MNSFβ Regulates TNFα Production by Interacting with RC3H1 in Human Macrophages, and Dysfunction of MNSFβ in Decidual Macrophages Is Associated With Recurrent Pregnancy Loss

Decidual macrophages (dMϕ) are the second largest population of leukocytes at the maternal–fetal interface and play critical roles in maintaining pregnancy. Our previous studies demonstrated the active involvement of monoclonal nonspecific suppressor factor-β (MNSFβ) in embryonic implantation and pregnancy success. MNSFβ is a ubiquitously expressed ubiquitin-like protein that also exhibits immune regulatory potential, but its function in human dMϕ remains unknown. Here, we observed that the proportion of CD11chigh (CD11cHI) dMϕ was significantly increased in dMϕ derived from patients with recurrent pregnancy loss (RPL dMϕ) compared to those derived from normal pregnant women (Control dMϕ). The production of MNSFβ and TNFα by RPL dMϕ was also significantly increased compared to that by Control dMϕ. Conditioned medium from RPL dMϕ exerted an inhibitory effect on the invasiveness of human trophoblastic HTR8/SVneo cells, and this effect could be partially reversed by a neutralizing antibody against TNFα. Bioinformatics analysis indicated a potential interaction between MNSFβ and RC3H1, a suppressor of TNFα transcription. Immunoprecipitation experiments with human Mϕ differentiated from the human monocyte cell line Thp1 (Thp1-derived Mϕ) proved the binding of MNSFβ to RC3H1. Specific knockdown of MNSFβ in Thp1-derived Mϕ led to a marked decrease in TNFα production, which could be reversed by inhibiting RC3H1 expression. Interestingly, a significant decrease in the protein level of RC3H1 was observed in RPL dMϕ. Together, our findings indicate that aberrantly increased MNSFβ expression in dMϕ may promote TNFα production via its interaction with RC3H1, and these phenomena could result in the disruption of the immune balance at the maternal–fetal interface and thus pregnancy loss.

Decidual macrophages (dMf) are the second largest population of leukocytes at the maternal-fetal interface and play critical roles in maintaining pregnancy. Our previous studies demonstrated the active involvement of monoclonal nonspecific suppressor factor-b (MNSFb) in embryonic implantation and pregnancy success. MNSFb is a ubiquitously expressed ubiquitin-like protein that also exhibits immune regulatory potential, but its function in human dMf remains unknown. Here, we observed that the proportion of CD11c high (CD11cHI) dMf was significantly increased in dMf derived from patients with recurrent pregnancy loss (RPL dMf) compared to those derived from normal pregnant women (Control dMf). The production of MNSFb and TNFa by RPL dMf was also significantly increased compared to that by Control dMf. Conditioned medium from RPL dMf exerted an inhibitory effect on the invasiveness of human trophoblastic HTR8/ SVneo cells, and this effect could be partially reversed by a neutralizing antibody against TNFa. Bioinformatics analysis indicated a potential interaction between MNSFb and RC3H1, a suppressor of TNFa transcription. Immunoprecipitation experiments with human Mf differentiated from the human monocyte cell line Thp1 (Thp1-derived Mf) proved the binding of MNSFb to RC3H1. Specific knockdown of MNSFb in Thp1-derived Mf led to a marked decrease in TNFa production, which could be reversed by inhibiting RC3H1 expression. Interestingly, a significant decrease in the protein level of RC3H1 was observed in RPL dMf. Together, our findings indicate that aberrantly increased MNSFb expression in dMf may promote TNFa production via its interaction with RC3H1, and
MNSFb (monoclonal nonspecific suppressor factor-b), also known as Fau (Finkel-Biskis-Reilly murine sarcoma virusassociated ubiquitously expressed gene), is a 133-aa protein containing a ubiquitin-like (Ubi-L/FUBI) domain and a ribosomal protein S30 domain, and the homology between the murine and human MNSFb proteins is greater than 97.8% (6). MNSFb was originally identified as an inhibitor of the T-cellmediated immune response (7) because it inhibits not only the proliferation of T and B cells but also the secretion of interleukin-4 (IL-4) from type 2 helper T cells and bone marrow-derived mast cells (8,9). Furthermore, it has been reported that MNSFb promotes the apoptosis (10) but inhibits the phagocytosis (11) and TNFa production (12) of murine macrophages.
MNSFb was first shown to be involved in embryo implantation due to its differential expression between the implantation sites and interimplantation sites of endometrial tissues in mice (13). Subsequently, it was revealed by our previous studies that deficiency in MNSFb could lead to embryo implantation failure in mice (14,15), and the MNSFb expression levels in both the decidual and villus tissues of RPL patients were significantly decreased (16,17). In particular, MNSFb exerted stimulatory effects on the proliferation and migration of human EVTs, suggesting that insufficient MNSFb expression at the maternal-fetal interface might lead to early pregnancy loss by interfering with the invasion of EVTs (17).
Given that MNSFb is widely expressed in various tissues and cells (6), and it is involved in regulation of embryo implantation, as well as the immune response of macrophages, we supposed that MNSFb might contribute to the immune balance at the maternal-fetal interface by regulating activities of dMf, and malfunction of MNSFb in dMf might be associated with the early pregnancy failure. Thus, this study was carried out to investigate the alteration of MNSFb expression in dMf of RPL patients, and the effect and its underlying molecular pathway of abnormal MNSFb expression on activities of dMf by using bioinformatic analysis and a human macrophage model differentiated from immortalized human monocyte cells.

