Apoptotic vesicles of mesenchymal stem cells promote M2 polarization and alleviate early-onset preeclampsia via miR-191-5p

Abstract

Background

Macrophages play a crucial role in the development of early-onset preeclampsia (EOPE), which may be closely associated with an imbalance in macrophage M1/M2 polarization. Mesenchymal stem cell (MSC)-derived apoptotic vesicles (apoVs) have anti-inflammatory, tissue repair, and immunomodulatory functions. MSC-apoVs may ameliorate EOPE by regulating macrophage polarization, but the underlying mechanisms remain to be clarified.

Methods

Macrophage infiltration and M1/M2 polarization were first analyzed in the placentas of PE patients and normal pregancies to identify macrophage alterations in EOPE placentas. MSC-apoVs were extracted and characterized. The effects of MSC-apoVs on macrophage polarization and trophoblasts invasion were validated in vivo and in vitro. miRNA transcriptomic sequencing of MSC-apoVs was conducted to identify key miRNAs involved in macrophage M2 polarization and to investigate upstream and downstream regulation factors, which were further validated in vivo and in vitro.

Results

The proportion of M2 macrophages was significantly reduced in EOPE placentas. MSC-apoVs carrying high levels of miR-191-5p recruited macrophages, downregulated CDK6 protein expression, stabilized mitochondrial membrane potential (MMP), and promoted M2 polarization of macrophages. This enhanced the invasion of trophoblasts and improved EOPE pregnancy outcomes in mice, including reduced blood pressure, decreased urine protein, and improved embryo quality. Overexpression of miR-191-5p mimics in MSC-apoVs further alleviated EOPE-related symptoms, whereas inhibition of miR-191-5p reduced the therapeutic effect of MSC-apoVs. Further experiments confirmed that M2 macrophages polarized by MSC-apoVs promote trophoblasts invasion by secreting platelet-derived growth factor-AB (PDGF-AB), which binds to platelet-derived growth factor receptor-beta (PDGFR-β) on trophoblasts, directly activating the downstream PI3K-AKT-mTOR signaling pathway, thereby improving EOPE.

Conclusion

Our findings reveal the crucial role of M2 macrophages in the pathogenesis of EOPE. MSC-apoVs with high miR-191-5p recruit macrophages, downregulate CDK6, stabilize MMP, and promote M2 polarization, increasing PDGF-AB secretion, which enhances trophoblasts invasion and thereby treat EOPE. Therefore, MSC-apoVs therapy may serve as a promising strategy to improve the prognosis of EOPE.

Supplementary Information

The online version contains supplementary material available at 10.1186/s13287-025-04546-5.

Introduction

Preeclampsia (PE) is a severe pregnancy complication characterized by hypertension and proteinuria, which can lead to maternal and fetal mortality in severe cases. Based on the timing of onset, PE is classified into early-onset preeclampsia (EOPE) and late-onset preeclampsia (LOPE) [1]. The pathogenesis of EOPE is more complex and involves insufficient trophoblasts invasion and maternal immune system dysregulation. Recent studies have shown that the polarization status of macrophages (M1/M2) in the placental microenvironment is closely related to the development of PE [2, 3]. M1 macrophages (classically activated) have pro-inflammatory characteristics, while M2 macrophages (alternatively activated) play a key role in anti-inflammation and tissue repair [4]. The imbalance between M1/M2 macrophage subtypes leads to increased local inflammation in the decidua, which may be crucial in the pathogenesis of PE [5, 6].

Mesenchymal stem cells (MSCs) have become a potential therapeutic tool for various diseases due to their potent anti-inflammatory, tissue repair, and immune-regulatory abilities [7, 8]. Studies indicate that MSCs primarily mediate their biological functions through paracrine signaling, such as extracellular vesicles (EVs) [9]. These EVs carry bioactive molecules like proteins, lipids, and nucleic acids [10], capable of regulating the function of recipient cells and promoting macrophage polarization toward M2 [11, 12]. Apoptotic vesicles (apoVs), a novel subtype of EVs released during cell apoptosis, have garnered significant attention in recent years. A large body of literature suggests that apoptotic products from MSCs may be key to their therapeutic effects in vivo [13, 14]. MSC-derived apoVs serve as mediators of information transfer, offering advantages such as high yield, natural drug delivery capabilities, and the absence of immune rejection [15], showing great therapeutic potential. They have been widely applied in diseases such as wound healing [16], aging [17], sepsis [18], osteoporosis, and bone regeneration [19, 20]. Given that apoVs inherit MSC functions and can cross the placental barrier, we hypothesize that MSC-apoVs may regulate macrophage biological functions by carrying specific RNA molecules. Among these, microRNAs (miRNAs) are small molecules that are abundant in EVs, can survive stably in extracellular environments for long periods, and play an important role in intercellular communication [21, 22].

Although some studies have indicated that MSCs and their derived EVs have the potential to regulate macrophage polarization [11, 12], the mechanisms by which MSC-apoVs affect macrophages and the miRNA expression profiles in MSC-apoVs remain unclear. Our study found that miRNAs are highly enriched in MSC-apoVs, and that miR-191-5p can alter the polarization of macrophages, promoting M2 macrophages to secrete more PDGF-AB repair factors and improving the invasive ability of trophoblasts. Our findings reveal the potential development of a novel vesicle in PE and provide a new pathway for macrophage regulation of trophoblasts.

Materials and methods

Collection of human placental tissues and blood sample

Placental tissues were collected from patients with normal pregnancies (n = 5) and EOPE (n = 5) during cesarean sections. Patients with concurrent autoimmune diseases were excluded. Peripheral blood samples were collected from healthy pregnancy donors. Isolation and purification of untouched, highly enriched CD14⁺CD16⁻ monocytes were performed from fresh human peripheral blood mononuclear cells using the EasySep™ Human Monocyte Isolation Kit (19,359, StemCell Technologies). CD14⁺CD16⁻ monocytes were resuspended in complete culture medium and seeded into 6-well plates at a density of 2 × 10⁵ cells/mL. Human M-CSF (50 ng/mL) was added, and half of the medium was replaced every 3 days. On day 6, the macrophages were stimulated with MSC-apoVs for 24 h, followed by flow cytometry analysis to assess the expression of CD86⁺ M1 and CD206⁺ M2 markers. The study was conducted in accordance with ethical guidelines for human sample use and was approved by the Ethics Committee of the Third Affiliated Hospital of Sun Yat-sen University.

Transcriptome sequencing, single-cell RNA sequencing and Mendelian Randomization (MR) analysis of placental tissue

The preeclampsia single-cell RNA sequencing dataset GSE173193 was sourced from GEO (http://www.ncbi.nlm.nih.gov/geo/). During the quality control phase of scRNA-seq data, we filtered cells that expressed at least 200 genes and had a mitochondrial gene count proportion below 10% of the total gene count. Subsequently, we performed dimensionality reduction and clustering analysis using the Seurat package. A resolution of 0.8 was selected as the optimal value for cell clustering. Additionally, we downloaded placental marker genes from the ACT database and used the UCell tool to calculate the clustering annotation information. Furthermore, we extracted macrophages from the data and annotated the M1 and M2 macrophage subtypes based on marker genes identified in previous literature [23].

The transcriptomic dataset for preeclampsia was sourced from the GEO GSE190639 dataset. This study aimed to analyze the expression characteristics of M1 and M2 macrophages in preeclampsia. Based on previous literature [2], we selected specific cell type marker genes to define M1 and M2 macrophage subtypes. Using these marker genes, we used the Gene Set Variation Analysis (GSVA) method to score the M1 and M2 marker genes and assess their expression levels in the control group, EOPE, and LOPE.

