The role and underlying mechanism of FABP4 dysregulation mediated by m6A modification in unexplained recurrent spontaneous abortion.

From 2020 to 2023, we recruited 37 patients undergoing induced abortions (Normal group) and 28 patients with URSA group at Renmin Hospital of Wuhan University. The inclusion criteria were as follows: (1) Normal group: Patients with an ultrasound-confirmed intrauterine pregnancy presenting with a fetal heartbeat, no history of spontaneous abortion, and undergoing pregnancy termination for non-medical reasons. (2) URSA group: Patients with a history of two or more spontaneous abortions, an ultrasound-confirmed intrauterine pregnancy, and no hormone therapy in the preceding three months. The exclusion criteria for all participants were: (1) Uterine abnormalities (e.g., malformations) and tubal abnormalities detected by pelvic examination and ultrasound. (2) Genetic factors, such as chromosomal abnormalities in the embryo or either parent. (3) Diabetes mellitus and hypertension. (4) Autoimmune diseases, including antiphospholipid syndrome, and pre-thrombotic states. (5) Infections and malignancies. The collection and use of villus tissue were approved by the Research and Ethics Committee of Renmin Hospital of Wuhan University (WDRY2020-K218). Following surgical aspiration, chorionic tissue was washed with PBS to remove blood, half was immediately frozen in liquid nitrogen and stored at -80 °C, while the other half was fixed in 4% paraformaldehyde and stored in a light-protected environment. Fabp4 knockout mice were generously provided by Dr. Wang Zhihua, Director of the Central Laboratory at Renmin Hospital of Wuhan University. These animals were bred in SPF environment at the Animal Experimental Center of Renmin Hospital of Wuhan University. Breeding conditions were standardized, maintaining room temperature at 22–25 °C, relative humidity at 60–70%, and a 12-h light/dark cycle. Mice had free access to water and food. Breeding occurred in cages with a 2:1 male-to-female ratio. Female mice births were closely monitored, and the genotypes of pups and parents were recorded. Female mice of homozygous, heterozygous, and wild-type FABP4 genotypes were paired with male mice of corresponding genotypes at a 2:1 ratio. Pregnancy was confirmed by vaginal plug and dated as 0.5 days post-conception. On day 13.5 of pregnancy, mice were sacrificed, and embryonic resorption was observed. Embryos and placentas were collected, weighed, and divided for storage. Part of them was stored at -80 °C, while the remainder was fixed in 4% paraformaldehyde. All animal experiments adhered to the guidelines of the Experimental Animal Welfare Ethics Committee (IACUC) of Renmin Hospital of Wuhan University (WDRM Animal (f) No. 20201207). A total of 40 female and 20 male Wistar rats aged 5–7 weeks were procured from Beijing Weitonglihua Experimental Animal Technology Co., Ltd. All animal experiments were conducted following the strict standards of animal experimentation ethics and approved by the Animal Management and Use Committee of the Animal Experiments Center (Approval No. 20201207). The female and male rats were paired at a 2:1 ratio. On day 7.5 of gestation, blastocyst formation was confirmed by ultrasound, and subsequently, the rats were intraperitoneally injected with LPS (0.25 mg/kg, once every other day) or FB23-2 (2 mg/kg/day). On day 13.5 of gestation, the rats were sacrificed, and embryonic resorption was observed. Serum and placenta samples were collected for further analysis. The detection of LPS, FABP4, and SOD2 in rat serum follows a standardized protocol. Initially, rats are fasted for 5–6 h and anesthetized using a 1.5-2% sodium pentobarbital solution. Once anesthetized, blood (3.5-5 mL) is collected via puncturing the strongest heartbeat point using a disposable yellow blood collection tube with a separator. The tube is inverted several times and allowed to stand at room temperature for 1 h. Subsequently, the blood is centrifuged at 4 °C and 3000 rpm for 15 min, and the supernatant is collected, discarding any serum with visible hemolysis. The serum is then stored at -80 °C for future analysis. For the ELISA assay, the kit and samples are equilibrated at room temperature. Each well is loaded with 100 µL of standard working buffer (appropriately diluted) or sample and incubated at 37 °C for 80 min. Wells are washed three times with 200 µL of wash buffer, and then 100 µL of biotinylated antibody working solution is added to each well, followed by incubation at 37 °C for 50 min. After another three washes, 100 µL of HRP enzyme working solution is added to each well and incubated for another 50 min at 37 °C. Wells are washed five times with 200 µL of wash buffer, and 90 µL of TMB is added to each well, incubated for 20 min at 37 °C. Finally, 50 µL of stop solution is added, and the absorbance is immediately read at 450 nm to derive the results. RNA is extracted from samples using RT-qPCR methods and adjusted to a concentration of 100–300 ng/µL with DEPC water. Subsequently, essential reagents are prepared, including 1× Wash Buffer, capture antibody solution, detection antibody solution, diluted enhancer solution, and a positive control standard curve solution. For RNA binding, 80 µL of binding solution is dispensed into each well, along with 2 µL of negative control, 2 mL of diluted positive control, and 200 ng of sample RNA (1–8 µL). The plate is then covered and incubated at 37 °C for 90 min. After incubation, the BS solution is discarded, and wells are washed thrice with 150 µL of diluted WB solution, followed by removal of the excess solution. Next, the processed RNA samples are mixed with m6A antibodies at a precise ratio and incubated under optimized conditions to facilitate binding to m6A-modified RNA fragments. A colorimetric substrate is then added to the mixture, thoroughly mixed, and incubated in a constant temperature water bath for a specified duration. During this reaction, the m6A antibodies react specifically with the colorimetric substrate, leading to a distinct color change. Finally, the absorbance of each sample is measured using a microplate reader at a designated wavelength. A standard curve is constructed based on the absorbance and concentration values of the standards. The relative content of m6A modification in the samples is accurately determined by calculating the concentration corresponding to each sample’s absorbance value. HTR-8/SVneo cells were seeded in a 96-well plate at 2 × 10 4 cells/well, including two empty control wells. Each well was filled with 80 µL of culture medium and incubated overnight in a CO₂ incubator. Probe plate hydration was performed by adding 200 µL of preheated solution to the underlying wells. The probe plate was then covered and incubated overnight in a non-CO₂ environment. Once the cell density reached 85–90%, the culture medium was replaced with experimental medium and incubated for 1 h at 37 °C in a non-CO₂ incubator. OCR assay solution was prepared by mixing basal medium with glucose, L-glutamine, and sodium pyruvate. Three drug formulations (A, B, C) were subsequently created using varying concentrations of Olig, FCCP, Antimycin, and Rotenone. Drug administration involved carefully loading 25 µL/well of each formulation onto drug plates A, B, and C, corresponding to the probe plate’s well positions. Using appropriate pipetting equipment, the drugs were administered while replacing the loading plate. The plate was then incubated at 37 °C in a non-CO₂ environment, and then completing the machine operation. The spectral acquisitions were conducted using a cutting-edge WITec alpha 300R + Raman confocal microspectrometer (Ulm, Germany), which was comprehensively equipped with a precision piezoelectric stage (UHTS 300, WITec, GmbH.), a high-magnification 100X air objective lens (Zeiss EC EPIPLAN, N.A. = 0.9), a green solid-state excitation laser (λ = 532 nm, 32 mW, WITec, GmbH.), and an advanced imaging spectrograph (Newton, Andor Technology Ltd. UK). The spectrograph featured a 600 groove/mm grating and a thermoelectrically cooled charged-coupled detector (CCD) operating at 60 °C. The CCD was optically coupled to the objective lens via a single-mode silica fiber-optic cable with a diameter of 10 μm. The laser excitation spot size was precisely calibrated to approximately 350 nm, allowing for the acquisition of spectral data spanning a broad wavenumber range from 0 to 3600 cm − 1 . Consistency was paramount in our experimental protocol, as we maintained a constant excitation laser intensity across all sample scans. Furthermore, for tissue mapping, we adhered to a standardized integration time of 0.1 s and a step size of 400 nm, covering an area of 100 μm × 100 μm on each tissue slide. For single spectral acquisitions, we extended the integration time to 2 s to ensure optimal signal-to-noise ratio. This rigorous experimental setup and methodology ensured the accuracy and reproducibility of our spectral data. Paraffin sections underwent dewaxing in Dewaxing Solutions I-III (10 min each), absolute ethanol I-III (5 min each), and distilled water. For antigen retrieval, slides were heated in citrate buffer (pH 6.0) in a microwave oven, alternating between medium heat (8 min), no heat (8 min), and low-medium heat (7 min). After cooling, slides were washed thrice in PBS (5 min each) on a shaking platform. The region of interest was marked with a hydrogen peroxide-resistant pen, and sections were incubated in 3% hydrogen peroxide for 25 min at room temperature in the dark. After washing, 10% BSA was used for serum blocking (30 min). The FABP4 primary antibody was added, and sections were incubated overnight at 4 °C. Following washing, the HRP-labeled secondary antibody was applied and incubated for 50 min at room temperature. Signal amplification was achieved by adding TSA-CY3 to the delineated area and incubating for 10 min in the dark. Slides were washed thrice in TBST. For the second round of antigen retrieval, tissue sections were reheated in the microwave oven using the same protocol. The FTO primary antibody was added, and sections were incubated overnight at 4 °C. After washing, the fluorescent secondary antibody was applied and incubated for 50 min at room temperature in the dark. DAPI counterstaining was done by incubating the