Study on amniotic fluid metabolism in the second trimester of Trisomy 21

Abstract Background Trisomy 21 is a common aneuploid condition in humans and accounts for approximately one quarter of all aneuploid live births. To date, early diagnosis of Trisomy 21 remains a challenging task. Metabolomics may prove an innovative tool to study the early pathophysiology of Trisomy 21 at a functional level. Methods Ultra‐performance liquid chromatography coupled with mass spectrometer (UPLC‐MS) was used for untargeted metabolomic analysis of amniotic fluid samples from women having normal and trisomy 21 fetuses. Results Many significantly changed metabolites were identified between amniotic fluid samples from Trisomy 21 pregnancies and normal euploid pregnancies, such as generally lower levels of several steroid hormones and their derivatives, higher levels of glutathione catabolites coupled with lower levels of gamma‐glutamyl amino acids, and increased levels of phospholipid catabolites, sugars, and dicarboxylic acids. The identification of a human milk oligosaccharide in amniotic fluid may worth further investigation, since confirmation of this observation may have significant implications for regulation of fetal development. Conclusions The metabolisms in amniotic fluid from Trisomy 21 and normal pregnancies are quite different, and some of the significantly changed metabolites may be considered as candidates of early diagnostic biomarkers for Trisomy 21.

membranes. 7 Amniotic fluid has very low levels of proteins and enzymes in early pregnancy (<11 weeks), but in mid-to late pregnancy the protein levels in amniotic fluid increase, and the fluid accumulates many metabolites from the fetal metabolism due to free circulation from the fetal respiratory and digestive tracks. 8  Metabolomics studies of amniotic fluid have been reported for GDM of pregnant women, 9 the prediction of preterm labor, 10,11 and deformity. 12,13 Amniotic fluid collected in the second trimester of pregnancy was studied for prenatal disorders using NMR spectroscopy to search for metabolite biomarkers of preterm delivery. 9 Ultra-performance liquid chromatography coupled with mass spectrometry (UPLC-MS) was used to investigate the effects of preterm delivery on amniotic fluid metabolisms during the second trimester.
In the preterm delivery group, the decreases of certain amino acids and increase of a hexose (possibly glucose) suggest alterations in placental amino acid fluxes and possibility of hyperglycemia. 13 Proton NMR spectroscopy was performed on human amniotic fluid samples representing pregnancies at different maturation stages and with different fetal-maternal complications. 14 In another study of T21 amniocytes using matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF), six proteins (calumenin, nucleophosmin, elongation factor 1-beta, cathepsin D, platelet-activating factor acetylhydrolase IB subunit beta, and 14-3-3 protein beta/alpha) were found to be significantly accumulated. 15 Alterations in metabolic pathways of porphyrin metabolism, bile acid metabolism, hormone metabolism, and amino acid metabolism were discovered in amniotic fluid from fetuses with Down syndrome using UPLC-TOF-MS. 16 The abovementioned studies have only identified limited numbers of metabolites in amniotic fluid, and how to discover as many amniotic fluid metabolites as possible remains a challenge, especially with T21 amniotic fluid samples which have been less studied. The aim of this study was to explore potential biomarkers of T21 in amniotic fluid of second trimester using an untargeted metabolomics approach and try to understand the underlying metabolic mechanisms of T21 birth defect. Ultra-performance liquid chromatography coupled with mass spectrometry (UPLC-MS) was utilized to identify significantly changed metabolites between T21 and normal amniotic fluid samples.

| Sample preparation for metabolomics study
The amniotic fluid samples were processed using an automated MicroLab STAR ® liquid handling system from Hamilton Company.
Several recovery standards were added prior to the first step in the extraction process to evaluate the efficiency of sample extraction.

| uplc-ms analysis
The UPLC-MS analysis utilized a Waters ACQUITY ultra-performance liquid chromatography (UPLC) and Thermo Scientific Q-Exactive high resolution/accurate mass spectrometer which is interfaced with a heated electrospray ionization (HESI-II) source and an orbitrap mass analyzer operated at 35 000 mass resolution. Each sample contains a series of added chemical standards at fixed concentrations to ensure sample injection and chromatographic consistency. For the four aliquots of each extracted sample, the first aliquot is analyzed by liquid chromatography optimized for more hydrophilic compounds under positive ionization. In this method, the sample was gradient-eluted from a C18 reverse-phase column (Waters UPLC BEH C18-2.1 × 100 mm, 1.7 µm) using water/methanol mobile solutions containing 0.05% perfluoropentanoic acid (PFPA) and 0.1% formic acid (FA). The second aliquot was also analyzed under positive ionization, but the liquid chromatography is optimized for more hydrophobic compounds. In this method, the extract is gradienteluted from the aforementioned C18 reverse-phase column using methanol/acetonitrile/water mobile solutions containing 0.05% PFPA and 0.01% FA. The third aliquot is analyzed under negative ionization mode. Each sample is gradient-eluted from a C18 column using methanol/water mobile solutions containing 6.5 mmol/L ammonium bicarbonate at pH 8. The fourth aliquot is analyzed under negative ionization mode after the injection is gradient-eluted from a HILIC column (Waters UPLC BEH Amide 2.1 × 150 mm, 1.7 µm) using mobile solutions containing water/acetonitrile with 10 mmol/L ammonium formate at pH 10.8. The MS analysis in the QE orbitrap alternates between MS and data-dependent MS n scans using dynamic exclusion.

