2025 年 48 巻 10 号 p. 1503-1513
Endometriosis, a prevalent gynecological disorder marked by ectopic growth of endometrial-like tissue, demonstrates malignant tumor-like properties including aggressive adhesion and invasiveness. Emerging evidence implicates roundabout guidance receptor 3 (ROBO3) in cellular pathophysiology, yet its role in endometriosis remains unexplored. In this study, we first found abnormally high ROBO3 expression in endometriosis by bioinformatics analysis. Next, functional assays revealed that ROBO3 is a key regulator promoting the invasion and migration of endometriotic stromal cells (ESCs) in vitro. Mechanistically, ROBO3 activates the Wnt/β-catenin signaling pathway, evidenced by increased phosphorylation of glycogen synthase kinase 3β, nuclear β-catenin accumulation, and upregulated c-myc expression. Pharmacological inhibition of Wnt/β-catenin with MSAB (5 μM) reversed ROBO3-mediated pro-invasive and pro-migratory effects. Furthermore, we discovered that ROBO3 enhances the secretion of chemokines CCL2 and CCL5 in ESCs, which subsequently promote macrophage polarization toward the M2 phenotype, as indicated through elevated expression of interleukin-10 and Arg-1. Collectively, our findings elucidate a dual mechanism whereby ROBO3 drives endometriosis progression through both intrinsic activation of Wnt/β-catenin signaling and extrinsic modulation of tumor-associated macrophages, underscoring ROBO3 as a promising therapeutic target for endometriosis.
Endometriosis, an estrogen-dependent disorder marked by ectopic implantation of hormone-sensitive endometrial tissue, represents a prevalent gynecological condition affecting 6–10% of reproductive-aged women.1) This pathophysiological state exhibits a heterogeneous clinical spectrum ranging from asymptomatic presentation to multiple debilitating symptoms, including chronic pelvic pain, cyclical dysmenorrhea, dyspareunia, urinary dysfunction, and subfertility.2) Endometriosis is considered as a benign condition, yet it has the malignant manifestations of strong immune evasion, survival, epithelial–mesenchymal transition (EMT), adhesion, and invasion.3,4) Current main treatments for endometriosis focus on suppressing estrogen secretion, inhibiting cell proliferation, and reducing inflammation.5) However, hormonal treatments are not recommended for long-term use and disease recurrence in the clinic is very common.6,7) Thus, nonhormonal therapeutic approaches targeting endometriosis are required for improving patient conditions.5)
Emerging evidence highlights the pivotal involvement of ectopic endometrial stromal cells (ESCs) in endometriosis pathogenesis.8) The multistep progression of this disorder is driven by aberrant cellular dynamics in displaced endometrial tissue, which is characterized by enhanced peritoneal adhesion capacity, accelerated proliferation rates, and enhanced invasive potential.9) Restraining the abnormal proliferation, migration, and invasion might be an effective treatment for endometriosis. Additionally, immune dysregulation within the endometriotic microenvironment plays a pivotal role in disease pathogenesis.10) A hallmark of this aberrant immunological milieu is the pronounced macrophage recruitment observed in both peritoneal fluid and ectopic lesions. M1 macrophages are associated with the inhibition of endometriosis development, while M2 macrophages are linked to its promotion. Notably, the M2 phenotype is predominant in the peritoneal environment of women with endometriosis.11)
Roundabout guidance receptor (ROBO) family proteins are mainly involved in various signaling pathways, such as cell proliferation, cell motility, and chemotaxis of inflammatory cells.12,13) ROBO3, as a transmembrane protein, was reported to play an important role in the ovary.14) According to the relevant studies, the expression of ROBO3 is dysregulated in various diseases.15,16) ROBO3 overexpression facilitated tumor cell growth and invasion by activating the Wnt/β-catenin pathway.16) Furthermore, studies have demonstrated that the activation of the Wnt/β-catenin pathway is associated with the proliferation, migration, and invasion of ESCs.17) This activation has also been implicated in exacerbating endometriosis progression.18–20) Moreover, it has been reported that ROBO3 positively correlates with M2-polarized macrophage levels,13) and the Wnt/β-catenin pathway plays a key role in the recruitment and polarization of macrophages.21) Therefore, we hypothesized that ROBO3 might exert its pro-disease effect in endometriosis by influencing the invasion and migration of ESCs and the polarization of M2 macrophages by upregulating Wnt/β-catenin signaling.