Human Decidual Tissue Collection
Human decidual tissues from 24 RPL patients (RPL, 6-10 weeks of gestation) and 25 normal pregnant women in the first trimester (Control, 6-9 weeks of gestation) were collected at the Department of Gynecology and Obstetrics, the Second Hospital of Tianjin Medical University (Tianjin, China). These collected decidual tissues were immediately washed several times with sterile, glucose-free PBS solution until there were no obvious blood clots in the decidual tissues. Then, the tissues were immersed in ice-cold RPMI-160 medium. Cells were harvested from the tissues or the tissues were fixed within 3 h. Current pregnancy losses of the RPL patients were objectively confirmed by transvaginal ultrasound examination. Patients with classical risk factors, including abnormal parental karyotypes, uterine anatomical abnormalities, infectious diseases, luteal phase defects, diabetes mellitus, thyroid dysfunction, and hyperprolactinemia, were excluded from this study. In parallel, Control women who had no history of miscarriage and were undergoing legal, voluntary terminations of early pregnancy were enrolled and evaluated for classical risk factors for early pregnancy loss. The sample collection for this study was approved by the Medical Ethics Committees of The Second Hospital of Tianjin Medical University (KY2017K002) and Shanghai Institute for Biomedical and Pharmaceutical Technologies (former Shanghai Institute of Planned Parenthood Research) (Ref # 2018-05). All the samples were collected after informed consent was obtained. No significant differences in the average age or gestational week at sampling were observed between the RPL patients and Control women (Supplementary Table S1).

Isolation of Human Decidual Macrophages (dMf)
DMf were isolated as previously described (18,19) with modifications. Briefly, tissues were washed and crushed into small pieces by a tissue crusher (Gentle MACS Dissociator, Miltenyi Biotec, Germany). Pieces of decidual tissues were digested in 3 mg/ml collagenase Type IV (Gibco, USA), 100 IU/ml DNase I (Sigma-Aldrich, USA), 100 IU/ml penicillin, and 100 IU/ml streptomycin (Gibco) at 37°C for 30 min. Subsequently, the decidual cells that were released were filtered through 100, 200, and 400 mesh sieves (Corning, USA). To exclude any remaining red blood cells, the filtered cells were incubated with red blood cell lysis buffer (BD Biosciences, USA). DMf were obtained with CD14+ antibodies conjugated to magnetic beads (Miltenyi Biotec). The purity of the dMf (CD45 + CD14 + ), which was detected by flow cytometry, was approximately 90% (Supplementary Figure S1B).

