Clinical sample collection

Serum samples from pregnant women were collected from the biological specimen bank of the Shengjing Birth Cohort in Key Laboratory of Health Ministry for Congenital Malformation, Shenyang, China. Nineteen pregnant women were diagnosed as carrying fetuses with NTDs using prenatal ultrasound, and the diagnosis was confirmed by autopsy after induced labor or by physical examination after delivery. Serum samples from gestational age- and maternal age-matched controls were obtained from pregnant women (n = 19) carrying normal fetuses without any abnormalities. Every pregnant woman consumed folic acid regularly under guidance during pregnancy. The clinical characteristics of the patients are summarized in Online Resource 1.

Animal sample collection

SBA was induced in Wistar rats with all-trans-retinoic acid (atRA; Sigma, St. Louis, MO, USA; 4% [w/v] in olive oil; 140 mg/kg body weight) at E10 (vaginal smear containing sperm designated E0) by gavage, and normal controls were treated with the same volume of oil, as described previously [24, 26]. Briefly, animals were anesthetized with isoflurane, and blood was collected from the apex cordis of living rats; euthanasia was then carried out by CO2 asphyxiation. All experimental protocols involving animals were approved by the Medical Ethics Committee of the Shengjing Hospital of China Medical University (2016PS106K).

Fetal deformities were examined using stereomicroscopy (M165 FC; Leica, Mannheim, Germany; Online Resource 4. Fig. 1). Blood samples were collected into vacuum tubes (Vacutainer SST; Becton, Dickinson and Company, Franklin Lake, NJ, USA) and centrifuged (Sorvall ST8R Centrifuge; Thermo Fisher Scientific, Walther, MA, USA) at 2000 × g and 4 °C for 20 min for serum sampling. Serum samples were collected from normal pregnant rats with normal embryos at E12 (n = 12), E14 (n = 6), E16 (n = 6), and E18 (n = 25) and from pregnant rats with SBA embryos at E12 (n = 12), E14 (n = 6), E16 (n = 6), and E18 (n = 25). We also collected 12 treated E18 samples from normal (n = 3) rats treated with oil, those without SBA (n = 3) from rats treated with 140 mg/kg atRA, those with SBA (n = 3) from rats treated with 140 mg/kg atRA (SBA group 1), and those with SBA (n = 3) from rats treated with 110 mg/kg atRA (SBA group 2; Online Resource 2). The fetal neural tube tissue (E12) or spinal cord tissue (E18) (from the inferior margin of the forelimb bud to the tail bud) was isolated in cold phosphate-buffered saline (PBS). Samples were stored at − 80 °C, except tissues used for immunohistochemistry, which were preserved in 4% paraformaldehyde and then embedded in paraffin.

Fig. 1

Bioinformatics and comparative analyses in serum exosomes of pregnant rats at E18. a Bar chart of gene ontology (GO) biological process, cellular component, and molecular function categories. b Protein network of ACTR2, CORO1A, and DNM2. c Western blot analysis of ACTR2, CORO1A, and DNM2 expressions in serum exosomes at E18 among normal group (treated with oil, n = 3), without SBA group (treated with 140 mg/mL atRA, n = 3), SBA group 1 (treated with 140 mg/mL atRA, n = 3), and SBA group 2 (treated with 110 mg/mL atRA, n = 3). d Bar chart of the relative expression of ACTR2, CORO1A, and DNM2 among the four groups in serum exosomes at E18, as determined by western blotting