| Data pretreatment and chemometric analysis
After evaluation and passing the quality control (QC) inspection, the raw data are extracted and peak-identified using the proprietary IT For univariate statistical analysis, missing values were filled with the minimum observed value for each compound, and the data were transformed to the natural log for Welch's t test to identify metabolites that differed significantly between experimental groups. Multidimensional analyses including PCA, HCA, and Random Forest were performed using the R package.

| Patients characteristics
A total of 21 cases of fetal T21 and 21 controls were recruited in our study. The average maternal age of both the cases and controls is 34 years old; the average gestational age is 132 days for cases and 133 days for controls, respectively. The fetal genders were 11 males and 10 females in both groups (Table1).

| Metabolomics study results
A total of 621 compounds of known identity were discovered from the amniotic fluid samples through the UPLC-MS/MS metabolomics platform. Welch's two-way t tests were performed to identify metabolites that differed significantly between the case and control groups, and 151 metabolites were found to be significantly different between the two groups. Among the 151 metabolites, 73 are upregulated and 78 downregulated in the T21 case group (Table 2).

| Hierarchical cluster analysis
Broad individual sample variation in the data is displayed in the hierarchical cluster (HC) heat map plot (Figure 2A). Despite this large interindividual biological variance, which is very common in human biological samples, the HC plot is a valuable method for identification of compounds having significant correlations within the dataset. For example, many compounds common to particular pathways, including steroid hormones and a group of gamma-glutamyl derivatives of amino acids, are clustered together ( Figure 2B).

| Random forest analysis
Random Forest (RF) analysis is a supervised learning algorithm testing how accurately the obtained metabolite data can assign each sample into the proper experimental group. As shown in

| Significantly changed metabolites in T21 amniotic fluid
Comparing to their levels in the control group, 13 of the 20 identified steroids hormones are significantly changed in T21 amniotic fluid with 11 decreasing and 2 increasing (Table 3, Figures 4 and 5).
The decreased steroids hormones in T21 amniotic fluid include pregnenolones, corticosteroids, androgens, and estrogens. The progestin class compound 5alpha-pregnan-3beta 20alpha-diol monosulfate is significantly increased in T21 amniotic fluid (P = .054). increasing. Similarly, two hormones have been reported to be significantly altered in T21 amniotic fluid. 16 In humans, analysis of total urinary steroids was effective in detecting fetuses with Smith-Lemli-Opotz syndrome, 18 and progesterone levels were altered in 87% of maternal urine in the presence of a DS fetus. 19 Increased levels of cortisol and decreased levels of pregnenolone sulfate were also detected in DS fetuses, which may be associated with abnormal bone and brain development. 19 Steroid hormones are important in a variety of physiological processes, including development,  Because hormone levels fluctuate strongly during development as well as affected by maternal age, 24 we cannot rule out the possibility that these observed changes of steroid hormones in T21 reflect the kinetics of fetal maturation process.

| D ISCUSS I ON
Free amino acids levels were mostly lower in T21 amniotic fluid relative to the controls. In a previous study reported by Huang et al,16 the authors found that the T21 amniotic fluid contains less arginine, histidine, and glutamate than the regular amniotic fluids. These findings are also confirmed in our study. Other than these 3 amino acids, more are found to be decreased in our study, including glycine, serine, aspartate, asparagine, phenylalanine, tyrosine, tryptophan, leu- Several short-and mid-chain dicarboxylic acids are found to be significantly higher in the case samples. These compounds are formed by omega oxidation of short-and mid-chain fatty acids. They have been found to be elevated in amniotic fluid of pregnancies in which fetuses are affected by glutaric aciduria type II, a genetic disorder that can lead to brain malformations, enlarged liver, and other developmental problems. 28 Glutaric aciduria type II can result from deficiencies in one of three proteins, the alpha (ETFA) and beta (ETFB) subunits of electron transfer flavoprotein, and electron transfer flavoprotein dehydrogenase (ETFDH). 29 It is not known if Trisomy 21 may affect the expression of one or more of these functionalities.
The T21 amniotic fluid samples contain significantly higher levels of N1,N12-diacetylspermine comparing to its level in the normal pregnancy amniotic fluid. N1,N12-Diacetylspermine is a polyamine commonly found in human urine. 30 It has been reported that urinary N1,N12-Diacetylspermine can be used as a marker to efficiently detect colorectal and breast cancers at early stages. 31  known to promote gut maturation and immunomodulary functions in newborns, 33 but whether they have any function at the fetal development is not known. At present, we could not find any study that has reported the presence of HMOs in amniotic fluid.
Other interesting discoveries include a large increase (2.5-fold) of 2-piperidinone, a delta-lactam compound, in the T21 amniotic fluid. This compound is typically a xenobiotic, possibly of maternal dietary origin, but it can also be made through enzymatic oxidation and cyclization of cadaverine in mouse liver, 34 suggesting possible perturbations of fetal liver functions.

| CON CLUS IONS
In summary, the current metabolomics study identifies the largest discovering (but discovering more) more changes in amino acids and hormones that have been reported previously, 16 we have also discovered new changes in polyamine metabolites, sugar metabolites, and metabolites related to liver damage.

ACK N OWLED G M ENTS
The authors acknowledge the efforts and contributions of all the participating volunteers.