This study identifies ROBO3 as a key pathogenic driver in endometriosis, demonstrating its aberrant overexpression in ectopic lesions. Our findings describe a dual mechanism whereby ROBO3 drives endometriosis progression via both intrinsic activation of Wnt/β-catenin signaling and extrinsic modulation of tumor-associated macrophages, underscoring ROBO3 as a potential clinical marker of endometriosis.
Microarray datasets GSE168902 and GSE25628, focusing on endometriosis, were retrieved from the Gene Expression Omnibus database. Differentially expressed genes (DEGs) were identified using the thresholds of p < 0.05 and |Log2(fold change)| > 1. A heatmap was generated to visualize the DEGs in both datasets. To identify key genes consistently dysregulated across the 2 datasets, a Venn diagram analysis was performed to pinpoint genes that were either jointly upregulated or downregulated in GSE25628 and GSE168902. To explore the function and potential pathways of genes associated with endometriosis, gene ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analyses were performed. A focused heatmap specifically depicted the expression profiles of the ROBO gene family.
ESCs’ IsolationThe research was approved by the Ethics Committee of the General Hospital of Northern Theater Command, and written informed consent was obtained from each patient. Ectopic endometrial tissues were obtained from women who had regular menstrual cycles and had not received hormone therapy in the 3 months leading up to laparoscopic surgery. The samples were stored in sterile phosphate-buffered saline (PBS) supplemented with 1% penicillin and streptomycin. Endometrial tissues were minced and digested with collagenase. Then, they were inoculated and cultured in Dulbecco’s modified Eagle’s medium/F12 (DMEM/F12) medium (BL305A, Biosharp, Anhui, China) containing 10% fetal bovine serum (FBS) (11011-8611, Zhejiang Tianhang Biotechnology, Zhejiang, China).
Adenovirus Vector Infection and Cell TreatmentFor ROBO3 overexpression, the coding sequence of the ROBO3 was subcloned into the viral vector as overexpressing ROBO3 (oeROBO3). For ROBO3 silencing, 2 strands of short hairpin RNA (shRNA) targeting ROBO3 were subcloned into viral vectors as shROBO3-1 and shROBO3-2.
The medium of ESCs was replaced with adenovirus-containing medium, and cultivation was continued for 48 h. After adding 5 μM Wnt/β-catenin signaling pathway inhibitor MSAB (Macklin, China) for 24 h, the ESCs and the supernatant after centrifugation were collected for analysis.
THP-1 cells (iCell-h213, iCell Bioscience (Shanghai, China)) were cultured in RPMI-1640 medium (31800, Solarbio, Beijing, China) containing 10% FBS at 37°C with 5% CO2 in the incubator. THP-1 cells were differentiated into M0 macrophages by treatment with 100 ng/mL phorbol 12-myristate 13-acetate (P849986, Macklin) for 48 h. Infected ESCs were cultured in fresh medium for 48 h. Then, the culture medium was diluted with fresh medium (1 : 1) and added to the THP-1 cells. Cells were collected after 24 h for further testing.
ImmunofluorescenceCells were separately fixed in chamber slides. Then, the cells were permeabilized with 0.1% Triton-X 100 (ST795, Beyotime, Shanghai, China) for 30 min and blocked with bovine serum albumin (A602440-0050, Sangon Biotech (Shanghai), Shanghai, China). Cells were probed with the primary antibodies (1 : 100) against Arg-1 (A4923, ABclonal, Woburn, MA, U.S.A.), cytokeratin (DF2554, Affinity, San Francisco, CA, U.S.A.), vimentin (A19607, ABclonal), and β-catenin (A11932, ABclonal) overnight at 4°C, and subsequently with secondary antibody (1 : 200; ab6939, Abcam, Cambridge, U.K.) for 1 h. Upon counterstaining with 4′,6-diamidino-2-phenylindole (D106471-5 mg, Aladdin (Beijing, China)), the slides were visualized with a fluorescent microscope (BX53, OLYMPUS, Tokyo, Japan).