Cell Culture
Primary dMf was cultured at a final concentration of 1 × 10 6 cells/ml in RPMI 1640 medium supplemented with 10% fetal bovine serum (FBS), 100 U/ml penicillin, and 100 µg/ml streptomycin (Gibco) under standard culture conditions (37°C in a 5% humidified CO 2 incubator). The immortalized human first-trimester extravillous trophoblast cell line, HTR8/SVneo, characterized by the abilities of growth and invasion (20), were kindly provided by Prof. Hongmei Wang, Institute of Zoology, Chinese Academy of Sciences, China, and cultured in RPMI 1640 medium, supplemented with 10% FBS, 100 U/ml penicillin, and 100 µg/ml streptomycin (Gibco) under standard culture conditions (37°C in a 5% humidified CO 2 incubator). The immortalized human monocyte cell line, Thp1, characterized by the absence of immunoglobulins, and the ability to restore Tlymphocyte response to ConA (21), was purchased from the American Type Culture Collection (ATCC, Manassas, USA). Thp1 cells were cultured routinely in RPMI 1640 medium plus 10% FBS and 0.1% b-mercaptoethanol with antibiotics (Gibco) under standard culture conditions (37°C in a 5% humidified CO 2 incubator) and treated with 200 ng/ml phorbol myristate acetate (PMA) for 24 h to induce the differentiation of Thp1 cells into macrophages. Thp1-derived Mf cells differentiate normally and expressed high level of the macrophage markers CD11b and SPI1 (Supplementary Figure S2).

Immunofluorescence Staining and Confocal Microscopy
Fresh decidual tissues collected from physically normal pregnant women in the early stage of pregnancy (8W) were embedded in Tissue-Tek O.C.T. compound (Sakura Finetek, USA), and frozen sections were generated with a thickness of 8 µm. The frozen sections were fixed in 4% paraformaldehyde (Sigma-Aldrich) for 24 h at 4°C and treated with 0.1% Triton. Then, the frozen sections were incubated with antibodies against human MNSFb (0.5 mg/ml, prepared by our lab), CK7 (0.5 mg/ml, ZSGB-BIO, China), CD31 (1 mg/ml, Abcam, USA), or CD14 (1 mg/ml, Abcam). Then, the sections were incubated with FITCconjugated or TRITC-conjugated secondary antibodies (ZSGB-BIO, China), and the cell nuclei of the sections were stained with 4,6-diamidino-2-phenylindole (DAPI; Sigma Aldrich). The results were recorded using a laser confocal microscope (Leica, Germany) and processed with ZEN 2012 software (Zeiss).

Real-Time Reverse Transcription Polymerase Chain Reaction Analysis
Total cellular RNA was extracted using TRIzol (Invitrogen) according to the manufacturer's instructions. Reverse transcription was performed with 0.2 mg of total RNA using SuperScript II Reverse Transcriptase (Invitrogen) and an oligo-dT primer (Invitrogen). Detailed information about the sequences of the primers is listed in Supplementary Table S2. Real-time PCR was performed by using the SYBR II kit (TaKaRa, China) according to the manufacturer's instructions on a Light Cycler 480 real-time PCR System (Roche, USA). The relative mRNA expression levels were determined by the 2 −DDCt method and normalized to the expression levels of Gapdh.

Western Blotting Analysis
Total cellular proteins were extracted from dMf or decidual cells using TRIzol (Invitrogen, USA) or SDS lysis buffer (2% SDS, 50 mM Tris-HCI, pH 7.6, 2 mM EDTA, and 10% glycerol). The protein concentrations were quantified by using the BCA Protein Assay Kit (Thermo Pierce, USA). Then, 20 mg of protein was subjected to 10% SDS-PAGE and electrophoretically transferred to a PVDF membrane (Thermo Pierce). After blocking with 5% BSA, the membrane was incubated with different primary antibodies (Supplementary Table S3) overnight at 4°C. Then, the membrane was washed and incubated with the corresponding horseradish peroxidase-conjugated secondary antibodies for 1 h at room temperature. The results were captured with a Gene Gnome Imaging System (Syngene, UK). The relative densities of the target proteins were determined by normalization to the density of beta-actin (b-actin) in the same blot, and the results were analyzed with ImageJ software.