stain in the delineated area for 10 min in the dark. After three washes, an autofluorescence quencher was applied for 5 min, rinsed with running water for 10 min, and the slides were mounted with an antifade medium. Images were captured using specific excitation and emission wavelengths: DAPI (330–380 nm/420 nm), 488 nm laser (465–495 nm/515–555 nm), and CY3 (510–560 nm/590 nm). To assess trophoblastic cell invasion through Matrigel and migration without coating, Transwell chambers (Corning, Cambridge, MA) with 8.0 μm pores, pre-coated with Matrigel, were employed. For coating, 100 µL of a 1:10 Matrigel-PBS dilution was applied to Transwell inserts and incubated at 37 °C for 1 h, discarding unsolidified Matrigel. Prior to seeding, trophoblastic cells were serum-starved for 12–24 h, digested, washed with PBS, and resuspended in serum-free medium at 1 × 10 5 cells/mL. Cells (100 µL) were seeded in each chamber, with 500 µL of 10% serum medium in the lower chamber. After 48 h of incubation, chambers were rinsed in PBS, and non-migrated cells were removed. Cells were fixed with 4% methanol, stained with 0.1% crystal violet, and excess dye washed off. Migrated cells were observed, photographed, and counted in five random fields under an inverted microscope, repeated three times for accuracy. To investigate RNA stability, we initially seeded the HTR-8/SVneo cell line at a density of 1 × 10^5 cells/mL into 6-well plates. Once the cells attained a confluency of 80% to 90%, we administered actinomycin D at a concentration of 5 µg/mL to the cell culture plates. Subsequently, at both 6 and 12 h post-treatment, we harvested the cells and promptly proceeded with real-time quantitative polymerase chain reaction (RT-qPCR) analysis to assess RNA stability. This approach allowed us to accurately determine the effect of actinomycin D on RNA degradation over time. Total RNA was extracted from the samples using Trizol reagent, following standard protocols. The quality of the extracted RNA was rigorously assessed using a spectrophotometer to ensure its integrity and purity. Subsequently, the RNA was converted to cDNA using reverse transcriptase and specific primers designed for reverse transcription. The RT-qPCR reaction mixture was precisely prepared according to the manufacturer’s instructions (Shanghai Yisheng Biotechnology Co., LTD), incorporating the cDNA template, primers, fluorescent dye or probe, enzyme, and buffer. To ensure reproducibility and statistical significance, at least three biological replicates were included for each sample. The relative expression level of the target gene was calculated using the 2 ^−ΔΔCt method, based on the Ct (cycle threshold) values obtained from the RT-qPCR reactions. This approach allows for accurate and quantitative assessment of gene expression in the samples. The primer sequence is as Table S1 . For tissue samples, 200 mg of frozen tissue was fragmented and transferred to 1.5 mL tubes with grinding beads. The tissue was pulverized using a liquid nitrogen-cooled grinding device, followed by the addition of 1 mL of lysis buffer. The mixture was thoroughly mixed and chilled on ice for 30 min to facilitate lysis. For adherent cells, the culture medium was discarded, and the cells were washed three times with chilled PBS. Protein lysis buffer was then added to the culture dishes in varying volumes based on plate type (150–200 µL per well for 6-well plates, 50–60 µL for 12-well plates). The cells were incubated on ice with gentle agitation for 20 min, scraped off, and the lysate transferred to 1.5 mL centrifuge tubes. The lysate was homogenized using an ultrasonic homogenizer at 30% intensity, with 2-second pulses and 3-second intervals, for a total of 50 s. Centrifugation at 12,000 rpm for 10 min at 4 °C separated the proteins from cellular debris. The resulting supernatant, containing the extracted proteins, was carefully collected and transferred to fresh 1.5 mL centrifuge tubes. SDS-PAGE gels were prepared according to the manufacturer’s guidelines, including both resolving and stacking gels. Prepared protein samples and markers were loaded into the gel wells. Electrophoresis was initiated at 80 V for 15–20 min to compress the samples in the stacking gel. Once a clear separation line was observed, the voltage was increased to 120 V to complete the electrophoretic run, ensuring effective separation of proteins based on their size and charge. To accurately measure triglyceride levels, we followed a precise protocol. Initially, we added 0.1 mL of lysis buffer to the cells cultured in a 6-well plate, thoroughly mixed the solution, and allowed it to sit at room temperature for 10 min. Following this, we heated the plate at 70 °C for 10 min, noting that flocculent precipitates may form due to high cell density. Once heated, we centrifuged the mixture at 2000 rpm for 5 min at 4 °C to separate the supernatant. This supernatant, rich in triglycerides, was then used for further analysis. Next, we prepared a triglyceride detection solution by mixing R1 and R2 in a 4:1 ratio. To establish a reliable standard curve, we serially diluted a 4 mM glycerol standard with physiological saline to obtain concentrations ranging from 1000 to 7.8125 µmol/L. For the assay, we dispensed 190 µL of the triglyceride detection solution into each well of a 96-well plate, adding 10 µL of either the sample or a standard solution. We replicated each group three times to ensure the accuracy of our measurements. Using a multifunctional microplate reader, we measured the absorbance at OD562 for each well. Based on the absorbance values of the standards, we plotted a standard curve and calculated the triglyceride concentration in the samples accordingly. This comprehensive procedure ensures the reliability and accuracy of our triglyceride detection results. For the purpose of accurately quantifying ATP levels, we adhered to a stringent protocol. Initially, the culture medium was removed, and 200 µL of lysis buffer was added to each well of a 6-well plate to lyse the cells. Following lysis, the mixture was centrifuged at 12,000 g for 5 min at 4 °C to obtain a supernatant, which was then used for subsequent measurements. To establish a reliable standard curve, we thawed the necessary reagents on ice and diluted the ATP standard solution with ATP detection lysis buffer to create a series of appropriate concentration gradients (0.01, 0.03, 0.1, 0.3, 1, 3, and 10 µM). Next, we prepared the ATP detection working solution by mixing the ATP detection reagent with the ATP detection reagent diluent in a 1:4 ratio. This working solution was prepared in an amount sufficient for each sample or standard, and was temporarily stored on ice. For the ATP concentration measurement, 100 µL of the ATP detection working solution was added to a 96-well plate (with alternating empty wells). The plate was then incubated at room temperature for 3–5 min. Finally, a multifunctional microplate reader was used to measure the chemiluminescence intensity at 560 nm. This process allowed us to accurately determine the ATP concentration in our samples. To precisely quantify the levels of reactive oxygen species (ROS), we adhered to a rigorous protocol. Initially, we formulated a detection solution by diluting DCFH-DA in serum culture medium at a 1:1000 ratio, yielding a final concentration of 10 µmol/L. Subsequently, we suspended the collected cells in the diluted DCFH-DA, ensuring a cell concentration of 106 cells per milliliter. For positive control, we mixed the reactive oxygen positive control reagent with the cell suspension at a 1:1000 ratio. Meanwhile, a cell suspension without additional reagents served as the negative control. Next, the cells were incubated in a 37 °C cell culture incubator for 20 min, with gentle agitation every 3 to 5 min to facilitate thorough interaction between the probe and the cells. Following incubation, we thoroughly washed the cells three times with serum-free cell culture medium to remove any DCFH-DA that had not penetrated the cells. Finally, the samples from each group were resuspended in PBS and analyzed using a flow cytometer to accurately determine the ROS content. This comprehensive approach ensures the accurate and reliable measurement of ROS levels. To determine malondialdehyde (MDA) levels, we initiated the process by collecting the cell precipitate. Based on accurate cell counting, we added the necessary volume of extraction buffer. Subsequently, we disrupted the cells using ultrasonic waves to effectively release their intracellular contents. The resulting mixture was centrifuged at 8000 g for 10 min at 4 °C, and the clarified supernatant was kept chilled on ice for further examination. Adhering strictly to the manufacturer’s guidelines, we added the prescribed reagents to the supernatant. The mixture was then incubated in a water bath at 100 °C for 60 min to promote the desired chemical reactions. Once the incubation period was complete, the mixture was rapidly cooled in an ice bath to terminate any ongoing reactions. To separate any insoluble components, we performed a subsequent centrifugation step at 10,000 g for 10 min at room temperature. Finally, with precision, we transferred 200 µL of the clarified supernatant into a 96-well plate. Using a spectrophotometer, we measured the absorbance of each sample at 450 nm, 532 nm, and 600 nm. These measurements will serve as valuable indicators of the MDA levels present in our samples, providing crucial insights into our experimental findings. SPSS 26.0 software was applied. Numerical variables were uniformly presented as mean ± standard deviation, ensuring clarity and accuracy. For normally distributed independent samples, we applied the Student’s t-test, a well-established method, to compare differences. Conversely, non-parametric tests were employed for non-normally distributed data, allowing for valid comparisons between independent samples. For multi-group comparisons, One-way ANOVA was utilized, a potent technique for examining variations across three or more groups. To investigate the relationship between variables, we conducted Pearson correlation analysis, a widely accepted measure that quantifies the strength and direction of linear association. The statistical significance of all tests was established by setting a P -value threshold of < 0.05.

