Patient-derived decidual organoids reveal rhG-CSF-associated protective responses involving JAK2/STAT3 signaling in recurrent spontaneous abortion

Abstract

Background

Recurrent spontaneous abortion (RSA) is a complex reproductive disorder frequently lacks effective therapeutic interventions. Recombinant human granulocyte colony-stimulating factor (rhG-CSF) has shown promise in improving pregnancy outcomes in RSA, but its underlying mechanisms remain unclear. Organoid technology offers new opportunities for disease modeling and mechanistic research in reproductive medicine.

Methods

Decidual tissue samples from RSA patients were utilized to establish decidual organoids using three-dimensional (3D) culture techniques. These organoids were characterized via morphological observation, hematoxylin and eosin (H&E) and Periodic Acid-Schiff (PAS) staining, immunofluorescence, and immunohistochemistry to validate key marker expression and hormone responsiveness. The effects of rhG-CSF on organoid viability were evaluated using luminescence assays. Transcriptomic sequencing and bioinformatics analyses were performed to identify differentially expressed genes and pathways following rhG-CSF stimulation. Protein expression and signaling pathway activation were further assessed via Western blot analysis.

Results

Decidual organoids were successfully established and recapitulated the structural and functional features of primary tissue, including hormone responsiveness and marker expression. rhG-CSF significantly promoted organoid proliferation, enhanced anti-apoptotic and angiogenic capacity, and induced inflammatory responses. Transcriptomic and pathway analyses indicated an enrichment and activation of the JAK2/STAT3 signaling pathway following rhG-CSF treatment. Western blot indicated increased phosphorylation of JAK2 and STAT3, upregulation of PCNA, VEGFA, IL-1β, IL-6, and Bcl-2, and downregulation of Bax in response to rhG-CSF.

Conclusion

A patient-derived decidual organoid model was established and validated to recapitulate in vivo tissue characteristics. In this in vitro RSA model, rhG-CSF appeared to exert protective effects by promoting proliferation, inhibiting apoptosis, and enhancing angiogenesis—responses that are potentially associated with the activation of JAK2/STAT3 signaling. These findings provide new insights and experimental evidence that may inform future RSA treatment strategies.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1186/s13287-026-05187-y.

Introduction

Recurrent spontaneous abortion is clinically defined as the occurrence of two or more consecutive unexplained pregnancy losses, confirmed by ultrasound or histopathology, with the same partner [1]. Women with RSA are at a significantly increased risk of subsequent miscarriage and adverse pregnancy outcomes. For instance, the risk of another miscarriage can be as high as 70%–80% in women who have experienced three consecutive pregnancy losses [2, 3]. Currently, RSA lacks distinct clinical manifestations and effective therapeutic strategies, presenting considerable challenges to patients’ physical and mental health, as well as to clinical management [4, 5].

In recent years, rhG-CSF has shown potential in improving conception and live birth rates in patients with RSA [6–9]. G-CSF stimulates the proliferation and differentiation of neutrophils and plays an integral role in endometrial development, trophoblast function, and placental metabolism regulation [10–12]. Studies have demonstrated that G-CSF may enhance decidualization of endometrial stromal cells and increase the number of activated cells through paracrine mechanisms, thereby facilitating embryo implantation [13, 14]. Other research suggests that rhG-CSF may regulate molecular pathways in the endometrium related to embryo implantation, including vascular remodeling, local immune modulation, and cell adhesion [15]. Therefore, further evaluation of the therapeutic efficacy of rhG-CSF for RSA is of paramount clinical importance.

The uterine decidua is a specialized tissue derived from the endometrium during pregnancy, primarily composed of terminally differentiated stromal cells, blood vessels, and various immune cells [16, 17]. The decidua plays a pivotal role in embryo implantation and pregnancy maintenance [18]. Decidual cells regulate maternal-fetal signaling through the secretion of cytokines and chemokines, thereby promoting embryo implantation and early development [19]. In addition, the decidua acts as a physical barrier and modulates immune responses to protect the embryo from maternal immune attack, while also providing nutritional support for the developing embryo prior to placental formation [20]. Aberrations in decidual function are closely associated with RSA [21]. These findings highlight the significance of decidual research for advancing our understanding of RSA.

Organoids are three-dimensional cellular structures derived from stem cells that can self-organize and differentiate in vitro, recapitulating the morphological and functional features of their respective organs [22]. Organoid technology holds great promise in reproductive medicine. Compared with traditional two-dimensional cell culture systems, organoids better mimic the complex polarity and functional characteristics of in vivo tissues [23, 24], and are now widely used in drug screening, disease modeling, and personalized medicine [25, 26]. In 2017, researchers successfully established organoid models from both murine and human endometrial tissues. These endometrial organoids can be long-term expanded and faithfully recapitulate the molecular and histological phenotypes of the endometrial epithelium, maintain genetic stability, and, importantly, respond to steroid hormones to undergo decidualization-like processes [27, 28]. These features provide advanced tools for the study of endometrium-related diseases.