Total serum exosome extraction and characterization

A 200-μL aliquot of each serum sample was adjusted to 3 mL with phosphate-buffered saline (PBS) and utilized for extraction. The diluted serum was filtered through a 0.22-μm filter (MILLEX GP, Millipore Express PES Membrane; Millipore, Billerica, MA, USA) [27]. The filtered sample was centrifuged (Micro Ultracentrifuge CS120FNX; Hitachi Koki, Tokyo, Japan) at 10,000 × g and 4 °C for 1 h. Subsequently, the supernatant was transferred to a fresh tube and centrifuged at 100,000 × g and 4 °C for 4 h. The pellets were washed with PBS, ultracentrifuged at 100,000 × g and 4 °C for 1 h again to purify the exosomes [28]. The pellet was resuspended in 100 µL PBS and preserved at − 80 °C as previously described [29, 30]. Characterization of the extracted exosomes was confirmed by transmission electron microscopy (HT7800; Hitachi Koki, Tokyo, Japan), dynamic light scattering (Nano ZS90; Malvern Instruments, UK), and exosomal biomarkers (Alix, CD63, and CD9; Online Resource 4. Fig. 2).

Fig. 2

Analyses of ACTR2, CORO1A, and DNM2 expressions in serum exosomes and serum without isolation of exosomes of pregnant rats. a Western blot analysis of protein expression in serum exosomes at E18 (complete bands are shown in Online Resource 4. Fig. 3); bar chart of relative protein expression at E18 in the normal (n = 19) and SBA (n = 19) groups. b Western blot analysis of protein expression in serum exosomes at E16 (complete bands are shown in Online Resource 4. Fig. 4); bar chart of relative protein expression at E16 in the normal (n = 6) and SBA (n = 6) groups. c Western blot analysis of protein expression in serum exosomes at E14 (complete bands are shown in Online Resource 4. Fig. 5); bar chart of relative protein expression at E14 in the normal (n = 6) and SBA (n = 6) groups. d Western blot analysis of protein expression in serum exosomes at E12 (complete bands are shown in Online Resource 4. Fig. 6); bar chart of relative protein expression at E12 in the normal (n = 12) and SBA (n = 12) groups. e Western blot analysis of protein expression in serum without isolation of exosomes at E18 (complete bands are shown in Online Resource 4. Fig. 7); bar chart of relative protein expression in serum without isolation of exosomes at E18 in the normal (n = 6) and SBA (n = 6) groups. f Western blot analysis of protein expression in serum without isolation of exosomes at E12 (complete bands are shown in Online Resource 4. Fig. 8); bar chart of relative protein expression in serum without isolation of exosomes at E12 in the normal (n = 6) and SBA (n = 6) groups

FNE isolation

Exosomes from fetal neural sources were isolated as previously described [21, 22]. Briefly, 400 µL sample was incubated for 90 min at 20 °C with 50 µL of 3% bovine serum albumin (BSA; Solarbio, Beijing, China) containing 2 µg polyclonal goat IgG anti-rat Contactin-2/TAG1 antibody (AF4439; R&D Systems, Minneapolis, MIN, USA) that had been biotinylated (EZ-Link sulfo-NHS-biotin System; Thermo Fisher Scientific). Then, 10 µL Streptavidin-Plus UltraLink resin (Pierce; Thermo Fisher Scientific) in 40 µL of 3% BSA was added. After centrifugation, the supernatant was transferred to an Eppendorf tube and stored at − 80 °C.

LC–MS/MS for biomarker screening

Three pairs of E18 samples were screened for potential biomarkers using label-free LC–MS/MS [31, 32]. Samples were first subjected to immunoaffinity depletion of high-abundance serum proteins. Then, 200 μg protein for each sample was processed by filter-aided sample preparation digestion. The peptide content was estimated by determining the ultraviolet spectral density at 280 nm, calculated based on the frequency of tryptophan and tyrosine in vertebrate proteins [32]. Each fraction was then injected for nanoLC-MS/MS analysis. The peptide mixture was separated with a linear gradient controlled by IntelliFlow technology. LC–MS/MS analysis was performed on a Q Exactive mass spectrometer (Thermo Fisher Scientific) coupled to an Easy nLC (Proxeon Biosystems). The MS data were analyzed using MaxQuant software version 1.5.3.17 (Max Planck Institute of Biochemistry, Martinsried, Germany) [33].

