#Authors contributed equally
This work aimed to clarify the interaction between the fetus and pregnant patients with gestational diabetes mellitus (GDM), the lipid metabolomics analysis of the fetal umbilical cord blood of GDM patients and normal pregnant women were performed to screen out the specific lipid metabolites for pathogenesis of GDM. From 2019–2020, 21 patients with GDM and 22 normal pregnant women were enrolled in Hexian Memorial Hospital, Panyu District, Guangzhou. The general information such as weight, height, age, body mass index (BMI) before pregnancy were analyzed. Non-targeted metabonomic detection and analysis were performed in umbilical cord plasma using LC-MS method. The age, BMI, delivery methods, and infant weight were different between GDM and control. There were 167 lipid metabolites in umbilical cord blood associated with GDM. Among them, 158 upregulated and 9 downregulated in GDM. There were 13 dysregulated metabolites with C < 30, including Lyso-phosphatidyl-colines LPC 16:0, 18:2, 18:1, 18:0, 20:4 and 22:6, glycerophosphocholines PC O-16:1, oleoylcarnitine CAR 18:2 and 18:1, dihexosylceramides Hex2Cer 13:0;2O, phosphatidylethanolamine PE O-22:6_2:0 and PE O-22:6_3:0 and sphingomyelin SM 8:0; 2O/11:0. Those metabolites were associated with glycerophospholipid metabolism and sphingolipid metabolism. Therefore, Lyso-phosphatidyl-colines, glycerophosphocholines, oleoylcarnitine, dihexosylceramides, phosphatidylethanolamine, and sphingomyelin were main lipid metabolites of GDM, which might be used for diagnosis and treatment of GDM.
Gestational diabetes mellitus (GDM) is a type of diabetes that occurs or is discovered during pregnancy in women without a history of diabetes. It is a special type of diabetes mellitus (
There is increasing evidence that the fetus is not only passively affected by pregnancy diseases but may also play an active role. It has been found that phosphatidylcholinyl C32:1 and proline metabolites in umbilical cord blood of infants have potential effects on maternal hypertension in pregnancy (
Metabonomic is a powerful tool for elucidating the mechanism of metabolic abnormalities in susceptible individuals by systematically studying low molecular weight compounds in biological tissues and body fluids (
Metabonomic markers have been used to help predict the risk of postpartum diabetes in women with GDM, and diabetes related lipid metabolism disorders have also been widely concerned as a major health burden worldwide (
In this study, by comparing the lipid metabolomics analysis of the umbilical cord blood of the fetus delivered by GDM patients and normal pregnancy women with delivery fetus using the high-throughput metabonomic method, the effect of GDM on lipid metabolism were evaluated, and the specific lipid metabolites that can characterize GDM were screened out.
Umbilical cord blood samples were collected from Hexian Memorial Hospital, Panyu District, Guangzhou. Pregnant women were recruited in Guangzhou from 2019 to 2020. All the subjects signed the written consent, and this study was approved by the ethics committee of Hexian Memorial Hospital.
All pregnant women included in the study were given a card at the Perinatal Medicine Clinic of Hexian Memorial Hospital from 2019 to 2020 and followed up regularly to collect basic patient information, such as height, weight, age, pre-pregnancy body mass index (BMI). According to the results of oral glucose tolerance test (OGTT) and delivery conditions of 75 g at 24–28 weeks of gestation, we selected those who gave birth at full term. The 75G OGCT test was performed according to the standards of the American Diabetes Association. Before the test, the patient was fasted for eight hours, and taken 75 g of glucose within 5 min, and then test the blood glucose at 1 h and 2 h, respectively. Any patient with blood glucose level that meets or exceeds the following criteria was diagnosed as GDM: Fasting: 5.1 mmol/L; 1 h after meal: 10.0 mmol/L; 2 h after meal: 8.5 mmol/L (
The inclusion criteria of the study subjects are: (1) Establishing a maternity check-up health card before the 13th week of pregnancy; (2) 18–40 years old. Exclusion criteria: (1) multiple pregnancy; (2) smoking or drinking; (3) history of hypertension and family history of diabetes; (4) abnormal blood glucose in the first trimester; (5) delivery within 37 weeks of pregnancy; (6) received clinical treatment of GDM. Finally, 21 GDM patients were selected, and 22 healthy pregnant women matched with age of GDM patients were selected as control group for further study. Umbilical cord blood was collected and serum were used for metabonomic analysis within one minute.
