M2 macrophage‐derived G‐CSF promotes trophoblasts EMT, invasion and migration via activating PI3K/Akt/Erk1/2 pathway to mediate normal pregnancy

Abstract Trophoblasts are important parts of the placenta and exert vital roles in the maternal‐foetal crosstalk, and sufficient trophoblasts migration and invasion is critical for embryo implantation and normal pregnancy. Macrophages, as the major components of decidual microenvironment at maternal‐foetal interface, can interact with trophoblasts to participate in the regulation of normal pregnancy. Previously, our group have demonstrated that trophoblasts could induce macrophages polarization to M2 subtype by secreting interleukin‐6 (IL‐6); however, the understanding of macrophages regulating the migration and invasion of trophoblasts is limited. In the present study, we used the co‐cultured model to further investigate the effects of macrophages on trophoblasts migration and invasion. Our results showed that co‐culture with macrophages promoted epithelial‐to‐mesenchymal transition (EMT) of trophoblasts, thereby enhancing their migrative and invasive abilities. Further experiments revealed that M2 macrophage‐derived G‐CSF was a key factor, which promoted the EMT, migration and invasion of trophoblasts via activating PI3K/Akt/Erk1/2 signalling pathway. Clinically, G‐CSF was highly expressed in placental villous tissues of normal pregnancy patients compared to patients with recurrent spontaneous abortion, and its expression level was significantly correlation with EMT markers. Taken together, these findings indicate the important role of M2 macrophages in regulating trophoblasts EMT, migration and invasion, contributing to a new insight in concerning the crosstalk between macrophages and trophoblasts in the establishment and maintenance of normal pregnancy.


| INTRODUC TI ON
Normal pregnancy, as a complex physiological process, 1 depends not only on the proper maternal-foetal crosstalk and immune regulation, but also on trophoblast development. 2 Increasing and accumulating evidence indicates that trophoblast invasion promotes maternal placental blood flow and maternal spiral artery remodelling, thereby establishing favourable microenvironment for embryo implantation. 3 In contrast, impairment of trophoblast migration and invasion can induce utero-placental insufficiency, which eventually leads to a series of pregnancy complications, such as foetal growth restriction, pre-eclampsia and recurrent spontaneous abortion (RSA). [4][5][6][7] Therefore, exploring the factors affecting trophoblasts invasion and migration is of great significance for further understanding of the normal pregnancy process and the pathogenesis of pregnancy complications. Epithelial-to-mesenchymal transition (EMT), as a cellular process in which cells lose their epithelial characteristics and acquire mesenchymal features, has been demonstrated to play a vital role in maintaining the migrative and invasive abilities of trophoblasts in recent years. 8,9 We and others previously found that inhibition of trophoblasts EMT programme would impair its ability to invade and migrate, resulting in the occurrence of RSA. 7,10 However, the regulatory mechanisms of trophoblasts EMT still remain elusive.
The initiation and maintenance of EMT status are a multi-step complex processes cooperating by multiple factors. 9,11 Emerging studies have demonstrated that intercellular interactions in the maternal-foetal interface microenvironment play an important role in regulating EMT of trophoblasts. 9,12 During pregnancy, trophoblasts gain the advantage of invasion and migration by interacting with immune cells, thereby establishing a unique maternal-foetal microenvironment that contributes to foetal survival and development. 9,13 Macrophages, as the second largest group of immune cells at the maternal-foetal interface, 2 have been demonstrated to play critical roles in embryo implantation, embryonic development, placental formation and delivery processes. [13][14][15] Previous studies have showed that macrophages can establish 'crosstalk' with trophoblasts in the maternal-foetal interface microenvironment via a complex cytokine-based connection. On one hand, macrophages can secrete amounts of soluble mediators to regulate the biological behaviours of trophoblasts. 6,13,[16][17][18][19] On the other hand, macrophages can respond to various factors produced by trophoblasts to convert polarization status, thereby playing different biological functions. [20][21][22] Previously, we have demonstrated that trophoblasts could induce macrophages polarization to M2 subtype by secreting interleukin-6 (IL-6), thereby modulating the process of normal pregnancy. 23 Although the interplay between trophoblasts and macrophages has been established at the maternal-foetal interface, the understanding of macrophages regulating trophoblasts EMT programme is limited.
In the present study, we utilized an in vitro co-culture model to investigate the effects of macrophages on EMT, invasion and migration of trophoblasts. The results showed that co-culture with macrophages promoted EMT programme of trophoblasts, thereby enhancing their migration and migration. Further experiments revealed that M2 macrophage-derived G-CSF was a key factor, which promoted the EMT, migration and invasion of trophoblasts via activating PI3K/Akt/Erk1/2 signalling pathway. Clinically, G-CSF was highly expressed in placental villous tissues of normal pregnancy patients compared to patients with RSA, and its expression level was significantly correlation with EMT markers. Taken together, these findings indicate the important role of M2 macrophages in regulating trophoblasts EMT, migration and invasion, contributing to a new insight in concerning the crosstalk between macrophages and trophoblasts in the establishment and maintenance of normal pregnancy.