To further analyze the causal relationship between M1 and M2 macrophages and PE, we referred to genes related to M1 and M2 macrophages collected from previous literature [23]. We extracted genetic variations within a 1000 kilobase (kb) range of the coding sequence, which were robustly associated with gene expression. These expression quantitative trait locus (eQTL) summary statistics were obtained from OpenGWAS (https://gwas.mrcieu.ac.uk/). The analysis was performed using R software, focusing on assessing the strength of genetic instruments used for MR analysis. For each genetic instrument, we calculated the F statistic to evaluate its strength and validity, and instruments were filtered based on an F statistic threshold > 10. MR analysis was conducted using the mr function to provide a comprehensive assessment of the causal relationship between exposure and outcome. We first analyzed the MR result files to identify genes with P values < 0.05 in the inverse variance weighted (IVW) method. Among these significant genes, only those with consistent odds ratios (OR) across all five MR methods (all OR > 1 or all OR < 1) were retained. We processed the pleiotropy results to exclude genes with P values < 0.05 in the pleiotropy test. Finally, genes were selected based on IVW significance thresholds, consistent OR direction in MR methods, and no significant pleiotropic effects. P value < 0.05 was considered statistically significant.

miRNA transcriptomics of MSC-apoVs, data analysis and sample preparation

After treating MSCs with 500 nM staurosporine (STS) for 6 h, the supernatant was collected, and MSC-apoVs were extracted according to the protocol outlined below. MSCs and MSC-apoVs were subjected to miRNA-seq analysis (three biological replicates per group). The miRNA-seq experiments were conducted by Shanghai Biotechnology (China). Total RNA was extracted from MSCs and MSC-apoVs using TRIzol (Invitrogen). The miRNA libraries were sequenced using Illumina next-generation sequencing technology. Sequencing data were filtered using fastx (version 0.0.13), and miRNAs were categorized and annotated using Bowtie. To ensure comparability of miRNA expression levels across different miRNAs and samples, the read counts mapped to each miRNA were normalized and transformed into TPM (Additional file Table 1). Heatmaps were generated using R software.

Cell isolation, culture and hypoxia modeling

Umbilical cord mesenchymal stem cells were isolated using the explant method. Sterile umbilical cord tissues from normal term pregnancy were collected during cesarean section. After washing with PBS and treating with 1% penicillin/streptomycin (P/S, Invitrogen), the amniotic membrane, as well as the umbilical artery and umbilical vein, were removed. The Wharton's jelly of the umbilical cord was cut into small blocks of 3–5 mm in edge length and placed in 10 cm dishes. The mesenchymal stem cell culture medium (Yocon, NC0103.S) was replaced every 3–5 days. After 12–14 days, the adherent cells were digested and MSCs were collected, cultured in mesenchymal stem cell culture medium (Yocon, NC0105.S).

Cell lines were obtained from ATCC. HTR-8/SVneo and THP-1 macrophages were cultured in RPMI 1640 medium, while JEG-3 cells were cultured in high-glucose DMEM. Both HTR-8/SVneo and JEG-3 cells were supplemented with 10% fetal bovine serum (FBS, ExCell Bio) and 1% P/S, while THP-1 macrophages were supplemented with 15% FBS and 1% P/S. To promote THP-1 macrophages adhesion, 100 ng/mL PMA (Invitrogen) was added in culture medium for 12 h before experiment.

In order to create a cellular hypoxia model, HTR-8/SVneo and JEG-3 cells were exposed to hypoxia for 8 h using 250 µM CoCl₂, and followed by reoxygenation (remove CoCl₂) for 16 h based on previous literature [24–26].

Isolation and characterization of apoVs

For apoptosis induction, the cells were washed twice with filtered PBS and basic DMEM medium (5 mL) with STS (500 nm for MSCs, Enzo Life Sciences) was added into the dishes to induce apoptosis for 6 h. Subsequently, culture supernatant was collected and centrifugated at 800 g for 10 min, 2000 g for 10 min at 4 °C to remove cell debris. Next, the supernatant was centrifuged at 16000 g for 30 min twice at 4 °C to isolate and resuspend apoVs with filtered PBS. Subsequently, apoVs were stained with CellMask Deep Red Plasma Membrane Stain (ThermoFisher) at 37 °C for 5 min and FITC Annexin V (BD Biosciences) in 1 × Annexin V binding buffer (BD Pharmingen) for 15 min at 4 °C. The total EV number generated by apoptotic cells were quantified by Nano-Flow Cytometry (U30, China). Immobilized apoVs were negatively stained and their structural features were observed under Transmission Electron Microscopy (TEM). The particle number and size of apoVs were analyzed using Nanoparticle Tracking Analysis (NTA) (Particle Metrix, Germany).

ApoVs uptake by cells in vitro and in vivo

ApoVs were labeled with PKH26 (MINI26, Sigma). Then, apoVs were isolated via centrifugation and washed twice with filtered PBS to further get rid of unbound PKH26. THP-1 macrophages incubated with PKH26-labeled apoVs 24 h and cytoplas stained by Cell Tracker Blue (ThermoFisher). Fluorescence imaging was performed by a confocal microscope (LSM 880, Zeiss, Germany) and analyzed using the ImageJ software.

PKH26-labeled apoVs were injected intravenously into C57BL/6 pregnant mice. The pregnant mice, uterus, embryos and placentas were imaged using the In Vivo Imaging System (IVIS) Spectrum (PerkinElmer) to assess the biodistribution of apoVs after 12 h.

Wound healing assay and transwell invasion assay

For the wound healing assay, HTR-8/SVneo, JEG-3 cells were cultured in 12-well plates at a density of 5 × 10⁵ cells/well. Wounds were made by scraping a 200µL pipette tip across the cell monolayer when the cells grow full. The detached cells were removed by washing with PBS. The gap area of the wound was photographed at 0 h and 24 h by optical microscope (Zeiss, Germany), and the wound width was measured using ImageJ software. The percentage of wound healed was calculated using the following formula: Wounded area filled (%) = 100% − (width after 24 h/width at the beginning) × 100%.

For invasion assays (HTR-8/SVneo and JEG-3 cells), matrigel was dissolved at 4℃ overnight and diluted 1:3 with basal medium. Diluted matrigel was added to the upper chamber of a pre-cooled transwell insert (8 µm) and incubated at 37 ℃ for 2 h. Excess matrigel was removed by gently washing with basal medium. The subsequent steps of the invasion assays were the same as those for the migration assays (THP-1 macrophages). First, medium (15% FBS + 1% P/S) was added to the lower chamber of a 24-well plate, and the transwell inserts were placed. Then medium containing 1 × 10⁵ cells (10% FBS + 1% P/S) was added to the upper chamber. After cell attachment, the related processes were applied. After 24 h, the transwell inserts were removed, and the migrated and invaded cells were fixed with 4% paraformaldehyde for 15 min. Cells on the upper side of the membrane were removed, and the remaining cells were stained with 1% crystal violet. Photographs were taken by optical microscope (Zeiss, Germany), and the number of cells in each field was counted.

The cytokines and neutralizing antibodies were added to the culture medium according to the experimental requirements: transforming growth factor-beta1 (TGF-β1) (5 ng/ml, 100-21C, PeproTech), PDGF-AB (10 ng/ml, 100-00AB, PeproTech), platelet-derived growth factors-BB (PDGF-BB) (5 ng/ml, 100-14B, PeproTech) and hPDGFR-β antibody (AF-385-SP, R&D).