The collected umbilical cord blood samples were centrifuged at 3000 rpm at 4°C for 10 min on the same day. All umbilical cord blood was centrifuged to obtain serum, which was immediately transferred to a centrifuge tube and stored at −80°C for testing. The ultra-high-pressure liquid-phase high-resolution mass spectrometer Agilent 6545A QTOF mass spectrometer (Agilent Technologies, Santa Clara, USA) was used for the test. The samples were centrifuged by methanol and MTBE vortex, and then ultrasonic processing. The organic phase and water phase were centrifuged, and then the organic phase and quality control (QC) sample were prepared. The mobile phase column temperature was controlled at 35°C, and the sample volume was 5 μL. Water: acetonitrile: formic acid (4:6), and acetonitrile: isopropanol (1:9) were used for positive and negative ions.
Agilent 6545a QTOF mass spectrometer is controlled by the control software (LC/MS data acquisition, version b.08.00) with auto MS/MS mode. The primary and secondary mass spectrometry data were collected, and the quality scanning range is m/Z (50–1100): (1) Chromatographic optimization for more hydrophilic compounds, using ultrapure water (A) containing 0.05% perfluorinated formic acid, 0.1% formic acid and methanol solution (B) on a C18 column for elution. (2) Chromatographic optimization for more hydrophobic compounds. This method uses ultrapure water (A) containing 0.05% perfluorinated formic acid, 0.01% formic acid and methanol/acetonitrile (B) on a C18 column. Remove, and operate under the condition of higher total organic content. (3) Use water (A) containing 6.5 mmol/L ammonium bicarbonate and 95% methanol (B) to elute on a C18 column. (4) Component 4 uses a mobile phase consisting of 10 mmol/L ammonium formate in water and acetonitrile to perform gradient elution on a hydrophilic interaction HILIC column. The positive and negative ion modes were used to collect respectively. The parameters of ESI ion source are set as follows: Ion source dry gas temperature (Gas Temp): 320°C, nitrogen flow (Gas Flow): 8 L/min, sheath gas flow rate (SheathGasFlow): 12 L/min, Sheath gas temperature (SheathGasTemp): 350°C; capillary voltage (VCap): 3500 V (negative ion mode), and 4000 V (positive ion mode).
The offline data is first converted to .abf format using Analysis Base File Converter, and then MSDIAL software (version 4.24) is used to perform data processing such as peak search and peak alignment on the converted abf file, and the identification results were obtained by searching the lipid blast database based on the primary and secondary maps. For the data identified by MSDIAL alignment, QC samples are used to control the quality of the test. The metabolites with more than 50% missing values in the original data were eliminated, and the sample index CV < 30% was controlled by QC samples, and the auto scaling method was used for normalization.
The univariate statistical analysis of metabolites was performed by Fold change analysis and T-test. The differential metabolite was screened by criteria:
Multivariate statistical analysis was performed. In this experiment, the PCA with unsupervised statistical model was used to analyze of GDM and normal samples.
PLS-DA models of GDM and normal groups were established, and the evaluation parameters of the model were obtained through interactive verification. The R2 and Q2 > 0.5 indicate a good prediction effect. The hierarchical clustering of samples in each group was carried out with the expression of qualitative and significant differential metabolites. Enrichment analysis of KEGG pathway was also performed. The analysis was performed by using Metabo analyst 4.0 software.
