LPMs Migrate Into Injured Endometrium

To examine the alterations in inflammation-associated immune cells within the injured endometrium, we induced endometrial damage in mice using an electric scratching tool, as previously described [11], at various time points: 3, 6, 12, 24, and 48 h prior to sample collection (Fig. 1A). Flow cytometry results showed that, after endometrium was injured, the proportion of neutrophils (CD11bintLy6G+) in uterus increased to the peak at 6 h and gradually disappeared (Fig. 1B). The increase in monocytes (CD11b+CCR2+) and uterine resident macrophages (CD11b+CD102−F4/80int) were slightly later, reach the peak at 12 h (Fig. 1C and D). Remarkably, there were a group of cells characterized as CD11bhigh, CD102+ and F4/80high. Their phenotype differed from that of neutrophils but increased at a similar rate (Fig. 1E). CD102 is the specific marker of cavity resident macrophages [27]. Given that the uterus is located in the peritoneal cavity, we supposed that these cells might be peritoneal macrophages.

Fig. 1

LPMs migrate into injured endometrium. A The endometrium of Balb/c female mice (8–10 weeks old) was injured using the electric scratching tool and samples were collected at 3-, 6-, 12-, 24- and 48-h post-injury. B-E The statistical figures of percentages of neutrophils (CD11bint Ly6G+), monocytes (CD11b+CCR2+), uterine resident macrophages (CD11b+F4/80intCD102−) and LPMs (CD11b+F4/80+CD102+) in uterus determined by flow cytometry assay. F The statistical figures of percentages of LPMs (CD11b+CD102+F4/80+) in peritoneal cavity determined by flow cytometry assay. G Flow cytometry assay data of the F4/80 and GATA6 expressing levels on cells from above gates. Among them, Q2 represented LPMs. H Immunofluorescence staining images (cross section) of CD102 and GATA6 expressions in uterus at 0 h (control) and 6 h after injured. CD102 (red) was located on the cell membrane and GATA6 (green) was located in the nucleus. Bar = 50 μm. I The statistical figures of percentages of LPMs (CD11b+CD102+F4/80+GATA6+) in uterus determined by flow cytometry assay. J Immunofluorescence staining images (cross section) of TUNEL in uterus at 0 h (control) and 6 h after injured. Bar = 100 μm. K The level of IL-1β, IL-6 and TNF-α in serum detected by ELISA assay. Bar graphs show the mean ± SEM; unpaired Student’s t test was used to compare the experimental groups. Bar = 50 μm. n = 4–6; *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001

Next, we measured the proportions of LPMs and SPMs in peritoneal cavity at each moment of endometrial injury by flow cytometry. The CD11b + myeloid cells in the peritoneal CD102 (red) was located on the cell membrane and GATA6 (green) was located in the nucleus. They were divided into three groups based on the levels of CD102 and MHCII, including LPMs (CD102+), SPMs (CD102−MHCII+) and an undefined subgroup called new cells (CD102−MHCII−) (Supplementary Fig. 1B). Among them, LPMs had the highest proportion (> 90%). Moreover, LPMs exhibited F4/80high and GATA6+ specifically, which was different from the other two populations (Supplementary Fig. 1 C and Supplementary Fig. 1D). Flow cytometry results showed that, after damage to the endometrium, the proportion of LPMs in peritoneal cavity was decreased at 3 h and came to the valley bottom at 6 h (Fig. 1F). Then the proportion gradually returned to normal. This pattern of change was just corresponded to that of CD11b+CD102+F4/80+ cells in injured endometrium (Fig. 1E). In addition, the proportion of SPMs was increased at 24 h after damage (Supplementary Fig. 1E) and new cells were increased to the peak at 6 h (Supplementary Fig. 1 F). These data indicated that LPMs may be the cell population which migrated from peritoneal cavity into the damaged endometrium.