Flow cytometry

For macrophage subtypes detection, THP-1 macrophages and human macrophages were stained with PE anti-human CD86 antibody (BU63, Biolegend), BV421 anti-human CD206 antibody (15–2, BioLegend) for 60 min at 4 °C and analyzed within 30 min by using flow cytometry (BECKMAN COULTER, Cyto Flex LX, USA). Placentas of mice were collected to analyze the percentage of immune cells (B cells, T cells, NK cells and macrophage subtypes). Immune cells were stained by APC-cy7 anti-mouse CD45 antibody (30-F11, Biolegend), FITC anti-mouse CD3 antibody (145-2C11, Biolegend), BV421 anti-mouse CD8a antibody (53–6.7, Biolegend), BV510 anti-mouse CD19 antibody (6D5, Biolegend), FITC anti-mouse TCR-β (H57-597, Biolegend), PE anti-mouse CD49a (HMα1, Biolegend), PE-cy7 anti-mouse CD49b (DX5, Biolegend), FITC anti-mouse CD19 antibody (6D5, Biolegend), FITC anti-mouse CD49b antibody (HMα2, Biolegend), BV510 anti-mouse CD11b antibody (M1/70, Biolegend), APC anti-mouse F4/80 antibody (BM8, Biolegend), PE-Cy7 anti-mouse CD86 antibody (GL-1, Biolegend) and PE anti-mouse CD206 (MMR) antibody (C068C2, Biolegend) for 60 min at 4 °C and analyzed within 30 min by using flow cytometry.

RNA isolation and qRT-PCR

Total RNA was extracted with Trizol reagent (Invitrogen). For qRT-PCR of mRNA, the cDNA was synthesized using the PrimeScript RT Reagent Kit (Vazyme). Then, qRT-PCR was conducted with SYBR Green Master Mix (Vazyme) and gene-specific primers. Quantification was performed by using GAPDH as the internal control and calculating the relative expression level of each gene with the 2^–ΔΔCT method. Values were expressed as fold changes. For qRT-PCR of miRNA, the cDNA was synthesized using PrimeScript RT Reagent Kit and Stem-loop RT primers. qRT-PCR was conducted with SYBR Green Master Mix, miRNA specific forward primer and the universal reverse primer. All the primer sequences were presented in Additional file Tables 2 and 3.

Animals experiments

C57BL/6 wild-type mice (5–6 weeks old) were purchased from Animal Center of Sun Yat-sen University. All animal experiments were compliant with the ethics committees of Sun Yat-sen University (SYSU-IACUC-2025–000023). Mice were maintained in a 12 h dark–light cycle at 22 ± 2 °C and 55 ± 10% humidity with food and water. Gestational ages were determined by monitoring the formation of vaginal plugs (counted as E0.5).

Pregnant mice participating in the experiment were randomized before starting treatment. Mice were divided into five groups: Control, L-NAME, L-NAME + apoVs, L-NAME + apoVs (miR-191-5p mimic) and L-NAME + apoVs (miR-191-5p inhibitor), each group had four pregnanct mice. L-NAME was administered to induce the EOPE mice model by intraperitoneal injection (60 mg/kg per day; HY-18729A, MCE) according to the protocol previously reported [27]. ApoVs was administrated to treat EOPE mice by tail vein injection (10 μg in 200 μL PBS per mouse) qod. The control mice were treated with equal volume of PBS. L-NAME was administered daily from E5.5 to E13.5 of pregnancy and apoVs on E5.5, E7.5, E9.5, E11.5, E13.5, E15.5 and E17.5.

The blood pressure of pregnant mice on E5.5, E7.5, E9.5, E11.5, E13.5, E15.5, and E17.5, was assessed noninvasively using the BP-2010 blood pressure analysis system (Softron). The urine was collected for 24 h and the total amount of urinary protein was measured using BCA protein assay kit (KGB2101, KeyGEN) on E5.5, E13.5 and E17.5 according to the experimental requirements. Pregnant mice were anesthetized by intraperitoneal injection of sodium pentobarbital (50 mg/kg) at E13.5 or E18.5 before sacrifice, and subsequently euthanized by cervical dislocation. Then the embryos and placentas were collected and weighed. The placenta tissues were fixed by 4% paraformaldehyde (PFA) for paraffin sectioning and digested for flow cytometry. This work has been reported in line with the ARRIVE 2.0 guidelines.

Enzyme-linked immunosorbent assay

The supernatant was collected from macrophages after centrifugation at 3000 rpm for 30 min at 4 °C. TGF-β1 was determined using ELISA Kit assay (88–8350, Invitrogen), while interleukin-10 (IL-10), epidermal growth factor (EGF), vascular endothelial growth factor (VEGF), PDGF were determined using ELISA Kit assay (RK04330, Ablconal) following the manufacturer’s instructions.

Histology and immunofluorescent staining

Human and mouse placental tissues were fixed with 4% PFA. The paraffin-embedded sections (5 μm) were prepared for staining. The routine procedure involves the use of xylene to remove paraffin and penetrate the tissue, followed by dehydration and removal of xylene using a gradient alcohol series. As for immunofluorescence staining of trophoblasts, removing the culture medium and fixing cells with 4% PFA. Tissue specimens and cells were blocked in 5% bovine serum albumin (BSA) for 1 h, and probed with the primary antibodies overnight at 4 °C. The primary antibodies used were as follows: anti-CK7 (ab181598, abcam), anti-CD68 (ab955, abcam), anti-F4/80 (30325S, CST), anti-CD86 (ab213045, abcam), anti-CD206 (ab64693, abcam), anti-PDGFR-β (3169S, CST). After primary antibody incubation, sections and cells were washed and incubated with secondary antibodies (K5007, DAKO) for 1 h at room temperature. For histological staining, the sections were washed, and the immune reaction was visualized using 3,3'-diaminobenzidine (DAB) (G1212-200 T, Servicebio), followed by counterstaining with hematoxylin. For immunofluorescent staining, the sections and cells were washed and incubated with the appropriate Alexa Fluor-conjugated secondary antibodies (ThermoFisher) for 1 h at room temperature. The sections and cells were washed again, and the nuclei were counterstained with DAPI. Images were captured using a confocal microscope (LSM 880, Zeiss) and analyzed with ImageJ software.

MMP measurement

THP-1 macrophages were incubated with JC-1 dye (HY-15534, MCE) at the manufacturer's recommended concentration for 30 min at 37 °C, followed by DAPI nuclear staining and imaging by confocal microscopy. A higher red-to-green fluorescence ratio indicates a greater MMP. THP-1 macrophages were stained with anti-CD206 antibody for 60 min at 4 °C, followed by incubation with TMRM dye (HY-D0984A, MCE) for 30 min at 37 °C and subsequently analyzed by flow cytometry. Increased red fluorescence intensity reflects higher MMP.

Luciferase assay

For the luciferase reporter assay, wild-type (WT-CDK6) and mutant (Mut-CDK6) dual-luciferase plasmids were constructed by QingZe Biotech Co., Ltd. (Guangzhou, China). HEK293T cells were seeded into 24-well plates at 40% confluence and transfected with 500 ng of WT-CDK6 or Mut-CDK6 plasmid alongside 50 nM miR-191-5p mimic or control mimic (GenePharma, Shanghai) using Lipofectamine 2000 (52,887, Invitrogen). At 24 h post-transfection, luciferase activity was measured with the Dual-Luciferase Reporter Assay System (E1910, Promega) following the manufacturer’s protocol. Firefly luciferase activity was normalized to Renilla luciferase activity. Data represent mean ± SD from three independent experiments.