RT-Quantitative PCR (RT-qPCR)Total RNA was isolated from cultured cells using Tripure reagent (RP1001, Bioteke, China) and reverse transcribed into cDNA using the All-in-One First-Strand SuperMix (MD80101, Magen, New York City, NY, U.S.A.). PCR amplification of cDNA was performed using the universal 2 × Fast Taq Plus PCR Master Mix (BL1014, Biosharp), and qPCR was measured using the SYBR Green (SR4110, Solarbio). β-Actin served as the endogenous control. Relative gene expression was calculated using the 2−ΔΔCT method. The primers used were as follows: ROBO3: forward, 5′-TGCGAGACCAAAGGAAA-3′ and reverse, 5′-GGAGGATGACAGGAGGC-3′; interleukin-10 (IL-10): forward, 5′-TGAGAACCAAGACCCAGAC-3′ and reverse, 5′-CATTCTTCACCTGCTCCAC-3′; NELL2: forward, 5′-TGAACAGCGAATGAATAGA-3′ and reverse, 5′-CATCCACATACGCAAGAG-3′; CD86: forward, 5′-CTCTGGTGCTGCTCCTCT-3′ and reverse, 5′-GGGTCCAACTGTCCGAAT-3′; iNOS: forward, 5′-TGGAGACGGGAAAGAAG-3′ and reverse, 5′-GGCAAGATTTGGACCTG-3′; CD206: forward, 5′-CTGGGTTGCTATCACTCT-3′ and reverse, 5′-CAAACTTGAACGGGAAT-3′; Arg-1: forward, 5′-TTTGCTGACATCCCTAAT-3′ and reverse, 5′-TTCCGTTCTTCTTGACTT-3′; CD163: forward, 5′-GAGACTGTTAGGGAAGGTG-3′ and reverse, 5′-TGTTTG-TTGCCTGGATT-3′; β-actin: forward, 5′-TCAGGGTGAGGATGCCTCTC-3′ and reverse, 5′-CTCGTCGTCGACAACGGCT-3′.
Western BlotESCs were lysed in RIPA lysis buffer (R0010, Solarbio) supplemented with 1 mM proteinase inhibitors (phenylmethylsulfonyl fluoride; P0100, Solarbio) to extract total proteins. After adding the cytoplasmic protein extraction reagent, the samples were centrifuged at 12000 × g for 10 min to obtain the supernatant as the cytoplasmic extract. The cell pellet was mixed with nuclear protein extraction reagent to obtain the nuclear extract. Subsequently, protein samples were electrophoresed on sodium dodecyl sulfate–polyacrylamide gel electrophoresis (D1010, Solarbio) and transferred to a polyvinylidene fluoride membrane (IPVH00010, Millipore, Bedford, MA, U.S.A.). The resolved proteins were blocked and probed overnight with the primary antibodies against ROBO3 (1 : 1000; 20220-1-AP, Proteintech, Rosemont, IL, U.S.A.), β-catenin (1 : 1000; A11932, ABclonal), glycogen synthase kinase 3β (GSK3β) (1 : 1000; A2018, ABclonal), p-GSK3β(Ser9) (1 : 500; AP0039, ABclonal), c-myc (1 : 400; AF6054, Affinity), histone H3 (1 : 5000; GTX122148, GeneTex, Irvine, CA, U.S.A.), NELL2 (1 : 1000; BF8498, Affinity), MMP2 (1 : 1000; AF5330, Affinity), MMP9 (1 : 1000; AF5228, Affinity), and β-actin (1 : 1000; sc-47778, Santa Cruz, Dallas, TX, U.S.A.). After washing with Tris-buffered saline with Tween-20, the membranes were incubated with horseradish peroxidase-conjugated anti-rabbit or anti-mouse secondary antibody (1 : 3000; SE134 or SE131, Solarbio). Finally, the bands were visualized using ECL reagent (PE0010, Solarbio).
Cell Proliferation AssayThe cell viability was tested using the Cell Counting Kit-8 (CCK-8; BS350A, Biosharp). Briefly, infected ESCs were inoculated into 96-well plates (5 × 103 cells/well) for the indicated times. CCK-8 solution (10 μL) was added to each well, and incubation continued for 2 h. The absorbance was determined at 450 nm by using a microplate reader (800TS, BioTek, Winooski, VT, U.S.A.).
Wound Scratch AssayThe migratory activity of ESCs was measured using a wound scratch assay. ESCs were cultured in the DMEM/F12 medium supplemented with mitomycin C (M0503, Sigma, St. Louis, MO, U.S.A.) for 1 h. Wounds were created by scraping the cell monolayer with a pipette tip across the center of the cells. Images were captured at 0 and 24 h after the wound was created using a microscope (IX53, Olympus, Tokyo, Japan).