Villous tissues were collected from URSA patients and normal pregnancy controls. The two groups were matched for maternal age, BMI, and gestational age, while the URSA group exhibited significantly higher abortion frequency and lower live birth rate (Table  1 ). Table 1 Baseline characteristics of URSA patients and normal controls Parameter Control( n  = 37) URSA( n  = 28) P Age (years) 30.43 ± 5.22 31.68 ± 4.71 0.7829 BMI (kg/m 2 ) 21.38 ± 3.22 21.13 ± 1.52 0.3314 Gestational weeks 6.68 ± 0.97 7.52 ± 1.24 0.2085 Spontaneous abortion (times) 0.08 ± 0.27 2.44 ± 0.69 < 0.0001 Number of live births (times) 1.41 ± 0.54 0.48 ± 0.68 < 0.0001 Baseline characteristics of URSA patients and normal controls FABP4 expression was significantly downregulated in URSA villous tissues at both mRNA and protein levels, as determined by RT-qPCR ( P  < 0.001, Fig.  1 A) and western blot ( P  = 0.02, Fig.  1 B). Immunofluorescence staining further localized this reduction to the cytoplasm of trophoblast cells (Fig.  1 C). Consistently, Raman spectroscopy revealed marked decreases in lipid droplets (1083 cm − 1 ) and triglycerides (1073 cm − 1 ) in URSA samples (Fig.  1 D- F), indicating disrupted lipid metabolism. These findings implicate FABP4 downregulation and associated lipid metabolic dysfunction in URSA pathogenesis. Fig. 1 FABP4 expression and lipid metabolism level in villus tissues of URSA patients. ( A ) The mRNA level of FABP4 in the villi of URSA patients and normal patients was detected by RT-qPCR; ( B ) FABP4 protein levels and relative quantitative results were determined by western blot; ( C ) The level and localization of FABP4 in villus tissues of URSA patients and normal group were detected by immunofluorescence. ( D ) Raman imaging of normal group and URSA in tissue sections; ( E ) Quantitative analysis of lipid droplet (LD) and triglyceride (TG) in Raman imaging; ( F ) Comparison of the wave peak of LD (1083 cm − 1 ) and TG (1073 cm − 1) between CTRL and URSA groups. Compared with the control, * means P  < 0.05; ** means P  < 0.01; *** means P  < 0.001 and **** means P  < 0.0001 FABP4 expression and lipid metabolism level in villus tissues of URSA patients. ( A ) The mRNA level of FABP4 in the villi of URSA patients and normal patients was detected by RT-qPCR; ( B ) FABP4 protein levels and relative quantitative results were determined by western blot; ( C ) The level and localization of FABP4 in villus tissues of URSA patients and normal group were detected by immunofluorescence. ( D ) Raman imaging of normal group and URSA in tissue sections; ( E ) Quantitative analysis of lipid droplet (LD) and triglyceride (TG) in Raman imaging; ( F ) Comparison of the wave peak of LD (1083 cm − 1 ) and TG (1073 cm − 1) between CTRL and URSA groups. Compared with the control, * means P  < 0.05; ** means P  < 0.01; *** means P  < 0.001 and **** means P  < 0.0001 To investigate the potential link between FABP4 and pregnancy loss, we conducted a study using Fabp4 knockout ( FABP4 −/− ) mice. In a 2:1 mating ratio with male mice, we observed that homozygous Fabp4 knockout ( FABP4 −/− ) female mice had a significantly lower pregnancy rate of 12.5% compared to 90% in wild-type females. This was accompanied by a higher non-pregnancy rate of 87.5% in FABP4 −/− females (Fig.  2 A). Furthermore, ultrasound imaging confirmed intrauterine pregnancies in both groups (Fig.  2 B). Additionally, we analyzed the genotypic and sex distribution of offspring from matings between FABP4 +/− and wild-type mice. Our findings indicated that wild-type females produced offspring with a balanced sex ratio, comprising approximately 48.6% male and 51.4% female pups. In contrast, heterozygous females produced offspring with a genotypic distribution of 20.8% wild-type, 57.3% heterozygous, and 21.9% homozygous, maintaining a near 1:1 female-to-male ratio, which aligns with Mendelian inheritance patterns (Fig.  2 C). In addition, we observed the embryo resorption situation. The results indicated that the embryos in the Fabp4 heterozygous group showed significant resorption at E13.5. The resorption rate of the heterozygous embryos was approximately 24.81%, and the non-resorption rate was about 75.19% (Fig. 2 D). Fig. 2 Pregnancy rate in Fabp4 knockout mice. ( A )Comparison of litter numbers between the two breeding pairs. Wild-type (WT) male and female mice (WT × WT), Homozygous Fabp4 knockout male and female mice ( FABP4 -/- × FABP4 -/- ); ( B ) Ultrasonic detection of pregnancy in mice; ( C ) Sex and genotypic division of the offspring of Fabp4 heterozygous ( FABP4 + /⁻) female mice and WT mice. ( D ) Comparison of embryo absorption between FABP4 + /⁻ female mice and WT female mice at 10 and 13.5 days gestation; ( E ) Comparison of placenta, embryo and embryo/placenta quality between FABP 4 + /⁻ and WT female mice; ( F ) Anatomical structure diagram and histological comparison of the placenta of FABP4 + /⁻ female mice and WT mice. Compared with wt, * means P < 0.05; *** means P < 0.001 Pregnancy rate in Fabp4 knockout mice. ( A )Comparison of litter numbers between the two breeding pairs. Wild-type (WT) male and female mice (WT × WT), Homozygous Fabp4 knockout male and female mice ( FABP4 -/- × FABP4 -/- ); ( B ) Ultrasonic detection of pregnancy in mice; ( C ) Sex and genotypic division of the offspring of Fabp4 heterozygous ( FABP4 + /⁻) female mice and WT mice. ( D ) Comparison of embryo absorption between FABP4 + /⁻ female mice and WT female mice at 10 and 13.5 days gestation; ( E ) Comparison of placenta, embryo and embryo/placenta quality between FABP 4 + /⁻ and WT female mice; ( F ) Anatomical structure diagram and histological comparison of the placenta of FABP4 + /⁻ female mice and WT mice. Compared with wt, * means P < 0.05; *** means P < 0.001 Upon detailed analysis of the non-resorbed embryos, we noted a significant difference in placental weight. The Fabp4 heterozygous embryos exhibited a placental weight of 0.115 ± 0.006 g, which was notably lower than the 0.137 ± 0.006 g recorded for wild-type embryos. Interestingly, when comparing the embryonic weights between the two groups, we found no significant statistical difference. However, the embryo-to-placenta weight ratio presented a different picture. The Fabp4 heterozygous group exhibited a significantly higher ratio compared to the wild-type group (Fig.  2 E). These compelling findings suggest a strong link between FABP4 and pregnancy loss. Specifically, the deficiency or downregulation of FABP4 appears to contribute to an increased risk of miscarriages and is closely associated with placental development. To further investigate this phenomenon, we compared the histological characteristics of the placentas between the two groups. Notably, the junctional zone (JZ) to labyrinth zone (LZ) area ratio in the placentas of wild-type female mice was larger than that of heterozygous female mice (Fig.  2 F). This finding suggests that abortion caused by FABP4 loss may be related to placental trophoblast cells, providing further evidence of the intricate relationship between FABP4 and placental development. To gain a deeper understanding of the relationship between FABP4 and trophoblast cell function, we initially screened three trophoblast cell lines: HTR-8/SVneo, JEG3, and JAR. Our comprehensive analysis revealed that there were no significant variations in the baseline levels of FABP4 protein or mRNA expression across these cell lines (Fig.  3 A). Given these results, we focused our subsequent studies on the widely utilized HTR-8/SVneo cell line. To modulate FABP4 expression in these cells, we used lentiviral vectors to introduce FABP4 plasmids and control plasmids into the HTR-8/SVneo cells. Western blot and RT-qPCR assays demonstrated that the FABP4-overexpressing (ov-FABP4) group exhibited significantly elevated levels of FABP4 protein and mRNA compared to the positive control (ov-NC) group (Fig.  3 B-C). Conversely, in the FABP4-knockdown (sh-FABP4) group, we observed significantly reduced levels of FABP4 protein and mRNA compared to the corresponding negative control (sh-NC) group (Fig.  3 D-E). These outcomes validate the successful establishment of both FABP4-overexpressing and FABP4-knockdown HTR-8/SVneo cell lines. All subsequent functional analyses were conducted using these modified cell lines and their respective control counterparts. Fig. 3 The influence of FABP4 on the phenotype of trophoblast cells. (A) Protein and mRNA expression of FABP4 in different trophoblast cell lines (HTR-8/SVneo cells, JEG3 cells and JAR cells); ( B-C ) Protein and mRNA expression of FABP4 in ov-NC and ov-FABP4 group; ( D-E ) Protein and mRNA expression of FABP4 in sh-NC and sh-FABP4 group. ( F ) Transwell detected the effect of FABP4 overexpression and knockdown on trophoblast invasion and migration and its quantitative analysis; ( G ) The effect of overexpression and knockdown of FABP4 on trophoblast proliferation was detected by CCK8; ( H ) The effect of FABP4 on invasion and migration molecules of MMP2, MMP9, TIMP1 and TIMP2 was detected by RT-qPCR. Compared with the control group, ns mean no statistical difference, * representing P  < 0.05, ** representing P  < 0.01, **** representing P  < 0.0001 The influence of FABP4 on the phenotype of trophoblast cells. (A) Protein and mRNA expression of FABP4 in different trophoblast cell lines (HTR-8/SVneo cells, JEG3 cells and JAR cells); ( B-C ) Protein and mRNA expression of FABP4 in ov-NC and ov-FABP4 group; ( D-E ) Protein and mRNA expression of FABP4 in sh-NC and sh-FABP4 group. ( F ) Transwell detected the effect of FABP4 overexpression and knockdown on trophoblast invasion and migration and its quantitative analysis; ( G ) The effect of overexpression and knockdown of FABP4 on trophoblast proliferation was detected by CCK8; ( H ) The effect of FABP4 on invasion and migration molecules of MMP2, MMP9, TIMP1 and TIMP2 was detected by RT-qPCR. Compared with the control group, ns mean no statistical difference, * representing P  < 0.05, ** representing P  < 0.01, **** representing