To validate this hypothesis, we conducted a detailed characterization of CD11b+, CD102+, and F4/80+ cells within the injured endometrium and discovered that these cells expressed GATA6, a specific marker for LPMs (Supplementary Fig. 1 A, Fig. 1G and H). Our flow cytometry analysis revealed that, subsequent to endometrial injury, the proportion of LPMs in the uterus increased notably at 3 h, reaching a peak at 6 h (Fig. 1I), followed by a gradual decline. Correspondingly, this treatment caused the increase of dead cells number in the endometrium (Fig. 1J) and the upregulation of inflammatory factors (IL-1β、IL-6 and TNF-α) in serum (Fig. 1K). All the above experiments were carried out on Balb/c mice. A similar phenomenon was also observed in C57BL/6 mice (Supplementary Fig. 2 A, B, C). Overall, these findings indicate that LPMs could transfer from peritoneal cavity to injured endometrium. This process was as rapid as the infiltration of neutrophils, and could be observed in different mouse strains.

LPMs Absence Enhances Inflammation and Accumulation of Dead Endometrial Cells

Our findings underscored the capacity of LPMs to swiftly migrate to the site of uterine injury. To delve deeper into the roles of LPMs in the injured uterus, we employed a strategy to deplete LPMs by administering clodronate liposomes (CLL) intraperitoneally to the mice, with con-CLL served as the control. Notably, a substantial majority of LPMs were effectively eliminated within 48 h post-injection (Supplementary Fig. 3 A). Additionally, we observed no significant alterations in the proportions of macrophages in the uterus, liver, or bone marrow, except for a marked reduction in macrophages in the spleen (Supplementary Fig. 3B). Moreover, level of IL-6 in serum was slightly increased, whereas the levels of IL-1β and TNF-α did not change significantly (Supplementary Fig. 3 C). Taken together, these results suggested that intraperitoneal injection of CLL could deplete LPMs from the peritoneal cavity of mice without causing too many additional effects.

We subsequently depleted LPMs in the peritoneal cavity using CLL 24 h prior to experimentally inducing endometrial damage in mice with our electric scratching tool, with samples collected 12 h later (Fig. 2A). Flow cytometry results indicated that CLL effectively hindered the migration of LPMs from the peritoneal cavity to the uterus, whereas con-CLL did not exert such an effect (Fig. 2B and C). The absence of LPMs prevented their accumulation in the injured uterus. Furthermore, the lack of LPMs exacerbated both the elevation of inflammatory factors in the serum resulting from uterine injury (Fig. 2D) and the accumulation of dead cells in the endometrium (Fig. 2E). Based on these observations, we hypothesized that LPMs may play a crucial role in phagocytosing dead cells, mitigating inflammation, and facilitating endometrial repair.

Fig. 2

LPMs absence enhances inflammation and accumulation of dead endometrial cells. A The schematic diagram of modelling process. Balb/c female mice (8–10 weeks old) were injected intraperitoneally with 100 μL PBS/con-CLL/CLL 24 h in advance. Next, endometrium of mice was injured using the electric scratching tool for 12 h. B The statistical figures of percentages of LPMs (CD11b+CD102+F4/80+) in peritoneal cavity determined by flow cytometry assay. C The statistical figures of percentages of LPMs (CD11b+CD102+F4/80.+) in uterus determined by flow cytometry assay. D The level of IL-1β, IL-6 and TNF-α in serum detected by ELISA assay. E Immunofluorescence staining images (cross section) of TUNEL in uterus. Bar graphs show the mean ± SEM; unpaired Student’s t test was used to compare the experimental groups. Bar = 100 μm. n = 4–6; *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001

LPMs Relieve Inflammation and Phagocytose Dead Cells in Injured Endometrium

To validate our hypothesis regarding the function of LPMs in the injured endometrium, we conducted a LPMs transplantation experiment. In this study, we accessed the peritoneal cavity of the mice and induced endometrial damage using a syringe needle. This procedure mirrored the establishment of the mechanical damage IUA mouse model described in our earlier publications [11, 28] (Fig. 3A). Our findings revealed that the introduction of LPMs was capable of abrogating the upregulation of serum inflammatory factors (IL-1β, IL-6, and TNF-α) triggered by endometrial injury (Fig. 3B). We also noted that LPMs were colocalized with dead cells within the endometrium (Fig. 3C), indicating their capacity to phagocytose these cells in that location. To further substantiate this observation, we conducted an in vitro phagocytosis assay. Given that endometrial stromal cells (ESCs) constitute the primary cell population in the endometrium, we utilized ESCs derived from mouse uteri as the target for phagocytosis. ESCs were seeded in well plates and subsequently treated with ethanol for 5 min to induce apoptosis (Supplementary Fig. 4 A and B). The immunofluorescence images obtained in vitro demonstrated that LPMs engaged in trogocytosis, shredding, and endocytosing apoptotic ESCs, confirming their ability to phagocytose these cells (Fig. 3D). Collectively, these results established that LPMs can alleviate inflammation and phagocytose dead cells in the injured endometrium, which is crucial for uterine repair.