The parameters recorded between the GDM and normal delivery women are shown in
Features | GDM (N = 21) | Normal (N = 22) | |
---|---|---|---|
Age | 32.09 (26–38) | 29.14 (22–34) | 0.2 |
BMI (kg/m2) | 28.86 (23.43–38.76) | 25.59 (21.15–31.25) | 0.002 |
Parity | 1.61 | 1.63 | 0.840 |
Fasting blood glucose | 5.14 | 4.61 | 0.097 |
Gestational age | 39.2 | 39.1 | 0.756 |
Proportion of male infants | 57% | 27% | 0.05 |
Delivery method-percentage of cesarean section | 90% | 9% | 0.000 |
Infant weight (g) | 3390 (2620–4400) | 3137 (2700–3840) | 0.029 |
Apgar score | 10 | 9.9 | 0.335 |
Principal component analysis (PCA) showed that QC samples were closely clustered, indicating a good repeatability of the experiments, with stable and reliable instrument analysis system (
The distribution of samples in principal components PC1 and PC2 shows that the samples of GDM and normal groups are separated (
The 13 metabolites with carbon chain length (C) <30, are shown in
Metabolites | INCHIKEY | Formula | SMILES |
---|---|---|---|
LPC 16:0 | ASWBNKHCZGQVJV-UHFFFAOYNA-N | C24H50NO7P | CCCCCCCCCCCCCCCC(=O)OCC(O)COP([O-])(=O)OCC[N+](C)(C)C |
LPC 18:2 | SPJFYYJXNPEZDW-UTJQPWESNA-N | C26H50NO7P | CCCCC\C=C/C\C=C/CCCCCCCC(=O)OCC(O)COP([O-])(=O)OCC[N+](C)(C)C |
LPC 18:1 | YAMUFBLWGFFICM-SEYXRHQNNA-N | C26H52NO7P | CCCCCCCC\C=C/CCCCCCCC(=O)OCC(O)COP([O-])(=O)OCC[N+](C)(C)C |
LPC 18:0 | IHNKQIMGVNPMTC-UHFFFAOYNA-N | C26H54NO7P | CCCCCCCCCCCCCCCCCC(=O)OCC(O)COP([O-])(=O)OCC[N+](C)(C)C |
LPC 20:4 | GOMVPVRDBLLHQC-DOFZRALJNA-N | C28H50NO7P | CC\C=C/C\C=C/C\C=C/C\C=C/CCCCCCC(=O)OCC(O)COP([O-])(=O)OCC[N+](C)(C)C |
LPC 22:6 | LSOWKZULVQWMLY-WSDBEMKQNA-N | C30H50NO7P | CC\C=C/C\C=C/C\C=C/C\C=C/C\C=C/C\C=C/CCC(=O)OCC(O)COP([O-])(=O)OCC[N+](C)(C)C |
PC O-16:1 | ZAPMAQAWNNDALV-KTKRTIGZNA-N | C24H48NO7P | CCC\C=C/CCCCCCCCOCC(COP([O-])(=O)OCC[N+](C)(C)C)OC(=O)CC |
CAR 18:2 | HQMPRWWWVKTZAS-HULFFUFUNA-O | C25H46NO4 | CCCCC\C=C\C=C\CCCCCCCCC(=O)OC(CC(O)=O)C[N+](C)(C)C |
CAR 18:1 | HOAMADDCQBUDDY-KHPPLWFENA-O | C25H48NO4 | CCCCCC\C=C/CCCCCCCCCC(=O)OC(CC(O)=O)C[N+](C)(C)C |
Hex2Cer 13:0;2O | ZKJIXTUCXJSUPH-UHFFFAOYNA-N | C25H47NO13 | CCCCCC(O)C(COC1OC(CO)C(OC2OC(CO)C(O)C(O)C2O)C(O)C1O)NC(=O)CCCC |
PE O-22:6_2:0 | XNSCNJIPDQULCK-JDPCYWKWNA-N | C29H48NO7P | CC\C=C/C\C=C/C\C=C/C\C=C/C\C=C/C\C=C/CCCOCC(COP(O)(=O)OCCN)OC(C)=O |
PE O-22:6_3:0 | JBXUVOQOEAOSJK-YNUSHXQLNA-N | C30H50NO7P | CC\C=C/C\C=C/C\C=C/C\C=C/C\C=C/C\C=C/CCCOCC(COP(O)(=O)OCCN)OC(=O)CC |
SM 8:0;2O/11:0 | YFCNXMQPCURICF-UHFFFAOYNA-N | C24H51N2O6P | CCCCCCCCCCC(=O)NC(COP([O-])(=O)OCC[N+](C)(C)C)C(O)CCCCC |
The results of enrichment analysis of KEGG pathway of all the 167 significant changed lipid metabolites showed those differential metabolites were involved in sphingolipid metabolism, and glycophoripid metabolism (
The KEGG pathway analysis of the 13 metabolites with C < 30 showed that LPC participated in the glycerophosphate metabolic pathway (map00564). PC participated in the glycerophosphate metabolic pathway (map00564) and the ether lipid metabolic pathway (map00565). Hex2cer participated in the glycerophosphate metabolic pathway (map00564) and the glycerolipid metabolism pathway (map00561). SM is involved in the glycerophosphate metabolic pathway (map00564), the ether lipid metabolism pathway (map00565), the glycerin metabolic pathway (map00561), the linoleic acid metabolism pathway (map00591), the sphingolipid metabolism pathway (map00600), the arachidonic acid metabolic pathway (map00590) and the α-linolenic acid metabolic pathway (map00592). These metabolites may be potential biomarkers for the identification of GDM, and the metabolites are closely associated with the pathway of glycerophosphate metabolism.
