The collection and use of human tissue samples were approved by the Human Research Ethics Committee of the Obstetrics and Gynecology Hospital, Fudan University, and followed the principles of the Helsinki Declaration (0423-10-HMO). Written informed consent was obtained from all women in this study. Endometrial tissues (n=20) were collected from patients with leiomyomas during hysterectomy. Patients aged between 25 and 40 years with regular menstrual cycles who had not received hormone therapy or took any medications were considered for this study. Normal decidua, villous tissues, and peripheral blood were obtained from women with clinically normal pregnancies (NPs; terminated for nonmedical reasons, gestational age: 6-10 weeks). All women in this study who had NPs never had spontaneous miscarriage and had at least one live birth before the current pregnancy. Decidual tissues and peripheral blood were collected from patients with URPL who received uterine curettage because of a lack of fetal heartbeat detected by ultrasound at 6-10 weeks of gestation. All URPL subjects had regular ovulatory menstrual cycles and a history of two or more consecutive miscarriages, excluding genetic, anatomic, or endocrine abnormalities; infections; immune disorders (anti-phospholipid antibody syndrome, thrombophilia, systemic lupus erythematosus, Hashimoto’s thyroiditis, Graves’ disease, rheumatoid arthritis); or poor health habits. Demographic details and characteristics of women with NPs (n=69) and URPL (n=25) are shown in Table 1 . All samples were obtained under sterile conditions and divided into two parts: One part was immediately fixed in 4% paraformaldehyde for immunohistochemistry (IHC) studies, and the other part was immediately collected into ice-cold DMEM/F12 medium (Gibco, USA), transported to the laboratory within 30 min, and washed in 1× sterile phosphate buffered saline (PBS) for cell isolation.

HTR-8 cells, an immortalized first-trimester trophoblast cell line, were purchased from the Cell Bank of the Chinese Academy of Sciences, Shanghai, China, and cultured in DMEM/F12 medium (Gibco, USA) supplemented with 10% fetal bovine serum (FBS) (Gibco, USA).

Endometrium tissues were washed twice in 1× PBS and cut and digested in DMEM/F12 medium supplemented with 1.0 mg/ml type IV collagenase (Sigma-Aldrich, USA) for 30 min at 37°C with gentle agitation. The suspension was filtered through a 40-μm nylon mesh (Falcon, USA) and centrifuged at 300×g for 8 min. The collected cells were cultured in phenol red-free DMEM/F12 medium (Genom, China) containing 10% charcoal-stripped FBS (BioSun, China), 1% ITS (Oricellbio, China), and 500 ng/mL puromycin for 24 h in a 37°C humidified incubator containing 5% CO₂. Adherent cells (ESCs) were digested and resuspended in complete medium for other treatments.

Decidual tissues were trimmed into 1-mm³ segments and submerged in DMEM/F12 medium supplemented with 1.0 mg/ml type IV collagenase (Sigma-Aldrich, USA) and 150 U/ml DNase I (Sigma-Aldrich, USA). Digestion was performed for 30 min at 37°C with gentle shaking. After digestion, the cells were washed in sterile PBS and filtered through 100, 70, and 40μm sieves. The filtered suspension was centrifuged at 300×g for 8 min and the collected cells were resuspended in DMEM/F12 medium. The suspension was layered on a discontinuous Percoll density gradient (20%/40%/60%; GE Healthcare, USA) and centrifuged for 30 min at 800×g. DSCs were isolated from the 20%/40% Percoll interface, and immune cells were distributed at the 40%/60% Percoll interface. The cells were then washed in sterile PBS for twice. DSCs were cultured in DMEM/F12 medium (Genom, China) containing 10% FBS (Gibco, USA) for 24 h in a 37°C humidified incubator containing 5% CO₂. Immune cells were collected and resuspended in Roswell Park Memorial Institute (RPMI) 1640 medium supplemented with 10% FBS for further experiments.