Transwell AssayThe upper transwell chamber (14341, Labselect, China) was pre-coated with Matrigel (356234, Corning, Corning, NY, U.S.A.), followed by cell seeding. To explore the invasion of ESCs, 800 μL of culture medium containing 10% FBS was added to the lower chamber, and 200 μL of ESC suspension was added to the upper chamber. After 24 h of culture, the cells in the lower chamber were fixed with 4% paraformaldehyde (C104188, Aladdin) and stained with 0.5% crystal violet (0528, Amresco, Solon, OH, U.S.A.).
To investigate the effect of ESCs on THP-1 chemotaxis, THP-1 cells were inoculated into the upper compartment (without Matrigel), and infected ESCs were inoculated into the lower compartment. After co-culturing for 48 h, cells were fixed and stained using the same method as above.
The number of invaded cells was quantified in 5 randomly selected areas using a microscope (IX53, OLYMPUS).
Enzyme-Linked Immunosorbent Assay (ELISA)The levels of CCL2 and CCL5 were quantified using commercial ELISA kits (EK187 and EK1129, Lianke Bio, Zhejiang, China) following the manufacturer's protocols. Absorbance was measured at 450/570 nm (reference) using an ELX-800 microplate reader.
Flow CytometryTHP-1-differentiated macrophages were precipitated by centrifugation and washed in PBS. The cells were resuspended in FIX & PERM Medium A (GAS001S5, Thermo Fisher Scientific, Shanghai, China) (100 μL), vortexed thoroughly, and incubated at room temperature for 15 min. Subsequently, FIX & PERM Medium B (GAS002S100, Thermo Fisher Scientific, Shanghai, China) (100 μL) and anti-CD206 antibody (5 μL) were added, and the mixture was vortexed and incubated at 4°C for 30 min. Data were acquired using a NovoCyte flow cytometer (Agilent (Santa Clara, CA, U.S.A.)).
StatisticsGraphs and statistical analyses (either t-test or ANOVA) were prepared with GraphPad Prism 9.0 (GraphPad Software Inc., San Diego, CA, U.S.A.). Data normality and homogeneity of variances were tested using the Shapiro–Wilk test and F-test. All data were expressed as mean ± standard deviation (S.D.). Significant difference is indicated by a p-value <0.05.
Two independent datasets were analyzed to identify DEGs between healthy controls and endometriosis patients, with their relative expression patterns visualized in a clustered heatmap (Fig. 1A). Comparative analysis revealed 286 significantly upregulated genes compared to 163 downregulated transcripts in endometriosis samples (Fig. 1B). To elucidate the biological significance of these DEGs, we conducted GO and KEGG analyses. KEGG analysis revealed that the DEGs were mainly enriched in cell adhesion molecules, mitogen-activated protein kinase (MAPK), phosphatidylinositol 3-kinase-Protein Kinase B (PI3K-Akt), Wnt, and cancer-related pathways (Fig. 1C). GO analysis revealed these DEGs mainly focused on the functions such as cell proliferation, cell migration, immune response, chemotaxis, and the Wnt pathway (Fig. 1D). These findings suggest potential mechanistic involvement of immune dysregulation, aberrant cellular motility, chemotactic processes, and the Wnt pathway in endometriosis progression. The ROBO gene family plays a pivotal role in reproductive disorders, prompting us to investigate its expression in endometriosis. Comparative analysis of GSE168902 and GSE25628 via heatmaps identified ROBO3 as the most significantly upregulated ROBO family member, suggesting its potential importance in disease pathogenesis (Fig. 1E). Based on these bioinformatic predictions, we subsequently conducted functional validation experiments to investigate ROBO3’s roles in endometriosis progression.