| Patients and tissue samples
Women conducted induced abortion for non-medical reasons were chose for control group, and patients suffered from recurrent spontaneous miscarriage for two or more times were regarded as RSA group. Twenty-one RSA patients and 28 normal pregnant women from Renmin Hospital of Wuhan University were included in the study. RSA patients with characteristics including endocrine or metabolic diseases (such as diabetes, hyperthyroidism, hypothyroidism), uterine abnormality, abnormal karyotype or infection (according to the leucorrhoea routine examination) were excluded from the study. The placental villous tissues and decidual tissue samples were collected at the time of surgery, fixed in 4% paraformaldehyde for paraffin-embedding in blocks or frozen and stored in liquid nitrogen. The procedure was performed with the approval of the internal review and ethics boards of Renmin Hospital of Wuhan University, and informed consent was obtained from all included patients.
The baseline characteristics of the patients were summarized in Table 1.

| Cell culture and reagents
The human monocyte cell line THP-1 was cultured and grown in RPMI-

| RNA isolation and quantitative real-time PCR (RT-PCR)
The total RNA was isolated using the TRIzol Reagent (Invitrogen,

| Western blot
Protein extraction and Western blot were performed as we previously reported. 7 The following primary antibodies were used: anti-E-cadherin

| Cytokine assays
The cytokine antibody array agented by Wayen Biotechnologies were expressed as pictograms per millilitre.

| Wound healing assay
The wound healing assay was used to evaluate the migration ability of trophoblasts. A wound was made by dragging the plastic pipette tip across the cell surface when cells were grown to 80% confluence in 24-well plates, and migrating cells at the wound front were photographed after 24h and 48h. The area of the wound was measured with Image J software (NIH, USA).

| Immunohistochemistry (IHC)
For IHC, serial sections from the human placental villous tissues and the decidual tissues were obtained. IHC was conducted as previously described. 24 The following primary antibodies were used: anti

| Co-culture with macrophages promote EMT, migration and invasion of trophoblasts
To determine the effect of macrophages on trophoblasts, we uti-  To evaluate whether G-CSF was critical for the EMT process of HTR-8, recombinant G-CSF was added in the culture medium of HTR-8. The results showed that G-CSF significantly reduced the expression of E-cadherin, while increased the expression of Vimentin in both protein and mRNA level ( Figure 2D-E).

| G-CSF is newly identified and validated as a key factor for M2 macrophage-induced trophoblasts EMT, migration and invasion
Furthermore, a G-CSF neutralizing antibody was used to confirm the regulatory effect of G-CSF in HTR-8. After adding the G-CSF neutralizing antibody into the co-culture model, the expression of Vimentin was decreased, while E-cadherin was increased when compared with G-CSF treatment group ( Figure 2D-E). Consistently, the G-CSF neutralizing antibody decreased the migratory ( Figure 2F) and invasive capacities ( Figure 2G) of HTR-8. Collectively, these findings indicate that M2 macrophage-derived G-CSF is newly identified and validated as a key factor for M2 macrophage-induced trophoblasts EMT.