dMφ were isolated from DICs using APC anti-human CD14 (BioLegend, USA) and Anti-APC MultiSort Kit (Miltenyi Biotec, Germany) according to the manufacturer’s protocol. Next, to separate CCR1^- and CCR1⁺ dMφ, the magnetic particles were first removed from dMφ by using the MultiSort Release Reagent (Miltenyi Biotec, Germany). Then dMφ were labeled with CCR1-fluorescein isothiocyanate (FITC) antibody (BioLegend, USA) followed by subsequent incubation with anti-FITC microbeads (Miltenyi Biotec, Germany). Then, magnetic separation was applied to positively select CCR1⁺ dMφ subset. The purity of CCR1⁺ dMφ was greater than 90%, as measured by FCM. The cells were cultured in DMEM-F12 medium supplemented with 10% FBS, 100 U/mL penicillin, and 100 mg/mL streptomycin.

PBMCs were isolated from peripheral blood samples of patients with NPs and URPL using Ficoll density gradient centrifugation (Solarbio, China) at 800×g for 20 min. CCR1⁺ peripheral monocytes (pMo) were isolated using the same method as that for CCR1⁺ dMφ. The cells were treated with 50 ng/ml recombinant human macrophage colony-stimulating factor (M-CSF) (MedChemExpress, China).

To induce in vitro decidualization, ESCs were plated on tissue culture plates, treated with complete medium containing 0.5 mM 8-bromoadenosine 3′,5′-cyclic monophosphate (8-Br-cAMP) (Sigma-Aldrich, USA) and 100 ng/ml MPA (Sigma-Aldrich, USA), and the medium was given fresh every other day and cells were treated for 4 days. The supernatant of cultured cells was collected and stored at −80°C for further analysis.

PBMCs were treated with 100 ng/ml rhCCL8 (R&D Systems, USA) with or without the administration of CCR1 antagonist (BX471) (MedChemExpress, China) at a concentration of 20 μM. A co-culture system of PBMCs and DSCs with or without the addition of 4 µg/mL CCL8 neutralizing antibody (R&D Systems, USA) was established via a 0.4-μm pore size Transwell co-culture system (Corning, USA). The neutralizing antibody was pretreated for 2 h. Briefly, PBMCs were placed in the lower chambers and DSCs were seeded in the upper chambers for 24 h before harvest.

A co-culture system of HTR-8 cells with CCR1^- dMφ or CCR1⁺ dMφ pretreated with control medium or recombinant human CCL8 (rhCCL8) (R&D Systems, USA) at concentrations of 10, 50, 100, and 200 ng/ml was established using a Transwell co-culture system (0.4-μm pore size, Corning, USA). Briefly, CCR1^- or CCR1⁺ dMφ were placed into the upper chambers and HTR-8 cells were seeded into the lower chambers. The ERK1/2 inhibitor PD98059 (MedChemExpress, China) was used at a concentration of 30 μM.

Cells were washed in PBS and incubated with fluorochrome-conjugated antibodies for 30 min at 4°C for cell surface staining. The following specific anti-human monoclonal antibodies were used: PerCP-Cy5.5-conjugated anti-CD45, FITC- or APC-conjugated anti-CD14, PE-Cy7- or FITC-conjugated anti-CCR1, AF700-conjugated anti-CD206, PE-conjugated anti-CD163, anti-CD80, and BV421-conjugated anti-CD86. Cells were fixed and permeabilized with BD Cytofix/Cytoperm™ Fixation/Permeabilization Kit (BD Biosciences, USA) according to the manufacturer’s protocol. The permeabilized cells were stained for intracellular cytokines as follows: APC-conjugated anti-IL-10, anti-TGF-β, and PE-conjugated anti-IL-8. All antibodies were purchased from BioLegend. FCM was performed using CytoFLEX (Beckman Coulter, USA), and the data were analyzed using FlowJo Version 6.1 software (TreeStar, USA).

The secretion of CCL8 in the cultured supernatant samples was determined using ELISA with the human CCL8 ELISA Kit (Abcam, UK) according to the manufacturer’s protocol. Absorbance was measured using a spectrophotometer (Biotek, Vermont, USA) at 450 nm.