Western blotting analysis

Cells or apoVs were harvested and lysed in RIPA (P0038, Beyotime) on ice. Total protein concentrations were measured by BCA protein assay kit (KGB2101, KeyGEN). Western blotting assay was performed according to protocols described below. Briefly, protein samples were separated by SDS–polyacrylamide gel electrophoresis gels and transferred to polyvinylidene difluoride membranes, followed by blocking with 5% BSA for 1 h at room temperature, and then incubated overnight at 4 °C with primary antibodies including anti-CD63 (ab59479, abcam), anti-CD9 (ab307085, abcam), anti-H3 (9715S, CST), anti-caspase-3 (ab184787, abcam), anti-β-tubulin (2128S, CST), anti-mTOR (2983S, CST), anti-phospho-mTOR (5536S, CST), anti-PI3K (4292S, CST), anti-phospho-PI3K (4228S, CST), anti-AKT (9272S, CST), anti-phospho-AKT (13038S, CST), anti-GAPDH (2118S, CST). After washing with Tris buffered saline (TBS) containing 0.1% Tween-20, the membranes were incubated with species-related peroxidase-conjugated secondary antibodies for 1 h at room temperature. The protein bands were detected using Signal Fire Elite ECL Reagent (12,757, CST) and imaging system (Tanon 5200CE).

Statistics

All the experimental data were expressed as means ± SD. For normally distributed data, comparisons between two groups were analyzed using independent unpaired two-tailed Student’s t tests, and comparisons between more than two groups were analyzed using one-way ANOVA with the Bonferroni adjustment. Other data were compared using the Mann–Whitney U test. P value < 0.05 was considered statistically significant. All other statistical analyses were conducted using GraphPad Prism 9.5.1.

Results

M2 macrophage proportion is decreased in human EOPE placenta

We first annotated and clustered the single-cell sequencing data of placental tissues from normal pregnancies and PE cases, resulting in 17 distinct cell subpopulations (Fig. 1a). We then analyzed the macrophage subpopulations and performed secondary clustering based on M1 and M2 marker genes (Fig. 1b). The bar chart showing proportions revealed a significant reduction in M2 macrophages in the PE group, while the proportion of M1 macrophages showed no statistically significant difference (Fig. 1c). To further elucidate the causal relationship between macrophages and PE, we conducted MR analysis using M1-related genes and M2-related genes as exposure factors and PE as the outcome. The results showed that the OR for M2 macrophage-related genes were all less than 1, with an inverse variance weighted P value < 0.05 (Fig. 1d), indicating a negative correlation between M2 macrophage-related genes and PE (Additional file1: Fig. S1A). In contrast, there was no statistically significant association between M1 macrophage-related genes and PE (Additional file1: Fig. S1B-C), suggesting that the reduction of M2 macrophages may be an exposure factor in PE. To explore whether the change in M2 macrophage proportion differs between EOPE and LOPE, we further analyzed the transcriptome sequencing data of placental tissues. We found that the significant reduction in M2 macrophages was mainly observed in EOPE, rather than in LOPE (Fig. 1e–f). Finally, we validated these findings using human EOPE placental tissues. Consistent with the bioinformatics analysis, the proportion of M2 macrophages was significantly decreased in the EOPE group, while there was no significant difference in M1 macrophages (Fig. 1g–j). These data strongly suggest that placental M2 macrophage polarization is closely associated with the development of PE, particularly EOPE. Therefore, we next aim to investigate whether MSC-apoVs can improve EOPE by inducing M2 macrophage polarization.

Fig. 1.

Differential expression of macrophages in placental tissues from normal pregnancy and preeclampsia (PE). a UMAP plot of single-cell RNA sequencing showed the dimensional distribution of different cell types in placental tissues. b Marking of M1 and M2 related genes in annotated macrophages. c Bar chart depicted the proportion of M1 macrophages and M2 macrophages. d Mendelian randomization analysis assessed the causal relationship between M2 macrophage-related genes and the development of PE. e, f Macrophage scores for M1 macrophages and M2 macrophages in normal pregnancy, EOPE, and LOPE groups. g, h The proportion of CD86⁺ M1 macrophages among total macrophages in placental tissues (n = 4). Scale bar, 20 μm. i, j The proportion of CD206⁺ M2 macrophages among total macrophages in placental tissues (n = 4). Scale bar, 20 μm. Error bars represent means ± SD. NS, not significant. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001

Induction of MSC apoptosis, extraction and characterization of MSC-apoVs

We isolated MSCs from human umbilical cord tissue and cultured them. Apoptosis was induced by STS. After apoptosis induction, the supernatant was collected, and cell debris was removed by low-speed centrifugation. MSC-apoVs were then precipitated by high-speed centrifugation and resuspended in PBS to obtain MSC-apoVs (Fig. 2a). To assess the multi-lineage differentiation potential of MSCs, we cultured MSCs in osteogenic and adipogenic media, resulting in the formation of distinct calcium salt crystals and lipid droplets (Fig. 2b). After STS-induced apoptosis, MSCs exhibited cell shrinkage, nuclear chromatin condensation, and membrane blebbing (Fig. 2c). TUNEL and DAPI staining revealed that the cell nuclei were double-positive (Fig. 2d–e). TEM images confirmed that MSC-apoVs exhibited a lipid bilayer structure (Fig. 2f). NTA showed that the particle concentration of MSC-apoVs was as high as 10¹¹ particles/ml, with an average diameter of 180 nm (Fig. 2g). Western blot analysis revealed high expression of the tetraspanin proteins CD63 and CD9 in MSC-apoVs, while the nuclear protein H3 was absent, confirming the EV structure of MSC-apoVs. The presence of phosphatidylserine externalization on the MSC-apoV membrane and the positive staining for cleaved caspase-3 further indicated that MSC-apoVs were apoptotic vesicles (Fig. 2h–i). These data verify the specific characteristics of MSC-derived apoVs.

Fig. 2.

Characterization of apoVs derived from STS-treated MSCs. a Schematic representation of the extraction process for apoVs. b Osteogenic differentiation of MSCs (Alizarin Red staining) and adipogenic differentiation (Oil Red O staining) capacity. Scale bar, 50 μm. c Bright-field images of MSCs before and after induction of apoptosis. Scale bar, 20 μm. d–e TUNEL (green) and DAPI (blue) fluorescence staining of MSCs following treatment with STS, along with quantitative analysis. Scale bar, 10 μm. f TEM image of apoVs. Scale bar, 100 nm. g NTA of apoVs, showing particle concentration and size distribution. h Nano-flow cytometry analysis of apoptosis in apoVs. i Western blot analysis of exosome markers (CD63, CD9), a negative EV marker (H3), and cleaved caspase-3 as an indicator of apoptosis in apoVs. Full-length blots were presented in Supplementary Figure S6. Error bars are means ± SD. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001

MSC-apoVs recruit and promote M2 polarization in macrophages, lower blood pressure, urinary protein, and improve embryo quality in EOPE pregnant mice

To investigate whether MSC-apoVs promote M2 polarization in macrophages, we first tested the affinity of macrophages for apoVs. After adding MSC-apoVs to macrophages, confocal microscopy showed that macrophages had a strong phagocytic ability for MSC-apoVs. After 24 h, the cytoplasm was filled with PKH26-labeled MSC-apoVs (Fig. 3a). Transwell assays indicated a significant migration of macrophages toward the lower chamber containing MSC-apoVs (Fig. 3b, c). Flow cytometry analysis revealed that, after co-culturing THP-1 macrophages with MSC-apoVs for 24 h, the proportion of CD206⁺ M2 macrophages significantly increased, while the proportion of CD86⁺ M1 macrophages remained unchanged (Fig. 3d, e, Additional file2: Fig. S2A). The results from human peripheral blood monocytes stimulated with MSC-apoVs were consistent with those from the THP-1 (Additional file2: Fig. S2B-C). qPCR showed no change in the mRNA levels of M1-related genes (TNF-α, iNOS), while the mRNA levels of M2-related genes (MRC1, Arg1) were significantly elevated. These results were consistent with the flow cytometry data (Additional file2: Fig. S2D-G). These findings suggest that macrophages can chemotax toward MSC-apoVs and efficiently phagocytize them, leading to M2 polarization upon engulfment of MSC-apoVs.