Flow Cytometry
Cells were incubated with Human TruStainFcX (BioLegend, USA) for 30 min to block the Fc receptors and subsequently stained with CD45-Pacific Blue, CD14-APC, and CD11c-FITC (BioLegend) for 30 min at 4°C. After washing with PBS three times, the cells were incubated with Cytofix/Cytoperm (BD Biosciences) for 20 min at 4°C. After washing with wash buffer (BD Biosciences) three times, the cells were incubated with specific antibodies against MNSFb for 30 min at 4°C. Then, the cells were incubated with Cyanine Cy ™ 5-conjugated secondary antibodies (Jackson, USA) for 30 min at 4°C. Approximately 100,000 cells were detected using FACS (BD, USA), and the data were analyzed with FlowJo V10.2.

Collection of Conditioned Media
Primary PRL dMf and Control dMf were cultured at a final concentration of 1 × 10 6 cells/ml in RPMI 1640 medium supplemented with 10% FBS, 100 U/ml penicillin, and 100 µg/ml streptomycin (Gibco) for 48 h. The supernatants of the cultured PRL dMf were collected and used as conditioned media from PRL dMf (PRL dMf CM), while those of cultured Control dMf were collected and used as the Control dMf CM. The collected CM samples were centrifuged at 12,000 g for 5 min at 4°C and subsequently stored at −80°C. Thp1-derived Mf transfected with NC or siMNSFb were cultured in RPMI 1640 medium with 10% FBS and 0.1% b-mercaptoethanol with antibiotics (Gibco) for 24 h. The conditioned media from siMNSFb-transfected (siMNSFb) or NC-transfected (NC) Thp1-derived Mf were collected. The samples were centrifuged at 12,000 g for 5 min at 4°C and stored at −80°C for subsequent experiments.

Transwell Assay
The invasive potential of HTR8/SVneo cells was assessed in vitro using a BD BioCoat Matrigel Invasion Chamber (BD Biosciences, Bedford, USA) as previously described (22)

Immunoprecipitation (IP)
The in vivo binding of MNSFb to RC3H1 was assayed by IP. Thp1-derived Mf were harvested in cold phosphate-buffered saline (PBS) and washed with PBS. Then, the cell pellets were suspended in lysis buffer (50 mM Tris-HCl, 120 mM NaCl, 1% Nonidet P-40, and protease inhibitor) and incubated on ice for 30 min. After centrifugation of the cell lysates (15 min, 12,000 g, 4°C), supernatant samples were used for IP. Samples containing total protein extract (1 mg of protein), 4 µg of anti-MNSFb antibody, anti-RC3H1 antibody or normal rabbit IgG, and 40 µl of protein A/G beads (50% slurry, Santa Cruz, SC-2003) were incubated at 4°C overnight with agitation. The beads were washed six times with wash buffer A (20 mM Tris-HCl, 1 mM EDTA, 900 mM NaCl, and 1% Nonidet P-40), and then washed once with wash buffer B with 100 mM NaCl before the elution (95°C, 5 min) of the bound proteins with gel-loading buffer.

Enzyme-Linked Immunosorbent Assays (ELISAs)
The conditioned media (CM) from primary dMf isolated from PRL patients (PRL dMf) or Control women (Control dMf) were collected, and the TNFa levels were measured by ELISA using a commercial sandwich ELISA kit according to the manufacturer's protocol (Quanticyto, China). The absorbance was read using an Infinite 200 Pro M Plex (TECAN). The absorbance readings were taken at 450 nm. A standard curve of TNFa was simultaneously analyzed in every plate using the dilution buffer provided by the manufacturer, and the TNFa concentrations in the samples were calculated based on the standard curve and dilution factor.

Statistical analysis
We established biological replicates during the processing of decidual tissues from each PRL patient or normal pregnant woman. All the experiments were repeated at least three independent times, and all the values are presented as the mean ± SD. All the statistical analyses were conducted with GraphPad Prism Version 6.0. Statistical analysis was carried out by two-sided Student's t-test, and differences were considered significant at p < 0.05.

Distribution of MNSFb in Decidual Macrophages at the Human Maternal-Fetal Interface
As the expression of MNSFb in human dMf has not yet been reported, immunofluorescent staining analysis was carried out to determine the localization of the MNSFb protein in human decidual tissues during early pregnancy (8W). In addition to decidual stromal cells (DSCs), CK7, CD31, and CD14 protein signals were used as markers of trophoblast cells, endothelial cells, and macrophages, respectively. CK7-positive staining indicated the implantation site (IS), whereas CK7-negative staining indicated the nonimplantation site (nIS). The results showed that the MNSFb protein signals were widely distributed in human decidual tissues, including decidual macrophages, during the first trimester ( Figure 1A). In addition, the protein expression of MNSFb in human DSCs and dMf during early pregnancy was also detected by Western blotting analysis ( Figure 1B).