Total RNA was extracted using TRIzol reagent (Invitrogen, USA) and reverse-transcribed into first-strand cDNA (TaKaRa Biotechnology, Japan) according to the manufacturer’s instructions. The synthesized cDNA was amplified using specific primers (Sagon, China) and SYBR Green (Yeasen, China) on the ABI PRISM 7900 Sequence Detection System (Applied Biosystems, USA). The reactions were run in duplicate using RNA samples. Fold change in the expression of each gene was calculated using the 2^−△△CT method, with actin as an internal control. According to the existing gene sequences in GenBank, primers were designed using computer assistance, and the primer sequences are shown in Table 2 .

IHC was performed on specimens of the human endometrium and decidual tissues from patients with NPs and URPL and villus tissues from patients with NPs. Paraffin-embedded sections were cut to a thickness of 3 µm. Tissue slides were incubated at 60°C for 2 h, and sections were deparaffinized and rehydrated in the order of using xylene and graded ethanol (100%, 95%, 85%, 75%, and 50%). They were soaked in 0.01 M citric acid (pH 6.0) for 20 min at 95°C for antigen retrieval. The slides were then incubated in 3% H₂O₂ for 10 min, blocked with 5% BSA for 20 min at room temperature, and incubated with rabbit anti-CCL8 antibody (Abcam, UK) at 4°C overnight in a humidity chamber. The samples were stained with an anti-rabbit IgG secondary antibody for 30 min and incubated in a DAB substrate solution until the desired staining intensity was reached at room temperature. The sections were then counterstained with hematoxylin, washed in running tap water for 30 min, dehydrated, and cleared through a graded series of ethanol (50%,75%, 85%, 95%, and 100%) and xylene. Finally, the sections were sealed with neutral resin and analyzed using an optical microscope (Olympus BX53, Japan).

Cell lysates were prepared in radioimmunoprecipitation assay lysis buffer (Beyotime, China) containing 1% proteinase inhibitor solution (Beyotime, China) and 1% phosphatase inhibitors (NCM Biotech, China). The protein yield was quantified using a bicinchoninic acid protein assay (Beyotime, China), and the protein was boiled with a 5× loading buffer (NCM Biotech, China). Total protein samples (20 μg) were separated using SDS–PAGE (Beyotime, China) and transferred to polyvinylidene difluoride membranes (Millipore, USA) for 1 h. Nonspecific binding sites were blocked by incubating the membranes with the QuickBlock™ Blocking Buffer (Beyotime, China) for WB for 10 min, followed by overnight incubation with primary antibodies against E-cadherin (1:5000; ProteinTech, USA), N-cadherin (1:2000; ProteinTech, USA), vimentin (1:1,000; PTM bio, China), ERK1/2 (1:1000, Proteintech, USA), p-ERK1/2 (1:1000, ProteinTech, USA), and tubulin (1:1000; Beyotime, China) diluted in blocking buffer (Beyotime, China) overnight at 4°C with gentle shaking. Primary antibodies were removed by washing the membranes three times in TBS-T, and the membranes were incubated for 1 h with secondary antibody (1:5000, arigo Biolaboratories Corp, China) at room temperature. After washing three times with TBS-T, immuno-positive bands on the blots were visualized using an enhanced chemiluminescence detection system (NCM Biotech, China).

Scratch wound healing assay was performed to assess cell motility. HTR-8 cells were seeded in six-well plates. CCR1^- and CCR1⁺ dMφ were pretreated with or without 100 ng/ml rhCCL8 for 24 h. When the HTR-8 cell density reached 80%, scratches were made with 1-ml pipette tips, and wounded monolayers were washed three times with PBS, followed by co-culture with pretreated CCR1^- and CCR1⁺ dMφ in a 0.4-μm pore size Transwell co-culture system with the addition of serum-free medium. Cells were incubated with 5% CO₂ at 37°C for 24 h, and the wound healing rates were determined and photographed. The images were analyzed and calculated using ImageJ software (NIH, USA).