| M2 Macrophage-derived G-CSF promotes EMT, migration and invasion of trophoblasts via activating PI3K/Akt/Erk1/2 pathway
To determine which downstream signals in HTR-8 responded to G-CSF secretion by M2 macrophages, we focused on PI3K, Erk1/2, Akt and P-38 pathway, which have been reported to be activated in trophoblasts treated with G-CSF. [25][26][27][28] The results found that the and p-Akt ( Figure 3B). In addition, both of LY294002 and PD98059 attenuated the promoting role of G-CSF or co-culture on the expression of Vimentin and also the inhibitory effect on E-cadherin ( Figure 3B). Moreover, the wound healing assay showed that in the presence of the Erk1/2 inhibitor or PI3K inhibitors, the gap closure was significantly reduced, compared with treatment of G-CSF or coculture model ( Figure 3C). Similar results were observed in invasion assay ( Figure 3D). These data demonstrate that M2 macrophage-de-

| High expression of G-CSF in placental villous tissues is associated with normal pregnancy
Furthermore, we evaluated the expression of G-CSF in the placental villous and decidual tissues from normal pregnancy. The results found that G-CSF was expressed on decidual stromal cells, with membrane patterns ( Figure 4A). To confirm the connection between G-CSF and M2 macrophages in the decidua, we investigated the expressions of G-CSF and CD163 by double immunofluorescence staining. As shown in Figure 4B, the results showed prolific G-CSF protein on CD163 + macrophages.

F I G U R E 4 High expression of G-CSF in placental villous
Then, IHC was performed to analysis the expression of G-CSF,

E-cadherin and Vimentin in the placental villous tissues in serial
sections. The results demonstrated that the expression of G-CSF was observed in placental villous tissues, and the G-CSF + cells were mainly located on cytotrophoblasts, a few on syncytiotrophoblasts ( Figure 4D). Furthermore, high level of G-CSF was associated with less E-cadherin and more Vimentin ( Figure 4D). Pearson test showed that there was a statistically significant negative correlation between the expression levels of G-CSF and E-cadherin and survival 28 ; however, the mechanism was not clear. G-CSF was involved in several physiological and pathological processes by complicated mechanisms, such as activation of MAPK and JAK2/ STAT3 pathway. 28,44,45 Moreover, it has been reported that G-CSF could increase migration of Swan71 cell via PI3K and MAPK activation. 27 In the present study, our results demonstrated that trophoblast-educated M2 macrophages promoted the EMT process, migration and invasion of trophoblasts via PI3K/Akt/Erk1/2 signal pathway. In addition, we observed a positive correlation between the expression of G-CSF and the percentage of CD163 + cells, which was in accordance with the result that high levels of G-CSF were associated with increased CD163 + macrophages in triple-negative breast cancer. 46 G-CSF has been regarded as a M2 marker, 47 and our previous study has confirmed that trophoblasts educated macrophages to M2 subtype, 23 so we considered that trophoblast-educated M2 macrophages secreted G-CSF to promote the invasion and migration of trophoblasts. In addition, a considerable number of G-CSF + CD163cells were found in the decidua, indicating that in addition to macrophages, other cells in the decidua, such as NK cells, were also sources of G-CSF. 48 The present study demonstrated high expression of G-CSF in placental villous tissues is associated with normal pregnancy, which highlights the importance of macrophages and trophoblasts crosstalk at the maternal-foetal interface. Given that the role of insufficient invasion of trophoblasts and decreased M2 macrophages in RSA has been confirmed by others' and our previous study, 6,7,17,49 we could speculate that decreased proportion of M2 macrophages reduces the level of G-CSF to a certain extent and then participates in the occurrence of RSA by affecting the invasion and migration of trophoblasts. However, the lower G-CSF in the placental villous tissues of RSA might be a consequence of miscarriage, as the pregnancy might have failed some time before the sample collection, which might be partially verified by animal experiments.

| CON CLUS ION
In summary, our current findings disclose the important role of macrophages in regulating the EMT programme, migration and invasion of trophoblasts, contributing to a new insight concerning the crosstalk between macrophages and trophoblasts in the establishment and maintenance of normal pregnancy.

CO N FLI C T O F I NTE R E S T
The authors declare that there is no conflict of interest.

DATA AVA I L A B I L I T Y S TAT E M E N T
All data generated or analysed during this study are included in this published article.