Increased MNSFb Expression in Decidual Macrophages Isolated From RPL Patients
In our previous studies, the MNSFb expression levels in both the decidual and villus tissues from RPL patients were observed to be significantly decreased compared to those in the tissues from normal pregnant women (16,17); thus, we isolated dMf from RPL patients (RPL dMf) and normal pregnant women (Control dMf) with~90% purity (Supplementary Figure S1A) by using CD14 and CD45 as biomarkers (19), and we hypothesized that the MNSFb expression level in RPL dMf would also be reduced. Unexpectedly, the RT-PCR (Figure 2A) and Western blotting ( Figure 2B) results showed that the MNSFb expression level in RPL dMf was significantly increased compared to that in Control dMf. The upregulated MNSFb expression in RPL dMf was further confirmed by flow cytometry ( Figure 2C). However, the MNSFb expression level in the total cells isolated from the decidual tissues from RPL patients was obviously reduced compared to that in the total cells isolated from the decidual tissues from Control women (Supplementary Figure S1B).

Proportion of CD11c high dMj Was Increased While That of CD11c low dMj Was Decreased in RPL Patients
Macrophages are usually classified into M1 and M2 subtypes; however, human dMf could not be clustered by M1 and M2 markers (18). It has been reported that macrophages in normal first-trimester decidual tissues could be categorized into CD11c high (~20%) and CD11c low (~80%) subsets (18,19). In the present study, we separated CD11c high dMf (CD11c hi) and CD11c low dMf (CD11c low) by flow cytometry and found that the proportions of the CD11c hi and CD11c low subsets in Control women were 27.0% ± 6.4% and 73.0% ± 6.4%, respectively, whereas those of the CD11c hi and CD11c low subsets in RPL patients were significantly increased to 58.4 ± 16.2% and dramatically decreased to 41.6 ± 16.2%, respectively ( Figure 3A). No significant difference in the MNSFb expression level was observed between the CD11c hi subset and CD11c low subset; however, the MNSFb expression levels in the CD11c hi and CD11c low subsets from RPL patients were significantly increased compared to that of Control women ( Figure 3B).

Expression and Secretion of TNFa Were Increased in dMf From RPL Patients
As it has been reported that MNSFb inhibits TNFa expression in murine macrophages (12), we reasonably thought that the TNFa expression level might be decreased in RPL dMf because MNSFb expression was upregulated. However, the RT-PCR results showed that the TNFa expression level in cultured primary dMf from RPL patients (RPL dMf) was significantly enhanced compared to that in cultured primary dMf from normal pregnant women (Control dMf) ( Figure 4A). The TNFa concentration in the conditioned media of RPL dMf (RPL dMf CM) was also higher than that in the conditioned media of Control dMf (Control dMf CM) ( Figure 4B), indicating that TNFa production and secretion levels were increased in the dMf from RPL patients.

Decidual Macrophages of RPL Patients Inhibited HTR8/SVneo Cell Invasion Partially Through TNFa
TNFa could inhibit the invasion of HTR8/SVneo cells (22), and we found that the level of TNFa secretion by RPL dMf was increased; thus, we also examined the effect of the conditioned media of dMf on the invasion of HTR8/SVneo cells. The Transwell assay results showed that the invasion of HTR8/ SVneo cells treated with RPL dMf CM was significantly reduced compared with that of cells treated with Control dMf CM ( Figure 4C), indicating that RPL dMf might inhibit the invasion of EVTs in a paracrine manner. More interestingly, the invasion of HTR8/SVneo cells was inhibited by Control dMf CM plus TNFa (TNFa+Control dMf CM), whereas the inhibitory effect of RPL dMf CM on the invasion of HTR8/ SVneo cells could be restored by the addition of an anti-TNFa antibody (anti-TNFa+ RPL dMf CM) ( Figure 4C); these results suggested that dMf might regulate the invasion of EVTs by secreting TNFa.