(A) Heatmap visualization of DEGs in 2 datasets (GSE25628 and GSE168902) from the GEO database. (B) Comparative analysis of up- and downregulated DEGs with Venn diagrams. (C) KEGG pathway enrichment analysis of common upregulated DEGs, with green bars denoting signaling pathways and red bars indicating associated cancer types. (D) GO enrichment analysis of common upregulated DEGs. (E) ROBO family expression in GSE168902 and GSE25628. ROBO3 shows consistent upregulation. Comparative analysis of ROBO3 expression levels in clinical specimens from 2 datasets: GSE168902: normal (n = 3 samples) vs. endometriosis (n = 3 samples); GSE25628: normal (n = 6 samples) vs. endometriosis (n = 7 samples). Data are presented as mean ± S.D. Comparisons were performed by unpaired t-test. DEG: differentially expressed gene.
Primary ESCs were isolated from clinical ectopic endometrial specimens and validated via immunofluorescence, demonstrating cytokeratin negativity and vimentin positivity, confirming stromal cell identity (Fig. 2A). To investigate the functional role of ROBO3, we transduced ESCs with adenoviral vectors for either ROBO3 overexpression or knockdown. Successful modulation was confirmed at both the protein (Western blot) and mRNA (RT-qPCR) levels (Figs. 2B and 2C). Functional assays revealed that ROBO3 knockdown significantly suppressed ESC viability, whereas ROBO3 overexpression promoted cell proliferation (Fig. 2D).
(A) Immunofluorescence staining of epithelial–mesenchymal transition markers vimentin and cytokeratin in ESCs. (B, C) Validation of ROBO3 knockdown or overexpression in ESCs with Western blot (B) and RT-qPCR (C). (D) The proliferation of ESCs was measured by CCK-8 assay. Data are presented as mean ± S.D. (n = 3 biological replicates per group). Comparisons were performed by one-way ANOVA (C) and two-way ANOVA (D). ESC: endometrial stromal cell.
To assess the functional role of ROBO3 in cell motility, we performed scratch-wound healing assays. ROBO3 knockdown significantly impaired wound closure compared to control (shNC), while ROBO3 overexpression accelerated scratch repair (Figs. 3A and 3C). Transwell invasion assays demonstrated that ROBO3 silencing markedly reduced ESC invasiveness, whereas ROBO3 overexpression enhanced invasive capacity relative to vector controls (Figs. 3B and 3D). Given MMPs’ established role in tissue remodeling, we examined whether ROBO3 regulates MMP2/MMP9 expression. Western blot analysis revealed that ROBO3 silencing significantly reduced both MMP2 and MMP9 protein levels, while overexpression upregulated both proteases, consistent with the observed invasion phenotypes (Fig. 3E). Chemokine profiling revealed that ROBO3 knockdown attenuated secretion of CCL2 and CCL5, while ROBO3 overexpression upregulated these inflammatory mediators (Fig. 3F). This ROBO3–chemokine axis suggests a potential mechanism for immune cell recruitment in endometriotic lesions. Collectively, these findings establish ROBO3 as a critical regulator of ESC proliferation, migration, and invasion, while simultaneously modulating the inflammatory microenvironment through chemokine regulation.
(A) Representative pictures of scratch-wound migration assay of ESCs. (B) Representative pictures of the transwell assay. (C) Quantitative analysis of wound closure rate. (D) Quantification of invaded cells. (E) Western blot analysis of MMP2 and MMP9. (F) ELISA measurement of chemokine secretion: CCL2 and CCL5 levels. Data are presented as mean ± S.D. (n = 3 biological replicates per group). Comparisons were performed by one-way ANOVA (for 3 groups) and unpaired t-test (for 2 groups).
To assess the effect of ROBO3 on the Wnt/β-catenin signaling pathway, we examined several key proteins associated with this pathway, including c-myc, GSK3β, and β-catenin. As displayed in Fig. 4A, ROBO3 knockdown downregulated c-myc expression, suppressed GSK3β phosphorylation, and diminished β-catenin levels in both cytoplasmic and nuclear compartments. Conversely, ROBO3 overexpression in ESCs elicited the opposite effects, upregulating these molecular markers. Immunofluorescence assays corroborated these findings, demonstrating that ROBO3 knockdown effectively suppressed β-catenin nuclear translocation in ESCs, whereas ROBO3 overexpression enhanced this process (Fig. 4B). Collectively, these data demonstrate that ROBO3 activates the canonical Wnt/β-catenin pathway through: (1) facilitating phosphorylation of GSK3β, leading to β-catenin stabilization, and (2) promoting β-catenin nuclear translocation to activate downstream target genes including c-myc.