P  < 0.0001 In this study, we investigated the influence of FABP4 on trophoblast cell invasion, migration, and proliferation. Our Transwell cell migration experiments revealed that, following 48 h of seeding, the ov-FABP4 group exhibited a slight increase in migration compared to the ov-NC group, although this difference was not statistically significant. However, the ov-FABP4 group demonstrated significantly higher invasiveness, indicating enhanced invasive capacity compared to the ov-NC group. Conversely, the sh-FABP4 group showed a significant reduction in both migration and invasiveness compared to the sh-NC group, indicating an inhibitory effect (Fig.  3 F). These findings suggest that FABP4 knockdown attenuates trophoblast invasion and migration, whereas its overexpression significantly enhances invasion. To further assess the role of FABP4 in trophoblast proliferation, we conducted CCK8 assays at 24, 48, 72, and 96 h. Our results indicate that FABP4 overexpression stimulates trophoblast proliferation in a time-dependent manner, whereas FABP4 knockdown does not significantly affect proliferation at earlier time points but exhibits a significant inhibitory effect at 72 h (Fig.  3 G). Additionally, RT-qPCR analysis suggested that overexpression of FABP4 promotes the expression of MMP2 and MMP9, while inhibiting the expression of TIMP1 and TIMP2 (Fig.  3 H). In summary, our experiments demonstrate that FABP4 plays a crucial role in regulating trophoblast invasion and proliferation. Specifically, FABP4 overexpression enhances these processes, while its knockdown exerts inhibitory effects on invasion and migration. However, the mechanism underlying the abnormal downregulation of FABP4 in the villus tissue of URSA patients requires further investigation. In the realm of post-transcriptional modifications, m6A stands as a prominent mechanism. This modification is primarily governed by three key regulators: the m6A methyltransferase, also known as the ‘writer’; the demethyltransferase, referred to as the “eraser”; and the m6A recognition protein, termed the “reader” [ 15 ]. Specifically, the m6A “writer” enzymes, including METTL3, METTL14, and WTAP, catalyze the formation of m6A modifications. Conversely, the m6A “eraser” enzymes, such as FTO and ALKBH5, are responsible for reversing or removing these modifications. However, the pivotal role in determining the fate of target mRNA lies with the m6A ‘readers’—proteins like IGF2BPs, YTHDCs, and YTHDFs. These m6A readers, by binding to mRNA marked with m6A, play a crucial role in influencing various aspects of mRNA metabolism, including nuclear transport, stability, splicing, and translation efficiency. Ultimately, this impacts the expression of target genes at the post-transcriptional level. Does m6A modification play a role in regulating trophoblastic biological activities? To examine the involvement of m6A modification in URSA, we analyzed the m6A levels in villus tissues from URSA patients and controls. Our results revealed that URSA patients had significantly higher m6A levels in their villus tissues compared to the normal control group (Fig.  4 A). To further elucidate the underlying mechanisms, we performed proteomic sequencing on villus tissues from URSA patients and normal controls. This analysis showed a notable reduction in the expression of the m6A eraser FTO in URSA patients. Additionally, we observed significant decreases in the expression of several m6A reader proteins, including YTHDF2, YTHDF3, SRSF2, IGF2BP3, HNRNPA2B1, and HNRNPC (Fig.  4 B). To validate these findings, we used RT-qPCR and Western blot to assess FTO expression in villus tissues. Our results demonstrated that URSA patients had significantly lower FTO protein expression compared to induced abortion patients (Fig.  4 C). Similarly, RT-qPCR analysis revealed that URSA patients exhibited lower FTO mRNA expression compared to normal controls (Fig.  4 D). In summary, our study indicates that URSA patients have significantly elevated m6A levels and concurrent reductions in FTO expression in their villus tissues. Fig. 4 Effects of m6A Reader FTO on FABP4 expression at cellular levels. (A) m6A levels in the villi of URSA patients and Normal group were quantitatively detected by m6A RNA methylation; ( B ) Heat maps of m6A related molecules were detected by protein sequencing; ( C ) Determination of FTO protein levels relative to GAPDH by western blot ( n  = 5) and relative quantitative results; ( D ) FTO mRNA levels were measured by RT-qPCR in villus tissues of URSA patients and normal patients; ( E-F ) Western Blot and RT-qPCR were used to detect the protein and mRNA expression of FTO in different trophoblast cell lines (HTR-8/SVneo cells, JEG3 cells and JAR cells). ( G ) Fluorescence microscopy was used to detect the transfection efficiency of different lentiviruses in HTR-8/SVneo cell lines; ( H ) The expression of FTO mRNA in ov-FTO and ov-NC group HTR-8/SVneo trophoblast cell lines was detected by RT-qPCR. ( I ) Western Blot analysis of FTO protein expression levels in ov-FTO and ov-NC HTR-8/SVneo trophoblast cell lines. ( J ) m6A kit was used to detect the effects of different FTO treatments on the overall m6A level of trophoblast cells; ( K ) The effects of ov-FTO, sh-FTO and FB23-2 groups on the expression of FTO and FABP4 mRNA were detected by RT-qPCR; ( L ) The effect of FTO on FABP4 mRNA m6A expression abundance in trophoblast cells was detected by meRIP-PCR; ( M ) Pearson method was used to analyze the correlation between FTO and FABP4. Compared with NC, * means P  < 0.05; ** means P  < 0.01; *** means P  < 0.001 and **** means P  < 0.0001 Effects of m6A Reader FTO on FABP4 expression at cellular levels. (A) m6A levels in the villi of URSA patients and Normal group were quantitatively detected by m6A RNA methylation; ( B ) Heat maps of m6A related molecules were detected by protein sequencing; ( C ) Determination of FTO protein levels relative to GAPDH by western blot ( n  = 5) and relative quantitative results; ( D ) FTO mRNA levels were measured by RT-qPCR in villus tissues of URSA patients and normal patients; ( E-F ) Western Blot and RT-qPCR were used to detect the protein and mRNA expression of FTO in different trophoblast cell lines (HTR-8/SVneo cells, JEG3 cells and JAR cells). ( G ) Fluorescence microscopy was used to detect the transfection efficiency of different lentiviruses in HTR-8/SVneo cell lines; ( H ) The expression of FTO mRNA in ov-FTO and ov-NC group HTR-8/SVneo trophoblast cell lines was detected by RT-qPCR. ( I ) Western Blot analysis of FTO protein expression levels in ov-FTO and ov-NC HTR-8/SVneo trophoblast cell lines. ( J ) m6A kit was used to detect the effects of different FTO treatments on the overall m6A level of trophoblast cells; ( K ) The effects of ov-FTO, sh-FTO and FB23-2 groups on the expression of FTO and FABP4 mRNA were detected by RT-qPCR; ( L ) The effect of FTO on FABP4 mRNA m6A expression abundance in trophoblast cells was detected by meRIP-PCR; ( M ) Pearson method was used to analyze the correlation between FTO and FABP4. Compared with NC, * means P  < 0.05; ** means P  < 0.01; *** means P  < 0.001 and **** means P  < 0.0001 To further investigate the role of FTO in regulating the m6A modification of FABP4, we successfully constructed trophoblast cells with both FTO overexpression and knockdown (Fig.  4 E-I). Furthermore, our findings reveal that the knockdown of FTO significantly impedes trophoblast migration and invasiveness compared to the control group, as illustrated in Figure S1 A-B. The CCK-8 assay indicates a time-dependent suppression of trophoblast proliferation following FTO knockdown (Figure S1 C). These results underscore the critical role of FTO in markedly promoting trophoblast invasion, migration, and proliferation. We subsequently analyzed how these alterations, along with the application of FB23-2 (a specific inhibitor of FTO activity ), impacted the m6A levels in FABP4 mRNA. Our key findings are as follows: overexpression of FTO led to a significant decrease in the overall m6A level in trophoblast cells, approximately 50% lower than the baseline. Conversely, knocking down or inhibiting FTO activity (FB23-2) resulted in a nearly 1.5-fold increase in the overall m6A level (Fig.  4 J). Moreover, our study revealed that FB23-2 was highly effective in suppressing FTO expression. Importantly, we observed that FTO overexpression significantly upregulated FABP4 mRNA expression, increasing it by approximately 3.87 times. Conversely, FTO knockdown significantly downregulated FABP4 expression, reducing it to approximately one-third of its original level (Fig.  4 K). These findings underscore the crucial role of FTO in modulating m6A abundance in trophoblast cells, with its overexpression promoting FABP4 expression and its knockdown or inhibition suppressing it. To enhance our understanding of the topic, we employed the meRIP-PCR kit in HTR-8/SVneo trophoblast cells. Our results revealed that both FTO knockdown (sh-FTO) and pharmacological inhibition of FTO activity (FB23-2) remarkably increased the abundance of m6A modifications in FABP4 mRNA (Fig.  4 L). Furthermore, a thorough correlation analysis between FTO and FABP4 mRNA expression levels provided additional insights. This analysis highlighted a robust positive correlation between FABP4 expression and FTO, evidenced by R 2 =0.857 (Fig.  4 M). Collectively, these findings highlight the positive correlation between FABP4 expression and FTO in trophoblast cells. Specifically, our data indicate that inhibiting FTO results in an increase in the abundance of m6A modifications in FABP4 mRNA, ultimately leading to the suppression of FABP4 expression. This discovery offers valuable insights into the regulatory mechanisms that govern FABP4 expression in trophoblast cells. To gain a deeper understanding of how FTO knockdown affects pregnancy outcomes, we established a rat model of LPS-induced abortion and employed FB23-2 to assess the consequences of FTO inhibition during pregnancy (Fig.  5 A). Our