Knockdown of MNSFb in Mf Resulted in Reduced TNFa Production and Enhanced HTR8/SVneo Cell Invasion
The results of the experiments mentioned above suggested that MNSFb expression in RPL dMf was increased, and RPL dMf might inhibit the invasion of EVTs via increased TNFa secretion, indicating a potential positive correlation between MNSFb expression levels and TNFa production levels in human dMf. Thus, we observed the effect of downregulated MNSFb expression on the TNFa production induced by LPS in human Thp1-derived Mf; a suitable in vitro model to investigate the Mf functions (23). The results showed that in Thp1-derived Mf, the MNSFb expression level could be significantly knocked down by transfection with a specific siRNA, and the TNFa expression level was also decreased in MNSFb-knockdown Mf ( Figures 5A, B). Furthermore, the invasion of HTR8/SVneo cells was enhanced by treatment with the CM of MNSFb-knockdown Mf (siMNSFb), and this stimulatory effect could be eliminated by the addition of TNFa ( Figure 5C); these results further indicated that dMf could inhibit the invasion of EVTs by secreting TNFa.

MNSFb Interacted With RC3H1 to Regulate TNFa Expression in Human Macrophages
To explore the molecular mechanism underlying the positive correlation between MNSFb expression and TNFa expression in human macrophages, we searched for and predict proteins that could potentially interact with MNSFb through BioGRID (https:// thebiogrid.org/), and RC3H1 was identified as a candidate (Supplementary Figure S3). Given that RC3H1 could inhibit TNFa expression (24), we hypothesized that MNSFb might promote TNFa production by weakening the inhibitory effect of RC3H1 on TNFa expression. Thus, we investigated the interaction between MNSFb and RC3H1 in human macrophages by co-IP assay. The results showed direct binding between the MNSFb and  RC3H1 proteins in Thp1-derived Mf ( Figure 5D). Then, we knocked down the expression of MNSFb or RC3H1 in Thp1derived Mf and found that the TNFa expression level was increased in RC3H1-knockdown (siRC3H1) cells but decreased in MNSFbknockdown (siMNSF) cells ( Figure 5E). These results suggested that MNSFb might promote the expression of TNFa by binding to RC3H1.

Protein Level of RC3H1 Was Decreased in dMf From RPL Patients
We also detected the RC3H1 expression level in primary human decidual macrophages. As we hypothesized, the protein expression of RC3H1 was decreased in RPL dMf compared with Control dMf ( Figure 6B). However, the RC3H1 mRNA expression level was not significantly different ( Figure 6A), indicating that MNSFb might only affect the protein level of RC3H1.