Fig. 3

LPMs relieve inflammation and phagocytose dead cells in injured endometrium. A The schematic diagram of modelling process. Balb/c female mice (8–10 weeks old) were carried out surgery and intrauterine injected LPMs (8 × 104). Samples were harvested after 12 h. B The level of IL-1β, IL-6 and TNF-α in serum detected by ELISA assay. C Immunofluorescence staining images (cross section) of CD102 (pink), GATA6 (green) and TUNEL (yellow) expressions in uterus from ‘injury + LPMs’ group mice. D Immunofluorescence staining images of LPMs (F4/80+, red) and ESCs (CFSE labeled, green) isolated in vitro. Bar graphs show the mean ± SEM; unpaired Student’s t test was used to compare the experimental groups. Bar = 50 μm. n = 3–5; *P < 0.05, **P < 0.01

Estradiol can Accelerate the Migration of LPMs to the Damaged Endometrium

We identified the estrous periods according to the cell morphology in mouse vaginal secretion smear of mice. In the vaginal secretions of proestrus mice, there were mainly nucleated, round, large cells, i.e. the normal endometrial epithelial cells. In the vaginal secretions of estrus mice, there were numerous large scaly cells, i.e. the keratinized endometrial epithelial cells (Supplementary Fig. 4 A). Indeed, the level of E2 was higher in the serum of proestrus mice than that in the serum of estrus mice (Supplementary Fig. 4B). We then divided mice into estrus and proestrus groups by the above characteristics, and damaged their endometrium with our electric scratching tool at 6 h before sample harvesting respectively (Fig. 4A). The results showed that the migration efficiency of LPMs in proestrus mice was significantly higher than that in estrus mice, although their migration to the uterus occurred in both stages of mice (Fig. 4B and C). Moreover, at 6 h after endometrium injury, the proportion of LPMs in peritoneal cavity of mice was inversely proportional to the level of E2 (Fig. 4D), but directly proportional in uterus (Fig. 4E). The serum levels of inflammation factors in proestrus mice were significantly lower than those in estrus mice (Fig. 4F). These data suggested that the different E2 levels may influence the migration of LPMs into the impaired uterus to suppress inflammation.

Fig. 4

The migration efficiency of LPMs into injured uterus is influenced by E2 level. A The schematic diagram of modelling process. Balb/c female mice (8–10 weeks old) were divided into estrus group and proestrus group. The endometrium was injured using the electric scratching tool for 6 h. B The statistical figures of percentages of LPMs (CD11b + CD102 + F4/80 +) in peritoneal cavity determined by flow cytometry assay. C The statistical figures of percentages of LPMs (CD11b + CD102 + F4/80 +) in uterus determined by flow cytometry assay. D Correlation analysis graph of percentage of LPMs in peritoneal cavity and E2 of serum. E Correlation analysis graph of percentage of LPMs in uterus and E2 of serum. F The level of IL-1β, IL-6 and TNF-α in serum detected by ELISA assay. Bar graphs show the mean ± SEM; unpaired Student’s t test was used to compare the experimental groups. n = 4–8; *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001