findings reveal that both LPS and FB23-2 significantly induced embryo resorption in these rats (Fig.  5 B). Further examination of placental tissues showed that LPS treatment substantially reduced the expression level of FTO and FABP4 mRNA. Similarly, compared to the DMSO control group, the FB23-2 group exhibited significantly lower expression levels of FTO and FABP4 mRNA (Figs.  5 C-D). These results indicate that both LPS and FB23-2 induce abortion in rats while concurrently lowering placental FTO and FABP4 mRNA expression. Fig. 5 Effects of FTO on FABP4 expression at animal levels. ( A ) Flow chart of rat abortion model induced by LPS and FB23-2; ( B ) Embryo absorption after treatment with LPS (URSA) and FB23-2; ( C ) The expression of FTO mRNA in placental tissues was detected by RT-qPCR. ( D ) The mRNA expression of FABP4 in placental tissues was detected by RT-qPCR. ( E ) The serum FABP4 and LPS level of rats was detected by ELISA; ( F ) ELISA was used to detect the level of FABP4 and LPS serum level in FB23-2 rats. Compared with the NP, * means P  < 0.05; ** means P  < 0.01; *** means P  < 0.001 and **** means P  < 0.0001 Effects of FTO on FABP4 expression at animal levels. ( A ) Flow chart of rat abortion model induced by LPS and FB23-2; ( B ) Embryo absorption after treatment with LPS (URSA) and FB23-2; ( C ) The expression of FTO mRNA in placental tissues was detected by RT-qPCR. ( D ) The mRNA expression of FABP4 in placental tissues was detected by RT-qPCR. ( E ) The serum FABP4 and LPS level of rats was detected by ELISA; ( F ) ELISA was used to detect the level of FABP4 and LPS serum level in FB23-2 rats. Compared with the NP, * means P  < 0.05; ** means P  < 0.01; *** means P  < 0.001 and **** means P  < 0.0001 When we analyzed serum levels of FABP4 and LPS in these models, we observed contrasting trends. In the LPS-induced abortion model, serum levels of FABP4 and LPS were significantly elevated compared to the control group (Fig.  5 E). However, in the FB23-2-treated rats, both FABP4 and LPS levels were significantly reduced (Fig.  5 F). It is noteworthy that in our animal experiments, the embryo resorption rate differed between the LPS-induced and FB23-2-induced abortion models. This discrepancy may be attributed to their distinct mechanisms of action. LPS is a potent inducer of systemic inflammation, mimicking the infectious or inflammatory etiology of abortion, which can lead to a strong and rapid activation of the maternal immune system, resulting in a high rate of embryo loss. In contrast, FB23-2 specifically inhibits FTO demethylase activity, directly affecting the epigenetic regulation of gene expression at the maternal-fetal interface. Its effect on FABP4 and subsequent placental dysfunction might be more gradual and localized, leading to a comparatively lower, yet still significant, rate of embryo resorption. Furthermore, the elevated serum FABP4 levels in the LPS model, contrasting with its placental downregulation, likely represent a compensatory systemic response to inflammation, a phenomenon not observed with the more targeted FTO inhibition by FB23-2. In the rat animal level, our finding further strengthens the positive correlation between FTO and FABP4 expression, indicating that FTO plays a pivotal role in regulating FABP4 levels in both placental tissue and the serum. To further investigate the role of m6A readers in URSA, we utilized RT-qPCR to analyze the expression of common m6A reader family genes, specifically IGF2BPs and YTHDFs. Our results revealed notable differences in the expression of IGF2BP2 and IGF2BP3 between normal and URSA groups. Specifically, IGF2BP2 expression was significantly upregulated in URSA patients, while IGF2BP3 expression was significantly downregulated (Fig.  6 A). Combined with the previous sequencing results, IGF2BP3 was selected for subsequent analysis (Fig.  6 B). To validate these findings, we conducted a single-cell RNA sequencing analysis of villi tissues from three normal individuals in early pregnancy (undergoing induced abortion) and three patients with URSA in the dataset GSE214607 . Specifically, we processed the data using Cellranger and employed the Seurat package for cell dimensionality reduction and clustering. To investigate the extravillous trophoblast (EVT) cell population, we leveraged characteristic genes. Further analysis of IGF2BP3 expression in EVT cell subsets, labeled as Villus cell trophoblast (VCT) and Syncytiotrophoblast (SVT), revealed a significant decrease in IGF2BP3 levels in these subsets from URSA patients (Fig.  6 C). This findings support that IGF2BP3 is downregulated in villus tissues of URSA patients, particularly in EVT trophoblast cells that are responsible for invasion and migration. Fig. 6 Effects of IGF2BP3 knockdown on FABP4 expression at cellular and animal levels. ( A ) The expression level of m6A reader related molecules (IGF2BPs and YTHDFs family) in villous tissues of normal group and URSA group; ( B ) The expression of IGF2BP3 in villi of normal group and URSA group was detected by RT-qPCR. ( C ) The expression levels of IGF2BP3 in EVT, VCT and SVT cells in normal group and URSA group in the GSE214607 dataset; ( D ) Effect of IGF2BP3 knockdown on FABP4 mRNA expression in HTR-8/SVneo trophoblast cells; ( E ) The expression level of FABP4 mRNA in trophoblast cells of sh-NC and sh-IGF2BP3 groups treated with actinomycin D was detected by RT-qPCR after 0, 6, and 12 h. ( F ) Absorption of mouse embryos and placenta weight after LPS treatment; ( G ) mRNA expression of IGF2BP3 and FABP4 in placental tissues was detected by RT-qPCR. Compared with the control (NP), * means P  < 0.05; ** means P  < 0.01; *** means P  < 0.001 and **** means P  < 0.0001 Effects of IGF2BP3 knockdown on FABP4 expression at cellular and animal levels. ( A ) The expression level of m6A reader related molecules (IGF2BPs and YTHDFs family) in villous tissues of normal group and URSA group; ( B ) The expression of IGF2BP3 in villi of normal group and URSA group was detected by RT-qPCR. ( C ) The expression levels of IGF2BP3 in EVT, VCT and SVT cells in normal group and URSA group in the GSE214607 dataset; ( D ) Effect of IGF2BP3 knockdown on FABP4 mRNA expression in HTR-8/SVneo trophoblast cells; ( E ) The expression level of FABP4 mRNA in trophoblast cells of sh-NC and sh-IGF2BP3 groups treated with actinomycin D was detected by RT-qPCR after 0, 6, and 12 h. ( F ) Absorption of mouse embryos and placenta weight after LPS treatment; ( G ) mRNA expression of IGF2BP3 and FABP4 in placental tissues was detected by RT-qPCR. Compared with the control (NP), * means P  < 0.05; ** means P  < 0.01; *** means P  < 0.001 and **** means P  < 0.0001 To delve deeper into the impact of IGF2BP3 on FABP4, we generated IGF2BP3 knockdown HTR-8/SVneo cells. Notably, this IGF2BP3 knockdown significantly reduced FABP4 mRNA expression (Fig.  6 D). As an m6A reader, IGF2BP3 is reported to modulate the mRNA stability of target genes through binding to m6A-modified regions of target transcripts, thereby regulating their post-transcriptional expression. To explore the potential regulatory relationship between IGF2BP3 and FABP4 in trophoblasts, we employed an actinomycin D experiment to assess the effect of IGF2BP3 knockdown on FABP4 mRNA stability at 0, 6, and 12 h in sh-IGF2BP3 and sh-NC groups. Our results indicated that IGF2BP3 knockdown significantly impaired the stability of FABP4 mRNA in trophoblasts (Fig.  6 E), suggesting a close association between IGF2BP3 expression and FABP4 mRNA stability. Transwell assays further suggested that reduced IGF2BP3 levels significantly inhibited trophoblast invasion and migration. Additionally, CCK8 assays indicated that IGF2BP3 knockdown could significantly inhibit the proliferation of trophoblast cells (Figure S2 ). In summary, IGF2BP3 knockdown is associated with reduced FABP4 mRNA stability and subsequent downregulation of FABP4 expression, which may jointly contribute to impaired trophoblast function and abortion. At the animal level, we conducted experiments using LPS-induced aborted mice and observed significant embryonic resorption, along with a notable reduction in placental weight compared to normal pregnancies (Fig.  6 F). Next, we examined the expression levels of IGF2BP3 and FABP4 in the placental tissues of these mice. Our results revealed that the mRNA expression of IGF2BP3 and FABP4 in the placental tissue was significantly lower in the URSA group (LPS induced abortion) compared to normal pregnant mice (Fig.  6 G). In summary, our animal experiments further confirm that the expression of both IGF2BP3 and FABP4 is significantly down-regulated in the placental tissues of LPS-induced aborted mice. Therefore, our findings indicate that the downregulation of IGF2BP3 is closely associated with reduced FABP4 expression, which is linked to the impaired stability of FABP4 mRNA in trophoblasts. Furthermore, we found that the expression levels of IGF2BP3 and FABP4 were notably decreased in placental tissue from LPS-induced aborted mice. These results suggest a critical role for IGF2BP3 and FABP4 in maintaining trophoblastic function and placental development. To explore the potential mechanisms by which FABP4 influences trophoblast invasion and migration, we conducted RNA-seq analysis on trophoblast cells with FABP4 overexpression (ov-FABP4 vs. ov-NC) and FABP4 knockdown (sh-FABP4 vs. sh-NC). KEGG analysis of differentially expressed genes from both datasets revealed consistent enrichment in oxidative phosphorylation and oxidative stress-related pathways (Fig.  7 A and Figure S3 A), suggesting that FABP4 is involved in regulating these processes. To further investigate the role of FABP4 in oxidative phosphorylation, we performed Seahorse metabolic flux analysis. Our results demonstrated that FABP4 knockdown significantly reduced the oxygen consumption rate (OCR) of trophoblast cells (Fig.  7 B).   