An 8-µm pore size filter (Corning, USA) was used to perform Transwell migration. CCR1⁺ pMo and dMφ were seeded into the upper chamber without serum and complete medium with 10% serum with or without 100 ng/ml rhCCL8 was added to the lower chamber. Cells migrated to the lower chamber after 48 h. Cells were gently washed with PBS and fixed with 4% paraformaldehyde and then stained with crystal violet. Images were taken using a microscope and counted with Image J software.

A Transwell system with 8-μm chambers (Corning, USA) in a 24-well plate was precoated with Matrigel matrix (Corning, USA) to evaluate the invasive ability of HTR-8 cells. CCR1^- and CCR1⁺ dMφ were pretreated with or without 100 ng/ml rhCCL8. Then, HTR-8 cells resuspended in serum-free medium were placed into the upper chamber and the lower chamber was filled with complete medium, CCR1^- dMφ, or CCR1⁺ dMφ. The plates were incubated at 37°C for 48 h. The inserts were washed with PBS, and non-invading cells were removed from the upper chambers. The inserts were then fixed in 4% paraformaldehyde and stained with crystal violet. The invaded cells were imaged by microscopy and quantified by counting cells in five random fields using ImageJ software.

Single-cell sequencing data were downloaded from the Gene Expression Omnibus (GEO, https://www.ncbi.nlm.nih.gov/geo/), Genome Sequence Archive (GSA) database, and ArrayExpress from EMBL-EBI. The original source of single-cell raw data is CRA002181 (32), which comes from the decidua of 15 healthy controls and 9 patients with URPL in the GSA database. The expression profile of the blood and normal decidua of early pregnancy was derived from E-MTAB-6678 (33) in the ArrayExpress database, and endometrial samples were derived from GSE111976 (34) in the GEO database.

Fastq files of raw data were downloaded from the GSA database, and Cellranger (10X genomics) was used to process, align, and generate the feature-barcode unique molecular identifier matrices with default parameters. The gene expression matrix was analyzed using the R package Seurat (CreateSeuratObject) (35). To filter out low-quality cells, we first removed cells with detected gene numbers (<500 or >3000) and high mitochondrial content (≥10%). The expression matrix was then normalized (NormalizeData), and highly variable genes were identified by fitting the mean–variance relationship (FindVariableGenes). Next, we performed principal component analysis using highly variable genes and integrated samples using the R package Harmony (RunHarmony). The same principal components were used to embed cells in the K-nearest neighbor graph (FindNeighbors) and cluster them using the Louvain algorithm at a resolution of 0.5 (FindClusters). To label the cell clusters, we used a set of classic marker genes to annotate each cell type.

To investigate the function of macrophage subsets, we calculated the fold change of all detected genes (FindMarkers) and performed gene set enrichment analysis (GSEA) to determine the function of macrophage subsets in the decidua and peripheral blood using the R package fgsea.

To infer cell–cell communication between DIC and other cell types, we used CellChat to calculate and visualize cell–cell communication.

Statistical analysis was performed using GraphPad Prism Version 7 (GraphPad, USA). All data were first tested for normality before statistical analysis. For parametric data, Student’s t-test for two-group comparisons or one-way ANOVA for multiple group comparisons were used. For non-parametric data, Mann-Whitney test were used to compare the difference between two groups. Data are presented as mean ± SEM. The criterion for statistical significance was set at p< 0.05.

