Study population
Singleton pregnant women between 20-36 weeks’ gestation were recruited from August 2017 to February 2018 at Beijing Obstetrics and Gynecology Hospital. Women who met the diagnostic criteria of PE were admitted as cases. Pregnant women without gestational mellitus, cardiovascular disease, hypertension, renal disease, autoimmune disease, metabolic disorders, previous history of PE or fetal growth restriction (FGR) were enrolled as controls. The PE diagnosis was determined using the diagnostic criteria proposed by the International Society for the Study of Hypertension in Pregnancy (ISSHP), which defines PE as gestational hypertension (systolic/diastolic blood pressure ≥140/90 mmHg) after 20 weeks’ gestation in previously normotensive women plus new onset of one or more of the following complications including proteinuria, renal insufficiency, liver disease, neurological problems hematological disturbances and FGR [13]. The study protocol (No.2017-KY-070-01) was approved by the Ethics Committee of Beijing Obstetrics and Gynecology Hospital on 18 July 2017 and was available in Supplementary Protocol. The informed consents were obtained from all the participants.
Sample preparation for metabolomics
The maternal blood samples (3 ml) were drawn from all PE patients and healthy controls by venipuncture, left to clot for 30 min, and centrifuged for 10 min at 3500 rpm. The serum aliquots (1 ml) were separated and stored at -80℃. The placenta tissue samples (100 mg) were collected at a maximum depth of 5 mm from the maternal central side of placenta (near the cord insertion) immediately after delivery and kept frozen at -80℃ [14].
For sample processing, 100 µl of each serum sample was mixed with 200 µl of acetonitrile: methanol (1: 1) solution and vortexed for 30 sec, followed by 10 min ultrasonication. Then the mixture was centrifuged at 12,000 rpm for 15 min at 4 ℃ and the supernatant was transferred into glass sample vials with screwed caps and stored at -80℃ until metabolomics analysis. As for the placenta, approximately 50 mg of placenta tissue from each patient was homogenized in 1 ml of cold mass spectrometry grade water with a plastic pestle. Then 200 µl of homogenate was mixed with 800 µl acetonitrile: methanol (1:1) solution, followed by vortex and ultrasonication. The mixture was subsequently centrifuged at 12,000 rpm for 15 min at 4 ℃. The supernatant was dried in vacuum, reconstituted with 100 µl of acetonitrile: water (1:1) solution and stored at -80℃ until metabolomics analysis.
LC-QTOF/MS analysis
In our study, the serum and placenta metabolic fingerprinting was acquired with the AB SCIEX Triple TOF 5600 mass spectrometry (MS) system. The Acquity UPLC HSS T3 C18 column (2.1 mm x 100 mm, 1.8 µm, Waters, Milford, MA) was used in the sample separation step with column temperature maintained at 40℃. The mobile phase consisted of ultrapure water with 0.1% v/v formic acid (phase A) and acetonitrile with 0.1% v/v formic acid (phase B). The following elution gradient program was applied in the liquid chromatography: 5% B for 0-1 min; 5-95% B for 1-14 min; 95% B for 14-17 min; re-equilibration for 3 min. The sample injection volume was 5 µl and the flow rate was 0.3 ml/min.
The MS analysis was performed in both positive and negative ion modes and the conditions of ion source and gas for ionization were as follows: ion source voltage, 5500V (4500V in negative ion mode) ; gas temperature, 550 ℃; curtain gas, 35 psi; gas1 (nebulizing gas), 50 psi; gas2 (heater gas), 55 psi. The declustering potential was set at 80 V on the orifice and the collisional energy was set between 20-50 V. The scan range was 100-1000 m/z in MS1 and 50-1000 m/z in MS2 respectively.
RNA isolation and quantification by RT-qPCR
With the metabolomics analyses in our study, 8 genes involved in GSH metabolism were chosen to test their expression levels in placenta. The real-time quantitative reverse transcription polymerase chain reaction (RT-qPCR) was performed and optimized as described previously [15]. Briefly, 20 tissues samples from PE (n=10) and normal pregnancies (n=10) were placed in liquid nitrogen and ground thoroughly with a mortar and pestle. The total RNA samples were isolated using TRIzol Reagent (Lot: 15596026, Thermo fisher, Carlsbad, CA) according to the manufacturer’s instructions. The RNAs were dissolved in diethylpyrocarbonate-treated water and reversely transcribed by the SuperScript III First Strand Synthesis Super Mix Kit (Lot: 18080051, Thermo fisher, Carlsbad, CA). The cDNA was quantified with quantitative reverse transcription PCR (RT-qPCR) using the Luna Universal qPCR Master Mix (M3003L, NEB). The relative quantification of the PCR products was performed after normalization against the Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) expression, using the comparative cycle threshold method. The qPCR primers sequences of glutathione cysteine ligase catalytic subunit (GCLC), glutathione cysteine ligase modulate subunit (GCLM), glutathione synthase (GSS), glutathione reductase (GSR), glutathione peroxidase-1 (GPx-1), GPx-4, cyclooxygenase-1 (COX-1) and COX-2 were provided in Supplementary Table S1. RT-qPCR reactions were performed on 96-well plates and run in the CFX 96 system (Bio-Rad Laboratories Inc), the relative expression was analyzed using Bio-Rad CFX Manager Software.
Data processing
The raw data of metabolic was collected and analyzed with the MassLynx software (Waters, Milford, MA). All the differential metabolites were identified by the in-house library with the aid of the reference standards and the open database of metabolic reaction network (MRN)-based recursive algorithm (MetDNA) [16]. The multivariate pattern recognition analyses such as principal component analysis (PCA) and orthogonal partial least squares-discriminant analysis (OPLS-DA) were performed in this study with the SMICA 14.1 software (Umetrics, Umea, Sweden). PCA provided a general overview of metabolic concentrations in biological samples. OPLS-DA was used to obtain the value of variable importance in the projection (VIP) of each metabolite. A permutation test was carried out to avoid OPLS-DA model over-fitting [17]. In the univariate statistical analyses, student’s t test was performed to calculate the p values and the corresponding fold change showed how the identified metabolites of diseased individuals varied from that of the healthy controls. With the aid of VIP ≥1 in OPLS-DA model and p <0.05 in student’s t test, the differential metabolites were eventually determined.
The clinical performance of these potential metabolic biomarkers in PE was further assessed by receiver operating characteristic (ROC) curves with MedCalc v11.4.2 (MedCalc Software, Ostend, Belgium) [18]. In addition, the metabolic pathway analysis of differential metabolites in serum and placenta was conducted by the web–based MetaboAnalyst 4.0 (https://www.metaboanalyst.ca) [19]. The gene expression of the enzymes involved in GSH metabolism and cyclooxygenase enzymes were analyzed by student’s t test and p <0.05 was considered statistically significant.