Fig. 3.

MSC-apoVs promoted M2 polarization of macrophages and improved symptoms in EOPE pregnant mice. a Representative confocal microscopy images showed PKH26-labeled apoVs (red) uptake by macrophages, with cytoplasm stained by Cell Tracker Blue (blue). Scale bar, 5 μm. b Schematic diagram of the co-culture setup using transwell between macrophages and apoVs. c Effect of apoVs on macrophages migration ability. Scale bar, 100 μm. d, e Flow cytometry analysis showed the percentages of CD86⁺ M1 macrophages and CD206⁺ M2 macrophages after co-culturing with apoVs. f Schematic diagram of the EOPE mouse model establishment (①) and apoVs intervention (②). g Representative images of embryos at E13.5 from the EOPE model, with red arrows indicating dead embryos (n = 4). h–j Blood pressure, urine protein levels, and embryo survival rate of E13.5 pregnant mice after L-NAME-induced EOPE. k In vivo imaging of PKH26-labeled apoVs (red) injected into pregnant mice 12 h post-injection, showing fluorescence intensity of the mice, uterus, and embryos. l Representative images of the uterus, embryos and placentas (red arrows indicate dead or resorbed embryos) from each group (n = 4). m–n Blood pressure and urine protein levels in pregnant mice from each group. o–q Embryo survival rate, embryo weight, and placenta weight in each group (E18.5). r–t Flow cytometry analysis and quantification of CD206⁺ M2 macrophages and CD86.⁺ M1 macrophages in the placenta from each group. Error bars are means ± SD. NS, not significant. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001

In our previous experiments, we observed a marked reduction in M2 macrophage proportions in EOPE, while MSC-apoVs promoted M2 polarization in macrophages. Therefore, we further validated the polarization effect and therapeutic potential of MSC-apoVs in EOPE mice model. We first induced EOPE-like symptoms in pregnant mice by intraperitoneal injection of L-NAME (Fig. 3f①). Macroscopic examination of the uterus and embryos at E13.5 showed that L-NAME injection resulted in slow embryo growth (Fig. 3g). Blood pressure and urinary protein levels in pregnant mice were significantly elevated (Fig. 3h, i), accompanied by reduced embryo and placental weights (Additional file2: Fig. S2H-I) and increased embryo mortality (Fig. 3j). These results indicate that our EOPE model in C57BL/6 pregnant mice was successfully established.

To investigate the effect of MSC-apoVs on placental macrophages in vivo, we first assessed whether MSC-apoVs could cross the placental barrier. MSC-apoVs were labeled with PKH26 (red), and then injected via tail vein into E16.5 pregnant mice. After 12 h, in vivo imaging showed strong PKH26⁺ fluorescence signals in the abdomen, uterus, and placenta of pregnant mice (Fig. 3k), indicating that MSC-apoVs can cross the placental barrier and enter the placental tissue. Next, we injected MSC-apoVs via tail vein in the EOPE mice model (Fig. 3f②). At E18.5, macroscopic examination and data showed that MSC-apoVs promoted weight gain in both EOPE pregnant mice and their embryos (Fig. 3l, Additional file2: Fig. S2J), reduced blood pressure and urinary protein in pregnant mice (Fig. 3m, n), and improved embryo weight and survival rate (Fig. 3o–q). Furthermore, flow cytometry analysis of the E18.5 placental tissue showed no statistically significant differences in total B cells (Additional file3: Fig. S3A-B), total T cells (Additional file3: Fig. S3C-D), and NK cells (Additional file3: Fig. S3E-F) among the groups. Further analysis of M1 and M2 macrophages (Additional file3: Fig. S3G) demonstrated a significant decrease in CD206⁺ M2 macrophages in the EOPE group, whereas MSC-apoVs injection resulted in a marked recovery of CD206⁺ M2 macrophages, with no effect on CD86⁺ M1 macrophages (Fig. 3r–t). These data indicate that MSC-apoVs can cross the placental barrier and improve EOPE by promoting M2 polarization in macrophages.

MSC-apoVs promote M2 polarization in macrophages through high expression of miR-191-5p and downregulation of CDK6 in vitro

In previous studies, we have confirmed that MSC-apoVs can promote M2 polarization in macrophages both in vitro and in vivo. Given that MSC-apoVs serve as carriers of information transmission, containing a large number of small molecules such as miRNAs, we hypothesized that MSC-apoVs may regulate macrophage polarization through the delivery of miRNAs. We performed miRNA transcriptome sequencing on MSCs and MSC-apoVs and found that the top 10 most highly expressed miRNA genes in MSC-apoVs were mostly consistent with those in MSCs, with seven miRNA genes being expressed at significantly higher levels in MSC-apoVs compared to their corresponding MSCs (Fig. 4a). Among these, miR-191-5p showed the most significant increase in MSC-apoVs (Fig. 4b). This result suggests that upon apoptosis induction, MSCs release MSC-apoVs containing more miRNAs for signaling transmission, further validating the parent-specific imprinting characteristics of MSC-apoVs and revealing miRNA enrichment. To identify the key miRNAs involved, we selected three statistically significant miRNAs (miR-191-5p, miR-125b-5p, miR-125a-5p) in MSC-apoVs relative to MSCs and performed experimental screening. We transfected MSCs with miRNA mimics and subsequently obtained the corresponding MSC-apoVs to co-culture with macrophages. The results showed that none of the above apoVs affected M1 polarization of macrophages (Additional file4: Fig. S4A-B), while apoVs containing miR-191-5p mimic significantly promoted macrophage polarization toward CD206⁺ M2, whereas apoVs containing miR-125b-5p mimic and miR-125a-5p mimic did not show a significant effect on M2 polarization (Fig. 4c, d).

Fig. 4.

MSC-apoVs miRNA sequencing analysis and its effect on macrophages. a Heatmap showed the differential expression of miRNAs between MSC-apoVs and MSCs. b qPCR validation of the 7 miRNAs that were highly expressed in MSC-apoVs. c, d Flow cytometry analysis assessed the effect of apoVs on the polarization of CD206⁺ M2 macrophages after MSC transfection with miRNA-191-5p mimic, miRNA-125b-5p mimic, or miRNA-125a-5p mimic, along with corresponding bar charts. e, f Transwell assay showed the effect of apoVs on macrophages migration. Scale bar, 100 μm. g, h Flow cytometry analysis showed the percentages of CD206⁺ M2 macrophages among macrophages from each group. i, j Flow cytometry analysis showed the percentages of CD86.⁺ M1 macrophages among macrophages from each group. k JC-1 fluorescence staining and quantification of THP-1 macrophages treated with apoVs. Scale bar, 5 μm. l Flow cytometry analysis of TMRM fluorescence intensity in M2 macrophages from THP-1 macrophages treated with apoVs, with corresponding quantification. m Venn diagram illustrating the intersection of miR-191-5p target genes and MMP-related genes. n–o Luciferase reporter assays demonstrating the binding interaction between CDK6 and miR-191-5p. p Western blot analysis revealed CDK6 protein expression levels in each group. Full-length blots were presented in Supplementary Figure S6. q JC-1 fluorescence staining of THP-1 macrophages across different treatment groups. Scale bar, 5 μm. r Flow cytometry analysis of TMRM fluorescence intensity in M2 macrophages from THP-1 macrophages across different treatment groups, with corresponding quantification. Error bars represent means ± SD. NS, not significant. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001

