All patients received written consent before surgery, and all procedures and processes were approved by the Ethics Committee of the First Affiliated Hospital of Anhui Medical University (Hefei, China)(authorization number: S20200021).

Samples were collected in the outpatient gynecological operating room of the First Affiliated Hospital of Anhui Medical University. The enrollment criteria were as follows: miscarriage in women was defined as a couple having one or more consecutive pregnancy losses before 13 gestational weeks. Elective terminations of pregnancies (ETPs) were confirmed by ultrasound, intrauterine pregnancy with embryo and cardiac tube pulsation. On the basis of 2017 Guidelines of the American Thyroid Association for the diagnosis and management of thyroid disease during pregnancy and postpartum, all the group patients’ TSH levels in the peripheral blood were less than 4 μIU/ml. Patients diagnosed with hypothyroidism, hyperthyroidism or other endocrine disorders were excluded. The mean age of each group was as follows: miscarriage 27.1 ± 6.1 years and ETP 26.3 ± 7.5 years (P > 0.05). The ETP was matched with miscarriage in gestational day (53.6 ± 10.2 days VS 57.1 ± 8.5 days). In cases of miscarriage, surgery was performed within the first 24 h after diagnosis. All women selected in this research had no medical or family history. Chromosomal abnormalities in villi were detected immediately after the operation.

The villi were collected immediately after uterine aspiration and curettage. After the sorting of villi, the villi were divided into three parts and washed separately with saline several times to avoid the contamination of blood or others. One part of the tissues was used for immunohistochemistry and immunofluorescence after fixed in 4% paraformaldehyde. Approximately 30 mg of the samples was immediately fitted into epoxy pipes containing l ml RNA Later solution (Invitrogen, United States) for mRNA detection and the rest was placed in pipes and quick-freezed in liquid nitrogen for Western blot analysis or other experiments.

JEG-3 cells were used to examine the effect of iopanoic acid (IOP), a specific activity inhibitor of Dio2, on the expression of PLGF and sFlt1. The cells were incubated in DMEM medium supplemented with 10% fetal bovine serum, 100U/ml of penicillin and 100 μg/ml streptomycin at 37℃ in 5% CO2-95% air. The cells then were treated with IOP (100 μM) for 24 h. The cells were then harvested for immunoblotting and RT-PCR to detect the expression of Dio2, PLGF and sFlt1.

The RNA Prep Pure Tissue Kit (TIANGEN, China) was used according to the manufacturer’s protocol for total RNA extraction in placental villi. After added isopropanol for precipitating of RNA and washed the RNA in pre-cooled 75% ethanol for three times. NanoDrop 2000 was used to detected the concentration of RNA the agarose gel electrophoresis was performed to detect integrity of RNA with Gel Red staining. A Prime Script RT reagent kit (Takara, Japan) was used for reverse-transcribe 500 ng of total RNA into cDNA in 10 μl of the reaction system. The Lightcycler480 Real-time software (Roche, Japan) was used to perform real-time quantitative PCR in a 20 μl of reaction system. 18 S RNA was used as the reference. The sequences of the gene-specific primers used are listed in Table 1. Relative mRNA expression levels were determined by the 2−ΔΔCt method.

Placental villi sections (3 μm) were dewaxed and rehydrated, and then subjected to sodium citrate buffer for epitope retrieval. After blocked with 3% H2O2 at room temperature for endogenous peroxidase activity, goat serum was used to eliminate the nonspecific binding of the primary antibodies (the antibodies were listed in Table 2). Sections were then incubated with specific primary antibodies overnight. Reactivity was detected by using the MAXIN Ultrasensitive SP kit (MAXIN KIT-9710, China) and counterstained with hematoxylin. The slides were assessed by two observers blinded to grouping. Five nonconsecutive fields of each slide were taken by Zeiss system at a magnification of ×400.

The proliferative activity of trophoblasts was determined by immunohistochemistry staining for Ki67, and apoptosis was determined by TUNEL assay. For semiquantitative analysis, the rate of positive cells in villi of per field was counted. CD34 was used as the hallmark of vascular endothelium to evaluate placental angiogenesis, and the number of CD34+ blood vessels per field was counted. For TTR, VEGF, THRα and THRβ, the percentage of positive cells per field was counted.

The placental villi Sect (3 μm) were dewaxed with xylene and rehydrated in gradient ethanol (from 100 to 50%) and then immersed in 0.01% Triton-X 100 solution. After blocking with 1% BSA, the rehydrated-sections were treated with specific primary antibody incubation (Dio2, 1:500) at 4℃ for 12 h. After washing 3 times with PBS, the secondary antibody was added and incubated for 1.5 h. Nuclear DNA was stained with DAPI. All the images were visualized on a full-field digital slice scanner (Pannoramic MIDI, Hungary).

Cytoplasmic protein of placenta villi and cells was isolated and the detailed method has been reported in previous study [16]. Steps were as follows: Placental villi and cells were homogenized in lysis buffer and homogenates were then centrifuged at 15,000 g for 15 min. Supernatants from each sample were added to a gel loading buffer and boiled for 10 min. After electrophoresis in 12.5% SDS-polyacrylamide gel, The gel was transferred electrophoretically onto a polyvinylidene fluoride membrane. The PVDF membranes were blocked by skim milk for 2 h at room temperature, and then incubated with primary antibodies (as listed in Table 2) for 3 h. After washes in DPBS containing 0.05% Tween-20 four times for 10 min each and PBS for 10 min once, the membranes were incubated with for 1.5 h at room temperature. The membranes were then washed four times in DPBS containing 0.05% Tween-20 for 10 min each and PBS for 10 min once, followed by signal development using an enhanced chemiluminescence detection kit.

All statistical analysis was performed with SPSS 23.0 software. All quantified data were presented as means ± standard deviation (SD) after determination if samples were normally distributed using 1-sample K-S test. The independent two sample t-test was performed to compare the difference if samples were normally distributed. If samples were non-normal distribution, data were present as quartile range, the difference was compared by Mann-Whitney U test. All experiments were performed at least three times. Difference was considered to be significant for P < 0.05.