(A) Western blot analysis of Wnt pathway components: GSK3β (phosphorylated/inactive form at Ser9: p-GSK3β), downstream target c-myc, and cytoplasmic and nuclear β-catenin (n = 3 biological replicates). (B) Nuclear translocation of β-catenin assessed by immunofluorescence (n = 3 biological replicates). Arrows indicate nuclear accumulation.
To establish the mechanistic dependence of ROBO3’s effects on Wnt/β-catenin signaling, we employed the specific β-catenin inhibitor MSAB in functional rescue experiments. CCK-8 assay findings demonstrated that the increased cell viability of ROBO3-overexpressing ESCs was attenuated following treatment with MSAB (Fig. 5A). In the wound scratch assay, MSAB treatment ameliorated the enhanced migratory capacity of ESCs induced by ROBO3 overexpression (Fig. 5B). Additionally, MSAB treatment abrogated the increased invasive capacity of ESCs with forced ROBO3 expression (Fig. 5C). Notably, the ROBO3-induced upregulation of CCL2 and CCL5 expression was also suppressed by MSAB treatment (Fig. 5D). While these results establish Wnt/β-catenin signaling as a major downstream mechanism mediating ROBO3’s effects, the residual MSAB-resistant activities suggest ROBO3 may additionally engage Wnt-independent pathways in ESCs.
(A) CCK-8 assay assessing ESC viability. (B) Representative pictures and quantitation of scratch-wound migration assay of ESCs. (C) Representative pictures and quantitation of transwell assay. (D) ELISA measurement of chemokine secretion: CCL2 and CCL5 levels. Data are presented as mean ± S.D. (n = 3 biological replicates per group). Comparisons were performed by one-way ANOVA.
Given the regulatory effect of ROBO3 on chemokine secretion, we further explored its role in modulating THP-1 macrophage chemotaxis and polarization. Transwell migration assays demonstrated that ROBO3 knockdown in ESCs suppressed THP-1 cell recruitment, whereas ROBO3 overexpression enhanced macrophage chemotaxis (Fig. 6A). Immunofluorescence staining confirmed that ROBO3 overexpression in ESCs enhanced Arg-1 expression in co-cultured THP-1 macrophages, while ROBO3 knockdown attenuated Arg-1 induction (Fig. 6B). Flow cytometric analysis of CD206, a well-established M2 macrophage marker, showed that ROBO3-deficient ESCs decreased the percentage of CD206+ THP-1 cells, whereas ROBO3-overexpressing ESCs increased this population (Fig. 6C). To comprehensively characterize macrophage polarization states, we performed RT-qPCR analysis of key M1 and M2 markers. The results demonstrated a distinct polarization pattern: ROBO3 knockdown significantly upregulated prototypical M1 markers (CD86 and iNOS), while simultaneously downregulating characteristic M2 markers (CD206, Arg-1, CD163, and IL-10). In contrast, ROBO3 overexpression drove macrophages toward an M2-polarized state, with opposite effects on these marker sets (Fig. 6D). Collectively, these findings demonstrate that ROBO3 expression in ESCs critically regulates macrophage recruitment and M2 polarization.
(A) THP-1 monocytes were differentiated into M0 macrophages with PMA (100 ng/mL) for 48 h. The transwell assay was used to detect the effect of co-culture with ROBO3 knockdown or overexpression ESCs on macrophage invasive capacity. ESCs with ROBO3 knockdown or overexpression were cultured in fresh medium for 48 h. Subsequently, the culture medium was collected, diluted 1 : 1 with fresh medium, and used to incubate PMA-differentiated THP-1 cells (M0 macrophages) for an additional 24 h. (B) Immunofluorescence staining of the M2 polarization marker (Arg-1) in THP-1-differentiated macrophages. (C) Percentage of CD206+ M2 macrophages was detected by flow cytometry. (D) RT-qPCR analysis of M1 (CD86, iNOS) and M2 (CD206, Arg-1, CD163, IL-10) polarization markers in macrophages co-cultured with ROBO3-modified ESCs. Data are presented as mean ± S.D. (n = 3 biological replicates per group). Comparisons were performed by one-way ANOVA. PMA: phorbol 12-myristate 13-acetate.