Detailed analysis of key Seahorse parameters revealed that the sh-FABP4 group exhibited significantly lower basal respiration, ATP production, and spare respiratory capacity compared to the sh-NC control group (Figure S3 B), indicating that FABP4 downregulation impairs mitochondrial oxidative phosphorylation and cellular energy metabolism. To further investigate the effects of FABP4 on oxidative stress, we detected the effects of FABP4 on ROS and MDA, and the results suggested that FABP4 knockdown significantly increased ROS and MDA levels (Fig.  7 C-D). Therefore, FABP4 downregulation increased oxidative stress levels in trophoblasts. Fig. 7 The effect of FABP4 Knockdown or of FTO Inhibition on Oxidative Phosphorylation Levels. ( A ) KEGG analysis results of the top20 differential expression genes of trophoblast RNA-seq of ov-NC and ov-FABP4. ( B ) Oxygen consumption rate (OCR) of sh-NC group and sh-FABP4 group was measured by Seahorse energy metabolism analysis. ( C ) ROS levels in sh-NC group and sh-FABP4 group were measured and quantitatively analyzed by flow cytometry; ( D ) MDA levels in sh-NC group and sh-FABP4 group were detected by MDA kit. ( E ) GO enrichment analysis of top20 differential expression genes of RNA-seq of sh-NC group and sh-FTO group. ( F ) The oxygen consumption rate (OCR) of trophoblast cells in sh-NC group, sh-FTO group, DMSO group and FB23-2 group was determined by Seahorse energy metabolism analysis. ( G ) MDA levels in sh-NC group, sh-FTO group, DMSO group and FB23-2 group were detected by MDA kit. (H)ATP production levels in sh-NC group, sh-FTO group, DMSO group and FB23-2 group were detected by ATP kit. Compared with the control, * P  < 0.05, ** P  < 0.01, **** P  < 0.0001 The effect of FABP4 Knockdown or of FTO Inhibition on Oxidative Phosphorylation Levels. ( A ) KEGG analysis results of the top20 differential expression genes of trophoblast RNA-seq of ov-NC and ov-FABP4. ( B ) Oxygen consumption rate (OCR) of sh-NC group and sh-FABP4 group was measured by Seahorse energy metabolism analysis. ( C ) ROS levels in sh-NC group and sh-FABP4 group were measured and quantitatively analyzed by flow cytometry; ( D ) MDA levels in sh-NC group and sh-FABP4 group were detected by MDA kit. ( E ) GO enrichment analysis of top20 differential expression genes of RNA-seq of sh-NC group and sh-FTO group. ( F ) The oxygen consumption rate (OCR) of trophoblast cells in sh-NC group, sh-FTO group, DMSO group and FB23-2 group was determined by Seahorse energy metabolism analysis. ( G ) MDA levels in sh-NC group, sh-FTO group, DMSO group and FB23-2 group were detected by MDA kit. (H)ATP production levels in sh-NC group, sh-FTO group, DMSO group and FB23-2 group were detected by ATP kit. Compared with the control, * P  < 0.05, ** P  < 0.01, **** P  < 0.0001 Moreover, we conducted RNA-seq on trophoblast cells with FTO knocked down, and the subsequent GO analysis suggested that FTO knockdown was closely related to embryonic development, positive regulation of oxidative stress, and other pathways (Fig.  7 E). To delve deeper into the functional implications of FTO modulation more thoroughly, we conducted Seahorse energy metabolism assays. Seahorse energy metabolism assays further revealed that both FTO knockdown and treatment with the specific FTO inhibitor FB23-2 resulted in a significant decrease in OCR efficiency (Fig.  7 F). Analysis of key parameters showed that the sh-FTO group exhibited significantly lower basal respiration, ATP production, and spare respiratory capacity compared to the sh-NC group. Similarly, the FB23-2 group showed significantly reduced basal respiration, ATP production, and spare respiratory capacity compared to the DMSO control group (Figure S3 C). MDA results further suggested that knocking down or inhibiting FTO could significantly promote MDA levels (Fig.  7 G). Consistent with the Seahorse analysis, both the sh-FTO and FB23-2 groups displayed significantly lower ATP production levels compared to their respective controls (sh-NC and DMSO) (Fig.  7 H). Collectively, these results emphasize the crucial role of FTO in regulating mitochondrial function and energy metabolism in trophoblast cells, highlighting its potential as a therapeutic target in related disorders. FABP4 plays a crucial role in mitochondrial metabolism process by facilitating the transport of free fatty acids to the mitochondria, where they undergo oxidative metabolism to generate energy. Disruptions in FABP4 function can lead to imbalances in lipid metabolism, potentially leading to the development of oxidative phosphorylation disorders. To explore the lipid droplet formation due to FABP4 modulation, we conducted experiments using Oil Red O staining. Our findings revealed that in trophoblast cells, overexpression of FABP4 significantly enhanced lipid droplet formation compared to the control group, while knocking down FABP4 resulted in a notable inhibition of lipid droplet levels (Fig.  8 A). Furthermore, using a triglyceride assay kit, we observed a significant increase in triglyceride formation in trophoblast cells with FABP4 overexpression. In contrast, FABP4 knockdown led to a substantial decrease in triglyceride levels (Fig.  8 B). Consistent with the Seahorse analysis, knocking down FABP4 inhibits ATP production, while overexpression has the opposite effect (Fig.  8 C). Expanding our investigation, we also examined the impact of FTO knockdown on lipid metabolism. Our findings demonstrated that FTO knockdown significantly inhibited lipid droplet formation in trophoblast cells. Additionally, analysis of intracellular triglyceride levels showed that FTO knockdown led to a marked reduction in triglyceride production compared to the control group (Figure S4 ). These outcomes indicate that FTO knockdown can induce lipid metabolic disorders in trophoblast cells. Fig. 8 The influence of FABP4 and FTO on lipid metabolism. (A- B) Determination and quantitative analysis of lipid droplets and triglycerides in FABP4-treated HTR-8/SVneo trophoblast cells by oil red O and triglyceride kit; ( C ) ATP production levels in ov-NC, ov-FABP4, sh-NC, sh-FABP4 group were detected by ATP kit; ( D ) Venn diagram showing the overlap of differentially expressed genes (DEGs) between two RNA-seq datasets, “FTO_DEGs” represent genes significantly altered upon FTO overexpression (ov-FTO vs. ov-NC), “FABP4_DEGs” represent genes significantly altered upon FABP4 knockdown (sh-FABP4 vs. sh-NC); ( E ) The effect of FABP4 on the mRNA expression of SOD2 was detected by RT-qPCR; ( F ) The effect of FB23-2 on the protein expression of SOD2 was detected by Western Blot. ( G ) The effect of LPS on SOD2 mRNA expression in rat placenta was detected by RT-qPCR; ( H ) ELISA was used to detect the effects of MDA levels and SOD2 secretion in URSA and FB23-2 rats. Compared with the control (NP), * means P  < 0.05; ** means P  < 0.01; *** means P  < 0.001 and **** means P  < 0.0001 The influence of FABP4 and FTO on lipid metabolism. (A- B) Determination and quantitative analysis of lipid droplets and triglycerides in FABP4-treated HTR-8/SVneo trophoblast cells by oil red O and triglyceride kit; ( C ) ATP production levels in ov-NC, ov-FABP4, sh-NC, sh-FABP4 group were detected by ATP kit; ( D ) Venn diagram showing the overlap of differentially expressed genes (DEGs) between two RNA-seq datasets, “FTO_DEGs” represent genes significantly altered upon FTO overexpression (ov-FTO vs. ov-NC), “FABP4_DEGs” represent genes significantly altered upon FABP4 knockdown (sh-FABP4 vs. sh-NC); ( E ) The effect of FABP4 on the mRNA expression of SOD2 was detected by RT-qPCR; ( F ) The effect of FB23-2 on the protein expression of SOD2 was detected by Western Blot. ( G ) The effect of LPS on SOD2 mRNA expression in rat placenta was detected by RT-qPCR; ( H ) ELISA was used to detect the effects of MDA levels and SOD2 secretion in URSA and FB23-2 rats. Compared with the control (NP), * means P  < 0.05; ** means P  < 0.01; *** means P  < 0.001 and **** means P  < 0.0001 Remarkably, by analyzing the FABP4 and FTO sequencing data, we identified both FTO and FABP4 as central regulators of oxidative stress. To further investigate how these molecules jointly control oxidative stress, we completed the intersection of the two groups of differentially expressed genes. Our analysis identified 18 common genes, with SOD2 standing out as the only gene directly linked to oxidative stress (Fig.  8 D). When we compared SOD2 expression levels between cells treated with sh-FABP4 and sh-NC, we observed a significant decrease in SOD2 mRNA levels (Fig.  8 E). Additionally, FB23-2 was observed to significantly inhibit SOD2 protein expression (Fig.  8 F). Collectively, these findings indicated that m6A-mediated FABP4 downregulation is closely correlated with SOD2 inhibition and subsequent trophoblast oxidative stress. In rat placental tissues, we investigated the potential connection between SOD2 and miscarriage. In LPS induced abortion, Although we observed a declining trend in SOD2 expression, this change did not reach statistical significance. However, we observed a significant downregulation of SOD2 expression in FB23-2 induced abortion (Fig.  8 G). In rat serum, we detected the MDA levels and SOD2 secretion. In LPS induced abortion, we found significant elevations in MDA levels, indicating oxidative stress. This oxidative stress was accompanied by a marked decrease in SOD2 levels, suggesting a potential role for SOD2 in mitigating oxidative damage. Correspondingly, the serum MDA levels in the FB23-2-treated rats were significantly elevated compared to controls, reinforcing the presence of oxidative stress. Moreover, SOD2 levels in the serum of these rats were conspicuously reduced (Fig.  8 H ) , indicating a potential mechanism for oxidative stress-induced miscarriage. Collectively, these findings suggest a potential mechanism in which FTO-mediated downregulation of FABP4 inhibits SOD2 expression, and increased oxidative stress that may lead to miscarriage. This suggests a potential link modulating SOD2 levels or the pathways involved in its regulation could provide therapeutic options for the prevention or management of miscarriage.