To further confirm the role of miR-191-5p in MSC-apoVs, we transfected MSCs with miR-191-5p mimic and miR-191-5p inhibitor to obtain the corresponding apoVs, and then treated macrophages. Transwell assays and flow cytometry analysis revealed that apoVs with miR-191-5p mimic accelerated macrophage aggregation and enhanced phagocytosis of apoVs (Fig. 4e, f) as well as their ability to convert into CD206⁺ M2 macrophages (Fig. 4g, h), with no significant effect on CD86⁺ M1 macrophages (Fig. 4i j). In contrast, apoVs derived from miR-191-5p inhibitor-transfected MSCs, by inhibiting the binding of miR-191-5p to its downstream target genes, resulted in a significant decrease in macrophage aggregation and M2 polarization ability (Fig. 4g–j). Therefore, we preliminarily conclude that miR-191-5p may be a key miRNA in MSC-apoVs that promotes macrophage polarization toward M2.

Mitochondrial function was closely associated with macrophage polarization. Stimulation with MSC-apoVs enhanced MMP in macrophages (Fig. 4k), particularly in M2 macrophages (Fig. 4l). Bioinformatic analysis predicted downstream target genes of miR-191-5p, which were intersected with MMP datasets (Fig. 4m). Notably, suppression of CDK6 increased mitochondrial quantity and ATP accumulation [28]. Luciferase reporter assays confirmed that CDK6 strongly bound to miR-191-5p (Fig. 4n–o), suggesting CDK6 as a critical target of miR-191-5p. Overexpression of CDK6 in THP-1 macrophages (Additional file4: Fig. S4C-D) significantly reduced MMP (Additional file4: Fig. S4E) and markedly decreased TMRM fluorescence intensity in M2 macrophages (Additional file4: Fig. S4F). Further functional rescue experiments revealed that MSC-apoVs combined with miR-191-5p mimic downregulated CDK6 protein expression and reversed the CDK6 overexpression-induced reduction in macrophage MMP (Fig. 4P-R, Additional file4: Fig. S4G). These findings indicated that miR-191-5p promoted M2 macrophage polarization by downregulating CDK6 protein expression, thereby elevating MMP.

In vivo validation of MSC-apoVs promoting macrophage M2 polarization through high expression of miR-191-5p and improving EOPE symptoms in pregnant mice

In vitro, we have already verified the effect of miR-191-5p on macrophage polarization. Next, we aim to further confirm this effect in vivo. Using the previously described method (Fig. 3f), we induced EOPE and injected MSC-apoVs with different treatments. Gross observations at E18.5 revealed that the apoVs (miR-191-5p mimic) group showed the most significant improvement in EOPE symptoms, with embryos being notably closer in size to the control group, while the apoVs (miR-191-5p inhibitor) group showed an effect between the L-NAME group and the L-NAME + apoVs group (Fig. 5a). Compared to the apoVs group, the apoVs (miR-191-5p mimic) group showed comprehensive improvement in blood pressure, urine protein, embryo survival rate, as well as embryo and placental weights in EOPE pregnant mice (Fig. 5b–f). Although the apoVs (miR-191-5p inhibitor) group did not completely block the effect of apoVs on fetal survival, it significantly inhibited the improvements in blood pressure, urine protein, and placental weight (Fig. 5b–f). Placental dysfunction is a root cause of PE, and trophoblast invasion into the uterine decidua is a key indicator of placental function [29]. Therefore, we focused on the ability of trophoblasts at the placental interface to invade the uterine decidua. Immunohistochemical staining results showed that trophoblasts in the control group invaded the decidua the furthest, followed by the apoVs (miR-191-5p mimic) and apoVs groups in the EOPE model, while the L-NAME group showed the shortest invasion distance (Fig. 5g–h). These data suggest that MSC-apoVs, particularly those enriched with miR-191-5p, promote trophoblast invasion at the placental interface.

Fig. 5.

Animal experiments verified that miR-191-5p promoted M2 polarization of macrophages and improved symptoms in EOPE pregnancy mice. a Representative images of the uterus, embryos size, and embryo death in mice from each group (red arrows indicate dead and resorbed embryos) (n = 4). b, c Blood pressure and urine protein levels in pregnant mice from each group. d–f Embryo survival rate, embryo weight, and placenta weight in pregnant mice from each group. g Immunohistochemistry staining showed the invasion distance of trophoblasts (CK7⁺) into the uterine decidua at the embryo interface in each group. Scale bar, 200 μm. h Statistical analysis of CK7 invasion distance in each group. i, j Flow cytometry analysis and quantification of CD206⁺ M2 macrophages among macrophages from each group in the placenta. k, l Representative confocal images and quantification showed F4/80⁺ macrophage cells (red), as well as the percentages of CD206.⁺ M2 macrophages (green) in the placenta from each group, counterstained with DAPI (blue). Scale bars, 20 μm. Error bars represent means ± SD. NS, not significant. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001

Flow cytometry analysis of the E18.5 placenta revealed that apoVs (miR-191-5p mimic) significantly promoted the polarization of placental macrophages toward CD206⁺ M2 macrophages, with a clear increase in the proportion of CD206⁺ M2 macrophages. The trends for the other groups were consistent with the previous cell experiments and trophoblast invasion data (Fig. 5i–j), with no significant effect on CD86⁺ M1 macrophages (Additional file4: Fig. S4H-I). We also performed immunofluorescence staining on the placenta, which corroborated the flow cytometry results. ApoVs (miR-191-5p mimic) promoted M2 macrophage polarization and enhanced the therapeutic effect of MSC-apoVs, while the use of miR-191-5p inhibitor significantly weakened the therapeutic effect of MSC-apoVs (Fig. 5k, l).

M2 macrophages promote trophoblast invasion by secreting PDGF-AB, activating the PI3K-AKT-mTOR signaling pathway

Our in vitro and in vivo experiments have shown that apoVs can deliver high levels of miR-191-5p, which is engulfed by macrophages and induces their polarization into the M2 phenotype. The increased number of M2 macrophages, in turn, enhances the invasion capacity of trophoblasts, significantly improving the symptoms and prognosis of EOPE in pregnant mice. However, how polarized M2 macrophages influence the invasion ability of trophoblasts will be the focus of our next research. M2 macrophages mainly regulate tissue inflammation through the secretion of cytokines. Therefore, we first performed ELISA to detect the cytokines secreted by polarized macrophages, including anti-inflammatory cytokines (TGF-β1, IL-10) and tissue repair and angiogenesis-related cytokines (EGF, VEGF, PDGF). The regulation of vascular function and the migratory invasive effects of PDGF factors on cells are mainly accomplished through the β-subunit [30, 31], so we chose to measure PDGF-AB and PDGF-BB. ELISA results indicated that the levels of IL-10 and EGF were lower in all groups, and the levels of VEGF and IL-10 in the macrophage supernatant decreased after co-culturing with apoVs. In contrast, the expression of TGF-β1 and PDGF-AB in the macrophage supernatant increased significantly after treatment with apoVs (Fig. 6a). We then selected TGF-β1, PDGF-AB, and PDGF-BB as candidate cytokines for further validation. We hypothesize that these cytokines may exert their effects by directly acting on trophoblasts. Therefore, we added TGF-β1, PDGF-AB and PDGF-BB cytokines to hypoxic trophoblasts (CoCl₂ group). Scratch wound healing assays showed that CoCl₂-treated trophoblasts had significantly reduced wound healing ability, both PDGF-AB and PDGF-BB cytokines promote trophoblasts migration, thereby facilitating the repair of the scratched wound (Fig. 6b–e). Transwell assay results indicate that PDGF-AB and PDGF-BB significantly enhance trophoblasts invasion (Fig. 6f–h). Moreover, we observed that a higher concentration of PDGF-AB promotes trophoblasts invasion more effectively than the relatively lower concentration of PDGF-BB. Therefore, we hypothesize that the key factor regulating trophoblasts invasion may be PDGF-AB, which is significantly elevated after macrophage M2 polarization, rather than the relatively lower concentration of PDGF-BB.