To characterize dMφ during early pregnancy, two published single-cell databases were integrated and analyzed (33, 34). The results showed that dMφ from the first trimester of pregnancy had a substantially higher expression of CCR1 than those in the proliferative and secretory endometrium ( Figure 1A ). Based on CCR1 expression, we divided macrophages into CCR1⁺ and CCR1⁻ subsets. A comparable percentage of CCR1⁺ Mφ was found in the proliferative and secretory endometrium; however, the infiltration of CCR1⁺ Mφ within the first trimester decidua considerably increased approximately 2–3 folds ( Figure 1B ). The decidua-specific expression of CCR1 in macrophages suggests the influence of a unique maternal–fetal microenvironment rather than reproductive hormones. We then conducted a gene ontology (GO) analysis to decipher the functional characteristics of CCR1⁺ dMφ. Compared with CCR1⁺ peripheral monocytes (pMo) ( Figure 1C ) and CCR1^- dMφ ( Figure 1D ), CCR1⁺ dMφ were strongly enriched in cell migration, tissue remodeling, anti-inflammatory responses, and vascularization functions. Thus, CCR1⁺ dMφ predominantly accumulate in the uterus during early pregnancy and are characterized by an immunosuppressive propensity.

The distinct CCR1⁺ dMφ population was clearly decreased in patients with URPL ( Figure 1E ). Moreover, in CCR1⁺ dMφ from patients with URPL, genes associated with anti-inflammatory responses (e.g., IL10, CD163, MRC1, TGFB1, and IL4) and extracellular matrix degradation (e.g., MMP9, MMP10, MMP14, MMP19, and ICAM1) were less enriched, whereas those responsible for activation and pro-inflammatory responses (e.g., CD86, IL1B, IL2, and TNF) were significantly enriched ( Figure 1F ). These sequencing results suggest a distinct CCR1⁺ dMφ population with a changed proportion and functional status in patients with URPL.

To validate the results from the single-cell profiling analysis, we performed FCM and found that dMφ had a significantly higher expression of CCR1 (59.98 ± 3.13) than did maternal pMo (28.04 ± 0.81) and nonpregnant endometrial Mφ (eMφ; 2.87 ± 0.79; Figure 1G ). Moreover, CCR1⁺ dMφ represented a more immunosuppressive phenotype with a higher expression of CD163, CD206, IL-10, IL-8, and TGF-β but lower expression of CD80 and CD86 than CCR1⁺ pMo ( Figure 1H ). In patients with URPL, remarkably decreased CCR1 expression was detected in both dMφ and pMo ( Figures 1I, J ). More importantly, the immunosuppressive phenotype of CCR1⁺ dMφ was remarkably attenuated in patients with URPL. In those with NP, CCR1⁺ dMφ exhibited a higher expression of CD163 and CD206 but lower expression of CD86 than CCR1^- dMφ. In patients with URPL, CCR1⁺ dMφ decreased the expression of CD163 and CD206 but increased the expression of CD80 and CD86, compared with CCR1⁺ dMφ from control donors with NPs ( Figure 1K ). The production of anti-inflammatory cytokines IL-10 and TGF-β and a proangiogenic factor, IL-8, was substantially upregulated in CCR1⁺ dMφ than in CCR1^- dMφ of normal controls but markedly reduced in CCR1⁺ dMφ of patients with URPL ( Figure 1K ). Collectively, these data demonstrate that CCR1⁺ dMφ exhibit an immunosuppressive and anti-inflammatory phenotype that is conducive to an immune-tolerant microenvironment, whereas decreased numbers of CCR1⁺ dMφ with dysfunction may be involved in the occurrence of URPL.