Previous studies have identified neural epidermal growth factor-like 1 (NELL2) as a ligand for ROBO3.22,23) To determine whether ROBO3 functions in a ligand-dependent manner in endometriosis, we first confirmed NELL2 expression in ESCs, THP-1 macrophages, and SW480 cells (positive control) by Western blot (Supplementary Fig. S1A). We then performed NELL2 knockdown in primary ESCs using siRNA, which significantly reduced NELL2 expression (verified by RT-qPCR; Supplementary Fig. S1B). Functional assays showed that NELL2 depletion blocked the enhanced migration (scratch assay) and invasion (Transwell assay) induced by ROBO3 overexpression (Supplementary Figs. S1C–S1F). Importantly, NELL2 knockdown also reversed the effects of ROBO3 on macrophage polarization: it restored CD86 (M1 marker) expression and reduced CD206 (M2 marker) levels in THP-1 cells (Supplementary Fig. S1G). These results demonstrate that NELL2 serves as a functional ligand for ROBO3 in our system. The partial inhibition of ROBO3-mediated phenotypes—including enhanced migration, invasion, and macrophage polarization—following NELL2 silencing provides strong evidence that ROBO3 signaling may be partially dependent on NELL2 ligand binding in endometriosis.
Endometriosis has a significant impact on patients’ QOL and has recently become a research hotspot.24) The complex pathogenesis of endometriosis makes it difficult to fully understand the specific mechanisms behind the disease.25,26) In patients with endometriosis, the ectopic endometrial tissues exhibit increased cell growth and enhanced migration and invasion abilities.27) In our study, we demonstrated increased expression of ROBO3 in patients with endometriosis (Fig. 1). Notably, ROBO3 knockdown inhibited ESC proliferation, migration, and invasion, potentially through its regulatory effects on the Wnt/β-catenin pathway (Figs. 2–5). Furthermore, ROBO3 silencing impaired ESC-mediated macrophage chemotaxis and M2 polarization (Fig. 6).
Currently, extensive studies focus on exploring the cell factors and their signaling pathways related to endometriosis.28–30) The canonical Wnt/β-catenin signaling pathway has been demonstrated to be critically involved in modulating multiple pathological processes associated with endometriosis pathogenesis.4,31) β-Catenin is a transcription factor and a vital component of the Wnt pathway. Under basal conditions without Wnt ligand stimulation, GSK3β-mediated phosphorylation targets β-catenin for proteasomal degradation, maintaining its cytoplasmic concentration at minimal levels.32) Upon binding to the receptor, GSK3β is inactivated. β-Catenin is dissociated from the degradation complex and accumulates in the cytoplasm.33) Following nuclear translocation, β-catenin forms a complex with T-cell factor/Lymphoid enhancer factor (TCF/LEF) family transcription factors, thereby upregulating expression of target genes such as c-myc that regulate critical cellular processes.34,35) In our study, Western blot revealed that ROBO3 overexpression in ESCs elevated p-GSK3β and c-myc protein levels (Fig. 4A). Furthermore, nuclear β-catenin accumulation was increased in oeROBO3-treated ESCs, indicative of Wnt/β-catenin pathway activation (Fig. 4B). Activation of this pathway is known to upregulate downstream targets such as MMPs, which are critically involved in endometriosis progression by facilitating tissue remodeling and invasion.36) Our finding that ROBO3 regulates MMP2/9 provides a direct mechanistic link to the elevated MMP activity observed in endometriotic lesions37) (Fig. 3E). This suggests ROBO3 inhibition could normalize the proteolytic imbalance characteristic of endometriosis progression. Moreover, abnormal stimulation of the Wnt/β-catenin pathway mediates proliferation, migration, and invasion of ESCs.4,36) Collectively, the Wnt/β-catenin pathway may be the key mechanism of ROBO3 promoting endometriosis.
Macrophages contribute significantly to endometriosis pathogenesis and symptom exacerbation.38) A recent study has demonstrated that the pharmacological β-catenin inhibitor downregulates the secretion of key chemokine CCL2, highlighting a critical β-catenin–CCL2 feedback loop in macrophage regulation.39) Additionally, β-catenin modulates CCL5 expression via the β-catenin/IKZF1/CCL5 axis.40) Our study demonstrated that ROBO3 overexpression in ESCs significantly upregulated the expression of CCL2 and CCL5—crucial chemokines known to mediate macrophage recruitment to endometriotic lesions (Fig. 3F). Importantly, these chemokines have been shown to promote macrophage differentiation into the M2 phenotype, thereby contributing to the formation of an immunosuppressive microenvironment within endometriotic lesions.38) This immunosuppressive environment may further facilitate disease progression.3,41) Therefore, activated polarization of M2 macrophages may be promoted by ROBO3 through upregulating the expression of CCL2 and CCL5 in ESCs.