Our study identified 13 lipid metabolites (C < 30) associated with GDM in umbilical cord blood, which can be divided into six categories including lysophosphatidylcolines, glycerophosphocholine, oleoyl carnitine, dihexylceramides, phosphatidylethanolamine, and sphingomyelin. Consistent to our results, it was reported that the metabolic spectrum of lysophosphatidylcholine, sphingomyelin and other lipids in GDM patients have changed (
Based on LC-MS tandem analysis, lysophosphatidylcolines was evaluated as a potential biomarker of cancer (
Glycophorophosphocholine is an ether lipid with a 1-o-alk-1’-alkenyl ether bond at the sn-1 position of the main chain of glycerol. Abnormal levels of ether glycerophosphatidylcholine (ether PCs) have been associated with cell dysfunction and various human diseases (
Acylcarnitine is a kind of metabolites formed by the combination of carnitine and fatty acids, which are similar in structure to each other. They are widely present in various tissues and body fluids. Its biological functions mainly include: (1) transferring long-chain fatty acids from cytoplasm to mitochondrial matrix for β oxidation, (2) promoting the production of peroxidase β oxidation, acetyl-CoA to enter mitochondria for oxidation, (3) assisting the transportation of short-chain and medium-chain fatty acids in mitochondria, (4) being an integral part of lipid metabolism and membrane integrity (
Dihexosylceramides is one of the ceramides. Using non-targeted lipomics based on LC-MS, studied have showed that the characteristics of lipid metabolism disorder in the heart of mice are related to the accumulation of glycerin, phospholipid and ceramide (
Phosphatidylethanolamine is one of the most abundant phospholipids in mammalian plasma membrane, which is second only to lecithin. Study has confirmed that phosphatidylethanolamine decreases in the offspring of GDM progenies (
Sphingomyelin (SM) biosynthesis may affect various important cellular processes, such as cell proliferation, cell survival and migration, even the normal physiology of organisms. The reduction of sphingomyelin and hexosylceramide is related to impaired sphingolipid metabolism, suggesting that endogenous adipogenesis may be the driving factor for the onset of diabetes (
Neonatal umbilical cord blood is directly affected by maternal metabolism, which has a strong correlation with maternal metabolic components. Compared with other biological samples, umbilical cord blood could reflect the fetal metabolism in a more direct and comprehensive manner. In this study, high-throughput, high-sensitivity and high-resolution ultra-performance liquid chromatography combined with mass spectrometry technology was used (
In conclusion, these screened lipid metabolites can be used as potential biomarkers associated with insulin resistance and GDM. There are changes of lipid metabolites in umbilical cord blood during the development of inflammation and insulin resistance. Further research should be done to explore the correlation between lipid metabolism disorders and the progress of GDM.
Supplementary Table 1 All the differential metabolites.