Besides the enhanced activated and inflammatory phenotype of CCR1⁺ dMφ in patients with URPL, CCR1⁺ dMφ in these patients have a decreased function in tissue remodeling and negative regulation on trophoblast cell migration, as revealed by single-cell sequencing analysis ( Figure 2A ). Next, we probed the effects of CCR1⁺ dMφ on the biological behavior of trophoblasts. CCR1^- dMφ and CCR1⁺ dMφ were isolated and co-cultured with HTR-8 cells (a human extravillous trophoblast cell line) respectively in a non-contact Transwell system. Compared to co-culture with CCR1^- dMφ, the migration and invasion of HTR-8 cells significantly increased when co-cultured with CCR1⁺ dMφ ( Figures 2B, C ). A recent study demonstrated that EMT plays an important role in the regulation of trophoblast migration and invasion, and this process is promoted by M2 macrophages but inhibited by M1 macrophages (7). WB was performed to analyze EMT markers and signaling pathways involved in HTR-8 after co-culture with CCR1⁺ dMφ or CCR1^- dMφ. In the CCR1⁺ dMφ co-culture group, the expression of the epithelial marker E-cadherin was markedly decreased, whereas that of the mesenchymal markers N-cadherin and vimentin was dramatically increased in HTR-8 cells ( Figure 2D ). The expression levels of E-cadherin, N-cadherin and vimentin were comparable between the HTR-8 cells cultured alone and those co-cultured with CCR1^- dMφ ( Figure 2D ). Moreover, CCR1⁺ dMφ co-culture increased the phosphorylation of ERK1/2 in HTR-8 cells in a time-dependent manner ( Figure 2E ). A specific EKR1/2 inhibitor, PD98059, was applied to the CCR1⁺ dMφ-HTR-8 co-culture system and effectively abrogated the CCR1⁺ dMφ-induced EMT of trophoblasts by reducing the expression of N-cadherin and vimentin but increasing the expression of E-cadherin ( Figure 2F ). Furthermore, the migratory and invasive activities of HTR-8 cells were determined following treatment with the EKR1/2 inhibitor. As shown in Figures 2G, H , ERK1/2 inhibition attenuated the migration and invasion of CCR1⁺ dMφ cells. These data suggest that CCR1⁺ dMφ induce EMT to promote trophoblast migration and invasion by activating the ERK1/2 pathway. Consistent with the decreased and dysfunctional CCR1⁺ dMφ in patients with URPL, deficient EMT of trophoblasts was observed in patients with URPL, which was manifested as upregulated expression of E-cadherin and downregulated expression of N-cadherin in placental villous tissues, mainly in cytotrophoblasts ( Figure 2I ). Together, our findings suggest that dysregulated CCR1⁺ dMφ affect EMT, leading to inadequate trophoblast migration and invasion, which is involved in the occurrence of URPL.

Next, we examined the possible regulator(s) of CCR1⁺ dMφ and underlying mechanisms in early pregnancy. We investigated the interaction between dMφ and other decidual cells using single-cell sequencing analysis (33, 34). CellChat analysis of single-cell sequencing in the endometrium and decidua revealed a potentially increased interaction between the stromal cells and macrophages in early pregnancy compared with that in the endometrium ( Figures 3A, B ). We also compared the overall communication probabilities of the two databases (33, 34). Intriguingly, 43 out of 66 pathways were highly active, albeit at different levels, in the decidual tissues ( Figure 3C ). The CCL signaling pathway was activated in the decidua ( Figure 3C ). As depicted in Figure 3D , the CCL pathway exhibited abundant signaling interactions between DSCs and dMφ. We then analyzed changes in the expression of CCL molecules and found that the expression levels of CCL13, CCL23, CCL28, CCL3L3, CCL4, CCL5, and CCL8 were markedly higher in the decidua than in the proliferative and secretory endometrium ( Figure 3E ). Specifically, only two CCL pathways, CCL8-CCR1 and CCL8-CCR2, were significantly enriched in DSCs and dMφ ( Figure 3F ). In addition, single-cell sequencing suggested that CCL8, a ligand of CCR1, was highly expressed in DSCs compared with that in nonpregnant ESCs ( Figures 3E, G ). CCL8 can bind to several receptors, including CCR1, CCR2, CCR3, CCR5, and CCR8 (24). However, except for CCR1, other receptors were scarcely expressed on macrophages from either the nonpregnant endometrium or pregnant decidua ( Figure 3H ). Thus, decidua-derived CCL8 is a possible ligand that interacts with CCR1⁺ dMφ.