DISCUSSION
It was found in the present study that, at the maternal-fetal interface of RPL patients, dMf showed an inclination to CD11c high subtype, instead of CD11c low subtype, accompanied with the significantly increased productions of MNSFb and TNFa, and the reduced production of RC3H1. MNSFb might promote the secretion of TNFa by binding to RC3H1. Compared to dMf of Control women at early pregnancy, dMf of RPL patients could strongly inhibit the invasive activity of extravillous trophoblasts (EVTs) in a paracrine manner at least partially mediated by TNFa (Figure 7). Successful establishment of mammals' pregnancy depends on the generation of maternal-fetal immune tolerance and remodeling of the spiral arteries (25). Dysfunction in these processes has been correlated with adverse pregnancy outcomes including RPL (26). Although it has been well recognized that dMf participate in the immune modulation and spiral artery remodeling at the maternal-fetal interface (25,27), the exact roles of dMf in these events are still not well understood.
We previously found that, MNSFb expression was significantly decreased in both decidual and villus tissues from RPL patients (16,17). The knockdown of MNSFb expression could inhibit proliferation and migration of human EVTs (17), as well as proliferation (Supplementary Table S4) and decidualization (Supplementary Figure S4 whether MNSFb participates in the regulation of decidual immune cells, especially dMf, has been unknown. Given decidual MNSFb expression was reduced in RPL patients, and the secreted form of MNSFb was identified as an immune suppressor, it was reasonable to hypothesize that MNSFb expression in dMf of RPL patients might well be decreased to destroy the immunotolerance at the maternalfetal interface. However, unexpectedly, in this study, it was observed that, MNSFb expression in RPL dMf was obviously increased (Figure 2), whereas its expression in total decidual cells from RPL patients was significantly decreased as expected (Supplementary Figure S1C). These data suggested that MNSFb might play different roles in various cells at the maternal-fetal interface.
Thus, we hypothesized that MNSFb may be involved in the regulation of cell proliferation and invasion of EVTs, cell proliferation and decidualization of ESCs, and cell differentiation and secretion of DICs. Therefore, the decreased MNSFb expression in EVTs and DSCs might lead to pregnancy loss by interfering with the invasion of EVTs and generation of decidua; however, in dMf, the abnormally increased MNSFb expression might disrupt the immune homeostasis of the local microenvironment and result in pregnancy failure.
Macrophages are usually classified into the M1 subtype, a proinflammatory phenotype, or the M2 subtype, an antiinflammatory phenotype (28). Enhanced M1 polarization, characterized by increased TNFa expression, is associated with RPL or preeclampisa (29,30). However, due to the lack of appropriate biomarkers, we were unable to separate M1 and M2 dMf from the decidual tissues at a reasonably high purity. Fortunately, it was reported that two distinct subtypes of dMf, namely, CD11c hi and CD11c low, could be specifically separated by the cell surface marker CD11c (18,19). Although they secrete both pro-inflammatory and anti-inflammatory cytokines, CD11c hi dMf are thought to be involved in lipid metabolism and inflammation, whereas CD11c low dMf are related to the extracellular matrix formation and tissue growth (18). However, whether the alteration in CD11c dMf polarization associates with RPL is unclear. It was reported that, CD11c hi dMf are localized close to EVTs (19), and TNFa is highly expressed in CD11c hi Mf (31). TNFa could inhibit the invasiveness of EVTs (22), and increased decidual and peripheral TNFa levels were observed in RPL patients (32). Interestingly, we found that the proportion of CD11c hi dMf was significantly raised in RPL patients (Figure 3), and the production and secretion level of TNFa was obviously increased in dMf from RPL patients (Figure 4). In addition, compared to dMf from normal women early in pregnancy, dMf from RPL patients exerted a stronger inhibitory effect on the invasion of EVTs in a paracrine manner, and this effect could be effectively reversed by anti-TNFa antibody (Figure 4). These results indicated that the appropriate CD11c dMf polarization was important for the microhomeostasis at the maternal-fetal interface.
In this study, we found that knockdown of MNSFb led to the decreased TNFa expression in human Mf. This phenomenon conflicted with previous reports that MNSFb inhibits TNFa production in LPS-activated murine Mf (12). We hypothesized that MNSFb might regulate the production of TNFa by different molecular pathways in different cells. Although several molecules, such as Bcl-G (10), endocytosin II (11), HSPA8 (33), and Hsp60 (34), have been identified to bind with MNSFb in murine Mf, none of them could explain the promoting effect of MNSFb on promoting TNFa expression. Thus, we identified candidates in the BioGRID database (Supplementary Figure S3) and found that MNSFb potentially binds to RC3H1, which could inhibit the expression of TNFa by degrading its mRNA (24,35), and as an immune regulator, RC3H1 is involved in T-cell activation (36).
We supposed that, as a ubiquitin-like protein, MNSFb might eliminate the inhibitory effect of RC3H1 on TNFa expression by degrading the RC3H1 protein. Consistently, the direct interaction between MNSFb and RC3H1 was subsequently confirmed, and knockdown of RC3H1 expression in human Mf could lead to increased TNFa expression ( Figure 5). More encouragingly, a significantly decreased level of the RC3H1 protein was detected in the dMf from RPL patients ( Figure 6B), suggesting that the increased MNSFb expression in dMf might promote the production and secretion of TNFa by binding to RC3H1, and the secreted TNFa could inhibit the invasion of EVTs, resulting in early pregnancy loss. Furthermore, we also observed that the secretive level of MNSFb protein was simultaneously increased in dMf from RPL patients (Supplementary Figure S5). The increased secretion of MNSFb in dMf might have synergistic effects with secreted TNFa on invasive activities of EVTs during early pregnancy, but need further investigation.  In addition, although it was found that dMf had no effect on the invasion of primary EVTs (37), it was also reported that LPSstimulated peripheral blood monocytes could inhibit HTR-8/SVneo cell invasion by the secretion of TNFa (38), and M2 macrophages showed an enhanced promotion effect on trophoblast cell motility (39); these suggest that the dMf might inhibit EVTs invasion in early pregnancy. However, in this study, we found that the dMf from normal women in early pregnancy exerted a stimulatory effect on HTR-8/SVneo cell motility ( Figure 4C). DMf not only secrete TNF-a and IL-10, both of which could inhibit trophoblast motility, but also IL-1b and IL-8, both of which could stimulate trophoblast motility (40,41); enhanced EVT motility would promote their invasion into endometrial tissue, whereas reduced EVT motility would be required to prevent their excessive invasion; thus, we hypothesized that dMf might promote the motility of EVTs in the early stage of the first trimester, but inhibit their motility in the late stage of the first trimester to guarantee the appropriate invasion of EVTs into maternal uterine tissues. Such an exquisite and complex regulation of EVTs invasion by dMf is worthy of further investigation.
It should be noted that the decidual tissue samples from RPL patients were collected after the death of the fetus; thus, the differences in MNSFb expression and the ratio of the CD11c hi/ CD11c low subsets might be consequences of abortion instead of its pathogenic causes. In addition, we fully understood that, since MNSFb expression was increased in RPL dMf, we should observe the effect of upregulated MNSFb expression on the functions of macrophages. However, although we have successfully established the MNSFb overexpression HTR8/ SVneo cell model (17), we failed to establish MNSFboverexpressing cell models in both Thp1-derived Mf and T-HESCs for unknown reasons. Thus, in future investigations, we should establish a reasonably large prospective cohort of women in early pregnancy as well as a macrophage-specific MNSFb gene knock-in mouse model to systematically explore the roles of MNSFb in the pathogenesis of RPL.
In summary, it was found in this study that the MNSFb expression level and the proportion of CD11c hi cells among dMf, as well as the TNFa production and secretion level in dMf, were significantly increased in RPL patients. In vitro, RPL dMf showed a TNFa-mediated inhibitory effect on the invasion of HTR8/SVneo cells in a paracrine manner. In human Mf, MNSFb could promote the TNFa production by binding to RC3H1. Furthermore, the RC3H1 protein level in RPL dMf was significantly decreased. These data suggested that MNSFb played important roles in human dMf at least partially via the RC3H1-TNFa pathway. The abnormally increased MNSFb expression in human dMf might lead to early pregnancy loss by inducing the polarization of dMf toward a proinflammatory phenotype and promoting the secretion of TNFa at the maternal-fetal interface.