E2/ER-β Axis is Necessary for the Migration of LPMs to Apoptotic ESCs

Above results hinted E2 may drive the migration of LPMs to damaged uterus. To further explore roles of E2 in migration of LPMs, we intraperitoneally injected E2 to estrus mice, and then damaged their endometrium with the electric scratching tool at 6 h before sample harvesting (Fig. 5A). Dimethyl sulfoxide (DMSO) was used as control. As showed in flow cytometry assay results, E2 promoted the accumulation of LPMs in impaired uterus, while there was no difference in the percentage of LPMs in peritoneal cavity between ‘Injury + DMSO’ group and ‘Injury + E2’ group (Fig. 5B and C). In addition, the transcriptome RNA sequencing (RNA-seq) results showed that E2 increased the expression of many genes related to cytoskeleton, adhesion and migration in LPMs (Fig. 5D). Indeed, we found that E2 could elongate the cytoskeleton and increase pseudopodia of LPMs in vitro (Fig. 5E), which were both necessary for cells to migrate. To better simulate the migration of LPMs to the damaged uterus, an in vitro migration assay was designed. ESCs are considered as the main cells in endometrium. ESCs were thus seeded in well plates, and then treated with ethanol for 5 min to induce apoptosis. Correspondingly, LPMs were seeded in the upper transwells for co-culture with apoptotic ESCs. The results showed that E2 indeed promote the migration of LPMs to apoptotic ESCs in a dose-dependent manner (Fig. 5F).

Fig. 5

E2/ER-β axis is necessary for the migration of LPMs to apoptotic ESCs. A The schematic diagram of modelling process. The endometrium of Balb/c female mice (8–10 weeks old) which were in estrus period was injured using the electric scratching tool for 6 h. B The statistical figures of percentages of LPMs (CD11b + CD102 + F4/80 +) in peritoneal cavity determined by flow cytometry assay. C The statistical figures of percentages of LPMs (CD11b + CD102 + F4/80 +) in uterus determined by flow cytometry assay. D RNA-seq analysis of LPMs isolated from mouse peritoneal cavity treated with or without E2 (10 nM) for 24 h. The heat map represented the differential expression of genes, sorted from top to bottom according to fold change (log2). E Immunofluorescence staining images of LPMs isolated from mouse peritoneal cavity in vitro. LPMs were treated by E2 (10 nM) for 24 h. Phalloidin (green) stained the cytoskeleton. Bar = 30 μm. F Crystal violet staining images of LPMs which migrated from inside of the transwell to the bottom of that. LPMs in transwell were pre-treated by E2 (1, 10 or 100 nM) or not for 24 h. The transwell within LPMs was cocultured with Ap-ESC for 6 h. Ap-ESC represents the apoptotic ESCs. Bar = 100 μm. The column graph was the statistic of the purple cell number. G The expression level (2∆Ct) of Esr-1, Esr-2 and gpr30 in LPMs (isolated from estrus or proestrus mice) or 3T3-L1 cells detected by qRT-PCR assay. H Relative expression of mRNAs of Esr-2 in LPMs from those mice detected by qRT-PCR assay. The LPMs isolated from mice were treated by E2 (1, 10 or 100 nM) or not for 24 h. I Immunofluorescence staining images of LPMs isolated from mouse peritoneal cavity in vitro. The LPMs were treated by E2 (10 nM) or not for 24 h. J Crystal violet staining images of LPMs which migrated from inside of the transwell to the bottom of that. LPMs in transwell were pre-treated by fulvestrant (Ful, 100 nM) or not for 1 h and treated by E2 (10 nM) or not for 24 h. The transwell within LPMs was cocultured with Ap-ESC for 6 h. Bar = 100 μm. The column graph was the statistic of the purple cell number. Bar graphs show the mean ± SEM; unpaired Student’s t test was used to compare the experimental groups. n = 3; *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001

Accumulated evidence has revealed that there are three types of estrogen receptors (ERα, ERβ, and GPR30). Importantly, LPMs were detected to express only ER-β (gene named Esr-2) specifically, but not ER-α (gene named Esr-1) and GPR30 (gene named Gpr30) (Fig. 5G). Moreover, E2 (10 nM) significantly up-regulated the expression of ER-β in LPMs (Fig. 5H and I), indicating ER-β in LPMs specifically responded to the stimulation of E2. Estrogen receptor antagonist fulvestrant (Ful) is showed to completely inhibit estrogen-mediated changes in gene transcription. In in vitro migration assay, Ful eliminated the promotion of E2 on the migration of LPMs to apoptotic ESCs (Fig. 5J). Combined with all above results, we concluded that E2/ER-β axis was necessary for the migration of LPMs to apoptotic ESCs.