Fig. 6.

The effect of cytokines secreted by macrophages (M) on trophoblasts. a The heatmap showed the ELISA results of cytokine concentration (pg/ml) in apoVs and the supernatant of macrophages (n = 3), * indicates the comparison between macrophages and macrophages + apoVs. b, c Scratch assay assessed the effects of TGF-β1, PDGF-BB and PDGF-AB cytokines on the migration ability of trophoblasts treated with CoCl₂. Scale bars, 20 μm. d, e Statistical analysis of wound healing in the scratch assays for trophoblasts. f Transwell assay evaluated the effects of TGF-β1, PDGF-BB and PDGF-AB cytokines on the invasion ability of CoCl₂-treated trophoblasts. Scale bars, 100 μm. g, h Statistical analysis of the number of invaded cells for trophoblasts. i Immunofluorescence images showed PDGFR-β (green) expression in trophoblasts, counterstained with DAPI (blue). Scale bars, 5 μm. j Schematic representation of the co-culture of macrophages and trophoblasts after apoVs treatment for 24 h, where the macrophages referred to apoVs-treated macrophages. k, l Scratch assay assessed the effect of PDGFR-β-nAb on reversing the impact of apoVs-treated macrophages on the migration ability of hypoxic trophoblasts. Scale bars, 20 μm. m Statistical analysis of wound healing in the scratch assay. n, p Transwell assay assessed the effect of PDGFR-β-nAb on reversing the impact of apoVs-treated macrophages on the invasion ability of hypoxic trophoblasts. Scale bars, 100 μm. q Western blot analysis showed the protein expression levels of the PI3K-AKT-mTOR signaling pathway in trophoblasts. Full-length blots were presented in Supplementary Figure S6. Error bars represent means ± SD. NS, not significant. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001

We further examined the expression of the PDGF-β receptors on trophoblasts. Immunofluorescence results showed that both HTR-8/SVneo and JEG-3 trophoblasts expressed abundant PDGFR-β receptors on their cell membrane and cytoplasm (Fig. 6i). Additionally, PDGF-β receptors expression was also high in CK7⁺ trophoblasts in both human and mouse placental tissues (Additional file4: Fig. S4 J-K). Next, we attempted to block the binding of cytokines to their receptors using neutralizing antibodies. We first treated macrophages with apoVs for 24 h, then co-cultured the polarized macrophages with hypoxic trophoblasts (Fig. 6j). The results showed that the addition of PDGFR-β neutralizing antibodies inhibited the migration ability of trophoblasts, significantly reducing wound healing capacity (Fig. 6k–m), and also significantly decreased the invasion ability of trophoblasts toward the lower chamber (Fig. 6n–p). Western blotting analysis of proteins from co-cultured trophoblasts revealed activation of the PI3K-AKT-mTOR signaling pathway. However, following PDGF-PDGFR binding, the use of AG1296 to inhibit downstream tyrosine kinase activity of PDGFR substantially reduced trophoblast migration and invasion capabilities. Furthermore, the combined use of the PI3K inhibitor LY294002 further mitigated the effects of PDGF-AB secreted by polarized macrophages on trophoblasts (Additional file5: Fig. S5A-G). Western blot analysis of proteins provided additional confirmation of these findings (Additional file5: Fig. S5H).

Therefore, we propose that apoVs derived from MSCs are enriched with high levels of miR-191-5p. Upon uptake by macrophages, miR-191-5p downregulates CDK6 protein expression, increases MMP, and promotes M2 macrophage polarization. Polarized M2 macrophages secrete elevated levels of PDGF-AB, which binds to PDGFR-β receptors on trophoblasts, directly activating the PI3K-AKT-mTOR signaling pathway and promoting the migration and invasion of trophoblasts (Fig. 7).

Fig. 7.

Schematic representation of the main findings of this study. Apoptosis was induced in MSCs to produce apoVs, which were endocytosed by macrophages. The vesicles released miR-191-5p to suppress CDK6 protein expression and promoted M2 phenotype polarization by enhancing MMP. Polarized M2 macrophages released PDGF-AB, which bound to PDGF-R on trophoblasts, directly activating the intracellular PI3K-AKT-mTOR signaling pathway, promoting migration and invasion, and improving PE symptoms and prognosis

Discussion

EOPE accounts for approximately 12% of PE cases [32], but it is more likely to be associated with fetal growth restriction and adverse pregnancy outcomes [33]. Currently, no specific therapeutic agents are available, and pregnancy termination remains the most effective treatment [1], although the associated neonatal complications are a significant concern. The pathogenesis of EOPE is complex, with insufficient trophoblast invasion during early pregnancy leading to impaired remodeling of the uterine spiral arteries and placental ischemia and hypoxia, which are considered key factors in the development of EOPE [34]. The powerful anti-inflammatory, tissue repair, and immune regulatory functions of MSC-apoVs align with the pathological factors involved in EOPE. Therefore, exploring the effects of MSC-apoVs on EOPE is crucial for developing effective therapeutic strategies. Our study revealed that MSC-apoVs promote macrophage chemotaxis and M2 polarization by transferring miRNA signals, enhance trophoblasts migration and invasion, and alleviate EOPE symptoms. This further underscores the important role of miRNA molecules in the signaling functions of MSC-apoVs. We also found that macrophages regulate trophoblasts directly by secreting PDGF-AB, which facilitates a more comprehensive understanding of the pathogenesis of EOPE and the development of new therapeutic strategies.

MSCs play a regulatory role in inflammation and immune diseases [7, 8]. However, many studies have shown that less than 5% of MSCs injected into animals reach the disease site within a few hours, with most undergoing rapid apoptosis within 24 h [14]. Therefore, the mechanism of MSCs in disease therapy may be mediated through the release of apoptotic products. Apoptotic signals and products undoubtedly stimulate the immune regulatory mechanisms in the body to maintain host immune balance [35, 36]. In fact, previous studies have shown that MSC-apoVs can regulate macrophage polarization to the M2 phenotype, promoting skin wound healing and improving type 2 diabetes [16, 37]. Consistent with this, our results showed that MSC-apoVs can attract macrophages, which then engulf and polarize to the M2 anti-inflammatory phenotype. Surprisingly, MSC-apoVs did not enhance macrophage M1 polarization, providing strong evidence for their role in treating inflammatory diseases. Bioinformatic analysis revealed a significant decrease in M2 macrophage levels in PE, particularly in EOPE rather than LOPE. Mendelian randomization analysis further suggested that the reduction of M2 macrophage-related genes could be an exposure factor for the onset of PE. Thus, MSC-apoVs, which selectively induce macrophage polarization to the M2 phenotype, may be an effective tool for treating EOPE.