DATA AVAILABILITY STATEMENT
The original contributions presented in the study are included in the article/Supplementary Material. Further inquiries can be directed to the corresponding authors.

ETHICS STATEMENT
The studies involving human participants were reviewed and approved by The Medical Ethics Committees of The Second Hospital of Tianjin Medical University and The Medical Ethical Committee, Shanghai Institute for Biomedical and Pharmaceutical Technologies (former Shanghai Institute of Planned Parenthood Research) (Seal). The patients/participants provided their written informed consent to participate in this study.

AUTHOR CONTRIBUTIONS
Y-LW and JW designed and supervised the study, integrated the data, and revised the manuscript. X-XZ, Y-PH and LY performed experiments of the flow cytometry, WB, qPCR, co-IP, cell culture, and transwell assay. YG contributed to the collection of clinical samples. QY participated in immunofluorescence staining and confocal microscopy. W-WG contributed to data analysis. All authors contributed to the article and approved the submitted version.

ACKNOWLEDGMENTS
We are grateful to Prof. Hongmei Wang, State Key Laboratory of Stem Cell and Reproductive Biology, Institute of Zoology, Chinese Academy of Sciences, Beijing, China, for her kind guidance on the isolation of CD11c high dMf and CD11c low dMf by flow cytometry.

SUPPLEMENTARY MATERIAL
The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fimmu.2021. 691908/full#supplementary-material (Control dMf: dMf isolated from decidual tissues from normal women in early pregnancy; Control: total decidual cells isolated from decidual tissues of normal women in early pregnancy; RPL: total decidual cells isolated from decidual tissues from RPL patients, *P < 0.05).