We then measured CCL8 expression at the maternal–fetal interface. As shown in Figure 3I , compared with the endometrium and villous, higher CCL8 expression was detected in the decidua from normal early pregnancy. Furthermore, both mRNA and protein expression levels of CCL8 were dramatically increased during in vitro decidualization ( Figures 3J, K ). However, immunochemistry staining showed that the decidua from patients with URPL exhibited markedly downregulated CCL8 expression compared with that from NPs ( Figure 3L ). We also identified significantly lower levels of CCL8 in DSCs from patients with URPL ( Figures 3M, N ). These findings indicate that elevated CCL8 during stromal cell decidualization is a possible regulator of CCR1⁺ dMφ, and abnormal CCL8-CCR1⁺ dMφ communication may be implicated in URPL.

Next, we investigated the regulatory effects of DSC-derived CCL8 on CCR1⁺ dMφ. We first performed a chemotaxis assay and found that CCL8, as a chemokine, potently promoted the migration of both CCR1⁺ pMo and CCR1⁺ dMφ ( Figures 4A, B ), suggesting that CCL8 contributes to the recruitment and residence of CCR1⁺ dMφ in early pregnancy.

CCR1⁺ dMφ was characterized by an anti-inflammatory phenotype in women with NP, which changed to an activated and less immunosuppressive status in patients with URPL. Whether CCL8 influences the unique phenotype of CCR1⁺ dMφ during early pregnancy remains unknown. Accordingly, pMφ were treated with recombinant human CCL8 (rhCCL8), and phenotypic changes were analyzed using FCM. Higher expression of CD163 and CD206 and lower expression of CD86 were detected in CCR1⁺ pMφ treated with rhCCL8, whereas the expression of CD80 remained unchanged. In addition, CCL8-treated CCR1⁺ pMφ produced more anti-inflammatory cytokines, including IL-10 and TGF-β, and the proangiogenic factor IL-8 ( Figure 4C ). Significantly, a specific CCR1 inhibitor, BX471, was applied during the in vitro stimulation of pMφ with CCL8, and all CCL8-mediated effects on pMφ, including increased expression of immunosuppressive markers and anti-inflammatory cytokines, could be abrogated by BX471. These results suggest that the CCL8–CCR1 interaction is conducive to the inactive and anti-inflammatory phenotype of macrophages at the maternal–fetal interface.

Similar results were obtained in the indirect contact co-culture system of pMφ and DSCs using Transwell chambers. DSCs effectively shifted pMφ into an anti-inflammatory phenotype with a higher expression of CD163, CD206, IL-10, TGF-β, and IL-8 but lower expression of CD80 and CD86 ( Figure 4D ). This effect was partially blocked by a neutralizing antibody against CCL8 ( Figure 4D ). Collectively, DSC-derived CCL8 can recruit CCR1⁺ pMo from the maternal periphery and induce an immunosuppressive phenotype, contributing to maternal–fetal immune tolerance.

We have demonstrated that CCR1⁺ dMφ could promote the migration, invasion, and EMT of trophoblast cells. Whether CCL8 has an impact on this function of CCR1⁺ dMφ is unknown. Therefore, HTR-8 cells were co-cultured with rhCCL8 preteated-CCR1⁺ dMφ or CCR1^- dMφ, followed by migration and Transwell assays. Consistent with the above results, CCR1⁺ dMφ significantly facilitated the functions of HTR-8 cells, while CCR1^- dMφ, whether treated with rhCCL8 or not, had less effects on migration, invasion or EMT process of HTR-8 cells ( Figures 5A–C ). Compared with CCR1⁺ dMφ, CCL8-pretreated CCR1⁺ dMφ showed a significantly enhanced capacity to promote trophoblast migration and invasion ( Figures 5A, B ). Moreover, CCL8-pretreated CCR1⁺ dMφ co-culture further increased the expression of N-cadherin and vimentin and decreased the expression of E-cadherin in trophoblast cells ( Figure 5C ), indicating a more activated EMT process than that induced by CCR1⁺ dMφ. In addition, CCR1⁺ dMφ pretreated with different concentrations of CCL8 promoted the activation of the ERK1/2 pathway in trophoblast cells in a concentration-dependent manner ( Figure 5D ). These data suggest that CCL8 has a positive effect on the function of CCR1⁺ dMφ in promoting EMT-mediated trophoblast migration and invasion.