The functional rescue experiments utilizing MSAB provided crucial mechanistic insights: (1) The substantial reversal of ROBO3-induced phenotypic changes across all examined parameters—including proliferation, migration, invasion, and inflammatory marker expression—strongly implicates Wnt/β-catenin signaling as the predominant downstream effector pathway (Fig. 5). (2) ROBO3 activity was maintained to a lesser extent despite β-catenin inhibition, implying the concurrent engagement of complementary signaling cascades.
Our findings establish ROBO3 as a critical regulator of macrophage polarization within the endometriotic microenvironment. These functional alterations are clearly demonstrated by our experimental results showing that ROBO3 overexpression significantly suppresses the expression of M1 polarization markers (CD86 and iNOS) while simultaneously upregulating characteristic M2 markers (CD206, Arg-1, CD163, and IL-10) (Fig. 6). This reciprocal regulation of macrophage polarization markers provides compelling evidence that ROBO3 serves as a molecular switch driving macrophages toward an M2-polarized phenotype in the endometriotic microenvironment. Building upon prior structural studies of ROBO3–NELL2 interactions,22,23) we systematically characterized this ligand–receptor pair in endometriosis pathogenesis. NELL2 expression was confirmed across relevant cell types, and functional studies established its necessity for ROBO3-mediated effects. NELL2 depletion not only abolished ROBO3-driven ESC invasion but also partially reversed M2 polarization (Supplementary Fig. S1). However, the incomplete rescue suggests additional ligand–receptor interactions may contribute to ROBO3’s immunomodulatory functions. The relative contributions of NELL2-dependent vs. NELL2-independent mechanisms in shaping the endometriotic immune landscape warrant further investigation.
Consistent with previous reports, the results of GO enrichment analysis in this study indicated that P13K-Akt, nuclear factor kappa B (NF-κB), and MAPK pathways may be involved in endometriosis pathogenesis.42,43) The function of PI3K-Akt in ESCs is to inhibit GSK3β activity, which keeps c-myc pluripotency.44) The NF-κB pathway has a central role in inflammatory response and immune regulation.45) GSK-3β, affected by ROBO3, is critically important for regulating NF-κB signaling activity.46) Additionally, when ROBO3 binds to its ligand NELL2, it may influence some specific members of the MAPK family, such as extracellular signal-regulated kinase (ERK), c-Jun N-terminal kinase/stress-activated protein kinase (JNK/SAPK), and p38 kinase, to trigger the MAPK pathway.23,47) MAPK pathway has been recognized to participate in complex cellular programs like proliferation, differentiation, and transformation.48,49) ROBO3 may play a role by affecting the above pathways, which needs to be further verified by subsequent experiments. The current study focuses on establishing ROBO3’s causal roles through in vitro manipulation. While public dataset analyses show suggestive correlations, definitive clinical validation requires prospective collection of tissues with detailed phenotyping—a direction we are actively pursuing.
Several limitations should be noted. First, while MSAB is a specific β-catenin inhibitor, we cannot exclude off-target effects contributing to the observed phenotypes. Second, while our data delineate the relative contributions of Wnt signaling, the precise molecular effectors downstream of endogenous ROBO3 remain to be identified. Future studies will explore whether ROBO3 directly modulates cytoskeletal regulators (e.g., FAK) or inflammatory transcription factors (e.g., NF-κB) in Wnt-independent contexts.
Overall, our findings establish ROBO3 as a key regulator of endometriosis progression through a unique ESC-macrophage crosstalk mechanism. In endometriotic lesions, ROBO3 upregulation: (1) directly enhances ESC malignant behaviors, and (2) stimulates M2 macrophage polarization. These effects are mediated at least in part through activation of the Wnt/β-catenin signaling pathway, which not only advances our understanding of endometriosis pathogenesis but also suggests novel therapeutic strategies that simultaneously target multiple ROBO3 effector pathways.
The authors declare that no funds, grants, or other support were received during the preparation of this manuscript.
The authors declare no conflict of interest.
This article contains supplementary materials.