Bioinformatics analysis

Bioinformatics analysis was carried out using WebGestalt (http://www.webgestalt.org/) with over-representation analysis (ORA) in Rattus norvegicus for enrichment with the gene ontology (GO) biological process functional database. The significance level in the advanced parameters was adjusted to a false-discovery rate of less than 0.05. The proteins showing enrichment were further analyzed using the String database (https://string-db.org/) to define protein interaction networks.

Western blotting

Exosome-derived from serum (E12, E14, E16, E18), serum (E12, E18), neural tube tissue (E12), and spinal cord tissue (E18) were treated with 20 µL radioimmunoprecipitation assay buffer (Solarbio) per 100 µL sample and subjected to ultrasound pyrolysis. The supernatants were collected as the protein-containing fraction. Serum (E12, E18) was diluted by PBS before western blotting. Samples were quantified using a BCA Protein Assay Kit (Solarbio), and protein concentrations were adjusted using PBS. The diluted samples were mixed with loading buffer, denatured, separated on Bio-Rad gels, and transferred to 0.45-μm polyvinylidene difluoride membranes. The primary antibodies used were as follows: anti-Alix (3A9 mouse mAb; Cell Signaling Technology, Danvers, MA, USA; exosome internal reference), anti-CD9 (D3H4P rabbit mAb; Cell Signaling Technology), anti-CD63 (TS63 mouse mAb; Abcam, Cambridge, UK), anti-β-actin (66009–1-Ig rabbit mAb; Proteintech, Wuhan, China; neural tube internal reference), anti-actin-related protein 2 (ACTR2; 10922–1-AP rabbit pAb; Proteintech), anti-coronin 1A (CORO1A; EPR19467-36 rabbit mAb; Abcam), and anti-dynamin 2 (DNM2; EPR9053 rabbit mAb; Abcam). The secondary antibodies used were goat anti-mouse (G-21040; Invitrogen, Carlsbad, CA, USA) or goat anti-rabbit (G-21234; Invitrogen). Fast-staining Coomassie Brilliant Blue (Solarbio) served as an internal reference for serum. Bands were visualized using chemiluminescent horseradish peroxidase-based substrate (Immobilon Western; Millipore) and captured (cSeries 300; Azure Biosystems, Dublin, CA, USA). Representative samples are shown, and raw data for the bands are shown in the Online Resource 4.

Immunohistochemistry

An UltraSensitive SP Kit (MXB Biotechnologies, Fuzhou, China) was used for immunohistochemistry, according to the manufacturer’s protocol. The primary antibodies were the same as used for western blotting. Sections were washed in PBS between each process and finally stained with diaminobenzidine and hematoxylin, dehydrated, and sealed with resin. Images were captured using a microscope (ECLIPSE 80i; Nikon, Tokyo, Japan).

ELISA

Serum (50 µL) was diluted with PBS (50 µL), and samples were subjected to ELISA using a human DNM2 ELISA kit (abx250599; 96-well; Abbexa) and human CORO1A ELISA kit ab214032; 96-well; Abcam), according to the manufacturer’s protocol. Absorbance was measured at 450 nm using a microplate reader (M200 PRO; Tecan, Switzerland). The relative optical density at 450 nm (OD450; OD450 of each well–OD450 of the blank well) was calculated.

Statistical analysis

Data are expressed as means ± standard deviations. Western blotting bands were quantified using ImageJ (1.37c) as the ratio of the gray value of the biomarker protein/the gray value of the internal reference protein. Quantitative variables were analyzed using unpaired t-tests, paired t-tests, and one-way analysis of variance (ANOVA). Differences with P values less than 0.05 were considered significant. Statistical analyses were performed using GraphPad Prism 8.0 software. The diagnostic capacity of biomarkers was analyzed using receiver operating characteristic (ROC) curves, and the area under curve (AUC), specificity, and sensitivity were determined using MedCalc 19.3.1 in Statistics-ROC.