In summary, we have uncovered a novel mechanism governing FABP4 expression in trophoblast cells. Specifically, we found that downregulation of the m6A eraser FTO increases the m6A modification of FABP4 mRNA, thereby suppressing its expression. Additionally, the abnormal downregulation of the m6A reader IGF2BP3 regulates FABP4 mRNA, further enhancing its suppression. Importantly, this imbalance in FABP4 expression, orchestrated by m6A RNA methylation regulators, has far-reaching consequences. It induces lipid metabolic disruptions and inhibits mitochondrial oxidative phosphorylation, ultimately triggering oxidative stress in trophoblast cells by suppressing SOD2 expression. This cascade of events hampers the cells’ ability to invade and migrate. These findings are significant as they point to a potential role in the development of URSA.

FABP4, a lipid-binding protein predominantly expressed in adipose tissue and macrophages, plays a pivotal role in regulating lipid metabolism and insulin sensitivity [ 19 ]. During pregnancy, FABP4 expression is intricately modulated by various factors, including hormonal status, nutritional conditions, and inflammatory responses [ 20 ]. Recent research has identified FABP4 in placental tissues and trophoblasts, suggesting that its aberrant expression and function may be implicated in pregnancy-related disorders [ 9 , 10 ]. Firstly, elevated FABP4 levels have been implicated in the pathophysiology of gestational hypertension, potentially via promoting inflammatory reactions and endothelial dysfunction [ 21 ]. Secondly, significant increases in FABP4 expression have been observed in patients with gestational diabetes mellitus (GDM), correlating with insulin resistance and abnormal insulin secretion. This suggests that FABP4 may contribute to GDM development by modulating lipid metabolism and insulin signaling pathways [ 22 , 23 ]. Lastly, FABP4 may be involved in fetal growth restriction, preterm birth, and other complications by influencing placental lipid metabolism, trophoblast function, and placental vascular development [ 24 ]. Despite its potential significance, research on FABP4 in the context of URSA remains limited, with few studies exploring its underlying mechanisms. Preliminary clinical investigations have only hinted at increased serum FABP4 levels in URSA patients [ 14 ]. Consequently, the relationship between FABP4 and URSA warrants further exploration and in-depth investigation. In our current study, we observed a marked decrease in FABP4 levels in the villous tissues of URSA patients. To elucidate the potential association between FABP4 and miscarriage, we used Fabp4 gene knockout mice and found a significant increase in embryo loss rates among these mice. To further explore the relationship between FABP4 and miscarriage, we compared the pregnancy outcomes of Fabp4 heterozygous female mice. Our findings revealed that heterozygous Fabp4 knockout mice exhibited significantly higher rates of embryo resorption compared to their wild-type counterparts, along with a notable reduction in the JZ/LZ area ratio in the placenta. This suggests that the absence of FABP4 expression may be intricately linked to placental development and the trophoblast layer within the JZ. To substantiate these observations, we conducted experiments in trophoblast cells. We discovered that overexpression of FABP4 promoted trophoblast proliferation and invasion, whereas its knockdown inhibited trophoblast migration and invasion. However, the mechanisms underlying the aberrant downregulation of FABP4 in the villous tissues of URSA patients remain elusive and warrant further investigation. m6A is one of the most prevalent post-transcriptional modifications, intricately regulating gene expression. The m6A writers, including METTL3, METTL14, and WTAP, catalyze the formation of m6A modifications, while the erasers, FTO and ALKBH5, reverse this process by removing m6A marks. The m6A readers, such as IGF2BPs, YTHDCs, and YTHDFs, play a decisive role in determining the fate of target mRNAs [ 25 ]. In our study, using proteomic profiling of URSA tissues, bioinformatics analysis, and an array of functional experiments, we observed a pronounced elevation in the global m6A level within the villous tissues of URSA patients. Concurrently, we identified significant downregulation of both the m6A eraser FTO, which possesses enzymatic activity, and the m6A reader IGF2BP3. FTO, the pioneeringly identified protein with m6A demethylase activity, modulates the m6A modification levels of target mRNAs, thereby influencing their degradation rates and post-transcriptional regulation [ 26 ] [ 27 ]. Notably, reduced FTO expression and elevated m6A levels have been observed in the villous tissues of URSA patients [ 28 ], whereas in URSA, increased FTO expression inhibits MEG3 m6A modification, downregulates TGF-β, and subsequently suppresses trophoblast invasion and proliferation [ 29 ], emphasizing FTO’s pivotal role in URSA. Notably, our finding of decreased FABP4 expression in URSA placental tissues appears to contrast with previous reports showing elevated serum FABP4 levels in patients with abortion [ 14 ]. This discrepancy may be explained by the distinction between local tissue dysfunction and systemic inflammatory responses. FABP4 is an inflammatory adipokine that can be secreted by adipocytes and macrophages into the circulation. In our LPS-induced abortion rat model, we observed opposing trends between local placental FABP4 expression and systemic FABP4 levels. LPS administration via intraperitoneal injection leads to inevitable systemic LPS exposure, triggering a potent inflammatory response that increases serum FABP4 secretion. This elevation likely represents a compensatory systemic reaction to inflammation, rather than reflecting the local placental environment. In contrast, pharmacological inhibition of FTO with FB23-2 significantly reduced FABP4 expression in both placental tissues and systemic circulation, further supporting a positive correlation between FTO and FABP4. In recent years, a growing body of evidence has linked aberrant expression of IGF2BP3 to diverse pregnancy-related disorders. For instance, Li et al. reported significantly lower IGF2BP3 expression in preeclamptic placentas compared to gestationally matched normal placentas [ 30 ]. Our preliminary studies have uncovered distinct roles of IGF2BP3 in the context of pregnancy outcomes. Specifically, downregulation of IGF2BP3 in trophoblasts triggers an imbalance in M2/M1 macrophage polarization by activating the NF-κB pathway and reducing IL-10 expression, ultimately contributing to abortion [ 31 ]. Conversely, high expression of IGF2BP3 in decidual tissues can lead to impaired decidualization and abnormal cell cycle regulation, also resulting in abortion [ 32 ]. These findings underscore the intricate involvement of IGF2BP3 in maintaining pregnancy homeostasis and its potential as a therapeutic target for pregnancy complications. Further investigation using meRIP-PCR and actinomycin assays revealed that knockdown or inhibition of FTO increases the abundance of m6A modifications on FABP4 mRNA, thereby suppressing FABP4 expression. Additionally, inhibition of IGF2BP3 is significantly associated with the destabilization of FABP4 mRNA in trophoblasts, which further contributes to the downregulation of FABP4 expression. As a classic m6A reader, IGF2BP3 mediates target mRNA stability through direct binding to m6A-modified transcripts, but the direct physical binding between IGF2BP3 and FABP4 mRNA has not been verified in the present study, and the regulatory relationship between them may be direct or indirect, which needs to be further explored. At the animal level, administration of FB23-2 to rats led to increased embryonic resorption, accompanied by reduced expression and secretion of FABP4 in both the placenta and serum. Similarly, our findings in an LPS-induced mouse abortion model confirmed the significant downregulation of Igf2bp3 and Fabp4 in the placental tissues of aborted fetuses. Therefore, these observations collectively underscore the intricate regulatory roles of m6A modification, mediated by FTO and IGF2BP3, in modulating FABP4 expression and potentially resulting to URSA pathogenesis. How does m6A regulator mediate FABP4 expression imbalance regulate and affect trophoblastic biological function? Numerous studies have established a link between lipid metabolic dysregulation and URSA. For instance, PPAR deficiency impairs fatty acid uptake and transporter expression, promotes production and secretion of inflammatory cytokines, and thereby impairs placental development and function, ultimately contributing to elevated URSA rates [ 33 , 34 ]. Additionally, insulin resistance is associated with elevated triglyceride levels, which leads to increased CD3 + CD4 + ratios and CD3 + CD8 + lymphocyte counts, reduced insulin sensitivity, and the induction of metabolic inflammation, predisposing individuals to URSA [ 35 ]. Our team previously reviewed the intricate relationship between lipid metabolism disorders and pregnancy complications, including abortion, in a comprehensive analysis [ 36 ]. Dysregulation of lipid metabolism in trophoblasts is often accompanied by heightened chronic inflammation and oxidative stress, which can compromise trophoblast function, causing cellular damage and dysfunction, and ultimately triggering abortion. Vondra et