and Immunity, S0889–1591(16), 30098-8.
123