ApoVs, a type of apoptotic EV, are widely present in the body and play a crucial role in maintaining normal physiological states [17, 38, 39]. Due to their small particle size, EVs can carry bioactive molecules, such as proteins and miRNAs, that travel to distant organs and cross barriers such as the placental barriers and blood–brain barriers to deliver signals [40, 41]. MSC-EVs have shown therapeutic potential in inflammation, tumors, immune diseases, and tissue regeneration [42]. miRNA sequencing of MSCs and their secreted apoVs revealed that the top 10 miRNAs in apoVs have expression levels largely consistent with those in MSCs. Interestingly, we found that the expression of most miRNAs in MSC-apoVs was higher than in MSCs, indicating that under stressful conditions (such as apoptosis), cells transmit more signals to neighboring cells or distant organs to maintain body homeostasis [43, 44], a process of active information exchange from dead to living cells via apoVs. Our experimental results confirmed that MSC-apoVs could cross the placental barrier and reach the placental tissue, and we identified miR-191-5p as a key miRNA in MSC-apoVs that promotes M2 macrophage polarization in the placenta. The miR-191 family plays an important role in tumor progression, metabolic regulation, and reproductive biology [45, 46] and has been linked to metastasis in colorectal cancer, intrahepatic cholangiocarcinoma, and breast cancer [47–49]. Recent studies have shown that EVs derived from breast cancer cells carrying miR-191-5p stimulate macrophage M2 polarization and mediate communication between tumors and macrophages [50]. Our study found that MSC-apoVs carrying miR-191-5p similarly reshape macrophage M2 activation and promote trophoblast cell invasion.

MMP and oxidative phosphorylation are closely associated with macrophage polarization [51, 52]. Mitochondrial oxidative phosphorylation promotes macrophage polarization toward the M2 phenotype [53]. We also found that MSC-apoVs significantly enhance the MMP of macrophages, particularly M2 macrophages. Further experiments confirmed that miR-191-5p (especially mimic transfection) downregulates CDK6 protein expression in macrophages, significantly increasing MMP. CDK6, a cyclin-dependent kinase, when inhibited, leads to ATP and mitochondrial accumulation, stimulating oxidative phosphorylation and enhancing metabolism [28]. Mechanistic studies confirmed that MSC-apoVs, through miR-191-5p-mediated downregulation of CDK6 protein expression in macrophages, increase MMP, promoting polarization toward the M2 phenotype.

Macrophages not only regulate other immune cells through the secretion of cytokines to improve the immune microenvironment [54, 55] but can also directly affect non-immune target cells by influencing downstream gene expression [56]. Our results showed that the addition of PDGF-AB cytokine directly promotes trophoblasts migration and invasion, similar to the effects observed when M2-polarized macrophages are co-cultured with MSC-apoVs. This suggests that cytokines exert their effects through multiple pathways. As an important repair factor, PDGF is widely involved in cell metabolism, apoptosis, and regeneration [57, 58], as well as in the regulation of trophoblasts growth [59]. The PDGFR-β receptors, highly expressed in trophoblasts, can specifically bind to the B chain of PDGF-AB. Upon phosphorylation, the PDGFR binds to the SH2 domain-containing PI3K subunit p85 and p110, forming a complex that subsequently activates downstream phosphorylation, thereby directly stimulating the PI3K-AKT-mTOR signaling pathway [60]. This may represent a new mechanism by which M2 macrophages directly promote trophoblast invasion. Activation of the PI3K-AKT-mTOR pathway is widely involved in regulating various physiological processes, including cell growth, proliferation, metabolism, and migration [61, 62]. Research has shown that the efficacy of PDGF-AB binding to PDGFR-β receptors is weaker than that of PDGF-BB [63]. However, our study found that the secretion level of PDGF-AB and the amount induced by vesicles in M2 macrophages are significantly higher than those of PDGF-BB. Moreover, blocking PDGFR-β receptors and inhibiting PDGFR phosphorylation significantly weakened the ability of M2 macrophages to promote trophoblasts invasion. Therefore, we believe that the elevated PDGF-AB cytokine is the key driver of trophoblasts invasion. The fact that the PDGF-AB induced by M2 macrophages differs from the PDGF-BB, which has the strongest binding capacity to PDGFR-β, by a single B chain is an intriguing phenomenon. This raises important questions about how PDGFR receptors (including the α receptors) are involved in the mechanistic regulation of trophoblasts, which we will address in future studies. Although we found that MSC-apoVs have no significant effect on the number of total B cells, total T cells, and NK cells in the placenta, cytokines secreted by macrophages may influence the classification and function of other immune cell subtypes, However, this novel mechanism of PDGF-AB directly acting on trophoblasts enhances our comprehensive understanding of macrophage functions as therapeutic targets for early-onset preeclampsia (EOPE).

Conclusion

In summary, our study emphasizes the role of macrophage M2 polarization in the pathogenesis of EOPE and identifies miR-191-5p as a key molecule in MSC-apoVs that promotes M2 polarization. Macrophages can also directly influence trophoblasts through cytokine-receptor binding. These findings provide a theoretical basis for a new therapeutic approach to EOPE.

Abbreviations

PE

Preeclampsia

EOPE

Early-onset preeclampsia

LOPE

Late-onset preeclampsia

MSCs

Mesenchymal stem cells

ApoVs

Apoptotic vesicles

MSC-apoVs

Mesenchymal stem cell apoptotic vesicles

EVs

Extracellular vesicles

miRNAs

MicroRNAs

M

Macrophages

PDGF-AB

Platelet-derived growth factor-AB

PDGF-BB

Platelet-derived growth factor-BB

PGFR-β

Platelet-derived growth factor receptor-beta

TGF-β1

Transforming growth factor-beta1

IL-10

Interleukin-10

EGF

Epidermal growth factor

VEGF

Vascular endothelial growth factor

MMP

Mitochondrial membrane potential

nAb

PDGF-β-nAb

STS

Staurosporine

TEM

Transmission electron microscopy

NTA

Nanoparticle tracking analysis

MR

Mendelian randomization

GSVA

Gene set variation analysis

OR

Odds ratios

IVIS

In vivo imaging system

BSA

Bovine serum albumin

Author contributions

LL contributed to the study design, implementation, data collection, figure creation, and drafting of the manuscript. XL was involved in figure creation and animal experiments. QL assisted with cell experiments and data analysis. XW was involved in data analysis and figure creation. CJ designed the experiments, performed the animal experiments, and handled the manuscript formatting. CX designed the experiments, supervised result collection and data interpretation, and revised the final manuscript. All authors have read and approved the final version.

Funding

This study was supported by the National Natural Science Foundation of China (82070606) and the Natural Science Foundation of Guangdong Province (2024A1515010538).

Data availability

The miRNA sequencing data generated by this study are available from the corresponding author or first anthor on reasonable request. All data generated or analyzed during this study are included in this published article and its supplementary files.

Declarations

Ethics approval and consent to participate

The clinical study was conducted in accordance with the Declaration of Helsinki and approved by the Ethics Committee of the Third Affiliated Hospital of Sun Yat-sen University (Protocol code: II2024-138–02, Data of Approval: 20 May 2024). Title:"Mechanistic study of HucMSC apoptotic vesicles improving preeclampsia by modulating mitochondrial function."Written informed consent was obtained from all patients prior to enrollment. The animal study protocol was approved by the Ethics Committee of Sun Yat-sen University (protocol code: SYSU-IACUC-2025-000023, Data of Approval: 7 January 2025), and the approved project title is"Mechanistic study of MSC apoptotic vesicles improving preeclampsia by modulating mitochondrial function."

Consent for publication

All authors confrm their consent for publication.

Competing interests

The authors declare no conflict of interest.