al. observed accumulations of free and esterified cholesterol in the villous trophoblasts of URSA patients [ 37 ]. This lipid disorder can lead to diminished mitochondrial membrane potential, downregulation of oxidative phosphorylation, and mitochondrial dysfunction, thereby reducing trophoblast invasiveness [ 38 , 39 ]. Furthermore, lipid disturbances promote the accumulation of ROS, triggering oxidative stress and ferroptosis, a type of cell death implicated in URSA [ 40 , 41 ]. In summary, lipid metabolic disorders in placental trophoblasts impact trophoblast function by modulating mitochondrial oxidative phosphorylation and inducing oxidative stress, thereby leading to URSA. In this study, we observed a pronounced reduction in lipid droplet and triglyceride expression in URSA villous tissues compared to controls, indicating the presence of lipid metabolic disorders in these tissues. Both FTO and FABP4 are proteins intricately linked to lipid metabolism and obesity, albeit executing distinct yet interconnected roles within cells. FTO influences lipid metabolism primarily through modulating the expression of genes related to this pathway [ 42 ], while FABP4 directly participates in lipid transport, storage, and metabolism [ 43 ]. The polymorphisms in the FTO gene are closely associated with obesity development [ 44 ], and FABP4 overexpression has also been implicated in obesity and related metabolic disorders [ 45 ]. A handful of studies suggest indirect regulation of FABP4 expression by FTO. Specifically, in adipocytes, overexpression of m6A eraser FTO promotes autophagy, which enhances FABP4 expression and consequently facilitates adipogenesis [ 46 ]. Furthermore, FTO may indirectly modulate FABP4 expression and function by influencing insulin signaling pathway activity. In our study, we found that FTO and FABP4 are closely related to oxidative phosphorylation and oxidative stress via sequencing. Leveraging the lipid-metabolic regulatory roles of FTO and FABP4, we further examined their impacts on lipid metabolism. Our findings indicate that downregulation of either FTO or FABP4 markedly inhibited lipid droplet formation and triglyceride accumulation, and diminished trophoblast oxidative phosphorylation. Moreover, by intersecting the effects of FTO- and FABP4-manipulated trophoblasts, we identified SOD2 as a pivotal player in m6A-mediated FABP4 regulation of trophoblast oxidative stress. Specifically, m6A-induced downregulation of FABP4 may inhibit SOD2 expression and activate trophoblast oxidative stress. Therefore, dysregulation of FABP4 expression mediated by m6A RNA methylation modulators can induce lipid metabolic perturbations, inhibit mitochondrial oxidative phosphorylation, and enhance trophoblast oxidative stress. This cascade ultimately inhibits trophoblast invasion and migration, thereby contributing to the pathological processes underlying URSA. Despite the progress made in this study, Several limitations of this study should be acknowledged, particularly regarding the animal models used. Currently, there is no universally accepted animal model that fully recapitulates human URSA. The most commonly used models include the CBA/J × DBA/2 mouse model of spontaneous abortion and the LPS-induced inflammatory abortion model [ 47 ]. While the LPS-induced model offers advantages such as simplicity, high reproducibility, and relevance to inflammation-associated pregnancy loss [ 48 , 49 ], it does not fully capture the “unexplained” nature of human URSA. LPS administration via intraperitoneal injection leads to systemic inflammatory exposure, which differs from the localized and idiopathic characteristics of URSA. In our study, LPS treatment significantly elevated serum FABP4 levels—consistent with its role as an inflammatory adipokine secreted by adipocytes in response to systemic inflammation [ 50 ], while paradoxically reducing placental FTO and FABP4 expression. In contrast, the FTO inhibitor FB23-2 reduced FABP4 levels in both placental tissue and serum, suggesting a more targeted effect on the FTO-FABP4 axis. The maternal-fetal interface is a highly intricate microenvironment, characterized by complex interactions among trophoblasts, decidual stromal cells, and immune cells, which collectively foster an immune-tolerant milieu. While we have begun to explore the regulatory mechanisms of FABP4 in trophoblasts, its influence on decidual stromal cells and immune cells within this interface remains underexplored. Future research should aim to comprehensively assess the multifaceted role of FABP4 across the maternal-fetal interface to gain a deeper understanding of its contributions to pregnancy maintenance and complications.

Recurrent spontaneous abortion (RSA), defined as two or more consecutive pregnancy losses before the 28th week of gestation with the same partner, affects 1% ~5% of reproductive-aged couples, imposing severe physical and psychological burdens [ 1 ]. While RSA has been associated with genetic factors, endocrine disorders, and immune dysregulation, approximately 40 ~ 60% of cases remain idiopathic, which are defined as unexplained recurrent spontaneous abortion (URSA) [ 2 ]. In recent years, delayed childbearing and shifts in family planning policies have led to a rising population of advanced-age pregnant women. Clinical guidelines emphasize the need to address URSA recurrence risks during diagnosis. Hence, elucidating the pathogenesis of URSA has paramount clinical and societal importance. During the first trimester of pregnancy, placental trophoblast cells will invade the decidua and myometrium of the mother, securing sufficient maternal blood supply for fetal development. These extravillous trophoblasts remodel maternal spiral arteries, channeling blood flow into dilated villous regions. This transformation reduces vascular resistance, ensuring optimal perfusion pressure and meeting the elevated metabolic demands of the growing fetus [ 3 ]. Inadequate trophoblast invasion can result in shallow implantation, impaired arterial remodeling, and subsequent placental ischemia, hypoxia, and nutrient deprivation, ultimately leading to pregnancy loss [ 4 ]. The invasion process is tightly regulated by a complex interplay of factors, including gene expression, extracellular matrix composition, cell-cell interactions, signaling pathways, cell cycle dynamics, and environmental cues [ 5 , 6 ]. Thus, deciphering the mechanisms underlying insufficient trophoblast invasion is essential for advancing the treatment of URSA. Fatty acid-binding protein 4 (FABP4), a 14–15 kDa lipid transport protein, is primarily expressed in mature adipocytes, macrophages, and capillary endothelial cells [ 7 ]. Recent studies have identified its expression in additional cell types, including dendritic cells [ 8 ], endothelial cells [ 9 ], and placental trophoblast cells [ 10 ]. FABP4 has gained recognition as an important biomarker associated with various pregnancy complications, such as gestational diabetes [ 11 ], preeclampsia [ 12 ], and preterm birth [ 13 ]. However, FABP4’s role in URSA remains poorly understood. To date, only one study has reported significantly elevated serum FABP4 levels in abortion patients compared to healthy controls, hinting at its potential involvement [ 14 ]. Hence, exploring the role and mechanism of FABP4 in placental villi tissue will provide valuable evidence for understanding the occurrence and development of URSA. N6-methyladenosine (m6A), the most prevalent post-transcriptional modification of mRNA, primarily occurs near stop codons and in 3’ untranslated regions (3’-UTRs) with the conserved RRACH motif. This dynamic modification is regulated by three classes of proteins: methyltransferases (“writers,” e.g., METTL3/14-WTAP complex), demethylases (“erasers,” e.g., FTO and ALKBH5), and m6A-binding proteins (“readers,” e.g., IGF2BPs, YTHDCs, and YTHDFs) [ 15 ]. Emerging evidence implicates m6A methylation in various reproductive disorders, including ovarian cancer [ 16 ], endometriosis [ 17 ], and infertility [ 18 ]. Intriguingly, studies have demonstrated elevated m6A methylation levels in the endometrial tissue of URSA patients but decreased levels in villous tissue, suggesting its potential regulatory role in pregnancy maintenance. However, the involvement of m6A modifications in regulating FABP4 expression in trophoblasts and the specific molecular mediators remain unknown, warranting further investigation. In this study, we utilized villus tissues from URSA patients to examine FABP4 expression in trophoblasts, and performed Fabp4 gene knockout mice to explore its role in abortion. Our results revealed significant downregulation of FABP4 in URSA villus tissues. Through integrative omics and molecular analyses, we identified the m6A regulators FTO and IGF2BP3 as key mediators of FABP4 suppression in URSA. Mechanistically, FTO downregulation enhanced m6A modification of FABP4 mRNA, while reduced IGF2BP3 levels further destabilized it, collectively repressing FABP4 expression. Functional studies revealed that FABP4 deficiency disrupts lipid metabolism, leading to oxidative stress and impaired oxidative phosphorylation via SOD2 dysregulation, ultimately contributing to abortion. These findings advance the understanding of URSA pathogenesis and highlight potential therapeutic targets.

Below is the link to the electronic supplementary material. Supplementary Material 1 Supplementary Material 1