Although significant progress has been made in the development and application of endometrial organoids worldwide, studies investigating the role of organoids in RSA remain in the early stages. Previous research has indicated the potential value of rhG-CSF in improving pregnancy outcomes in patients with RSA; however, its underlying mechanisms remain largely unclear, thereby limiting its widespread clinical application. To address this gap, our study established decidual organoids derived from RSA patients and utilized this model to investigate the cellular responses and potential signaling pathways associated with rhG-CSF treatment. Our findings aim to provide further scientific evidence to optimize the diagnosis and treatment strategies for RSA.

Materials and methods

Source of samples

We collected decidual tissue samples from patients diagnosed with RSA at the First Affiliated Hospital of Guangxi Medical University [29]. The inclusion criteria were as follows: [1] a history of two or more consecutive spontaneous abortions with the same partner; [2] gestational age between 6 and 10 weeks at the time of sample collection, confirmed by ultrasound; and [3] patients aged between 18 and 35 years.

To ensure the homogeneity of the samples, we applied strict exclusion criteria: [1] embryonic chromosomal abnormalities confirmed by genetic testing; [2] uterine anatomical abnormalities (e.g., septate uterus, Asherman’s syndrome); [3] endocrine disorders (e.g., diabetes, thyroid dysfunction, hyperprolactinemia); [4] autoimmune diseases (e.g., antiphospholipid syndrome); and [5] acute or chronic reproductive tract infections.

We concurrently recruited a control group consisting of 20 women undergoing elective surgical termination of normal early pregnancy for non-medical reasons. The control group was matched to the RSA group in terms of age (± 2 years) and gestational age (± 1 week). Inclusion criteria for the control group were: [1] clinically confirmed intrauterine pregnancy with a viable embryo observed by ultrasound (6–10 weeks); and [2] a history of at least one live birth with no history of spontaneous abortion, preterm birth, or infertility. The exclusion criteria regarding systemic diseases and uterine abnormalities were identical to those of the RSA group.

None of the participants had received hormonal treatments such as mifepristone, misoprostol, or progesterone prior to the surgical evacuation. This study was approved by the Ethics Committee of the First Affiliated Hospital of Guangxi Medical University (No. 2023-K057-01). All participants signed written informed consent prior to surgery.

Specimen collection and processing

After surgical termination of pregnancy, the decidual tissue was repeatedly washed with pre-cooled PBS. Coagulated blood was removed using sterile forceps and ophthalmic scissors, and the tissue was washed until no visible blood clots remained. Each decidual tissue sample was divided into two parts: the first one was placed in a centrifuge tube containing pre-cooled transport medium, which consisted of 5% fetal bovine serum (Gibco, USA), 2% penicillin-streptomycin (Solarbio, China), and 93% DMEM/F-12 (Thermo Fisher Scientific, USA), and rapidly transferred to the laboratory for primary organoid culture. The other was fixed in 4% paraformaldehyde (PFA) in a 15 mL centrifuge tube for subsequent paraffin embedding, staining, and immunofluorescence analysis.

Organoid culture

Organoid culture was performed following the protocol described by Boretto et al. [27]. Decidual tissue obtained in the operating room was placed in a sterile specimen box and transported to the laboratory. After centrifugation at 400 × g for 5 min at 4 °C and removal of supernatant, the tissue was minced into 1–3 mm³ fragments. The fragments were then digested in an enzyme solution containing Dispase II (Sigma, USA), Collagenase IV (Sigma, USA), and Y-27,632 (Sigma, USA) at 37 °C for 1–2 h. Digestion was terminated, and the cell suspension was sequentially filtered through 100 μm and then 40 μm strainers. The cells retained on the 40 μm strainer were collected by inverting the strainer. The cell pellet was resuspended in Advanced DMEM/F-12 (Thermo Fisher Scientific, USA) medium, and the cell concentration was determined with a hemocytometer. Approximately 2 × 10⁴ cells were embedded in 25 µL of Matrigel (Corning, USA). Matrigel and Advanced DMEM/F-12 cell suspension were mixed at a 7:3 ratio. A 25 µL drop of the Matrigel-cell suspension was placed at the center of each well of a 48-well plate and incubated at 37 °C for 20 min to allow the gel to solidify. Then, 300 µL of pre-warmed organoid culture medium was added to each well. Organoid growth was monitored and recorded daily; the medium was changed every 2–3 days. The detailed composition of the organoid culture medium is provided in Supplementary Table S1. To assess long-term expansion potential and stability, a subset of organoids was continuously cultured up to passage 10. For all other functional experiments, organoids at passages 3–5 were used to ensure phenotypic consistency.

Organoid passage

After aspirating the culture medium from the 48-well plate, a gentle dissociation solution (TrypLE (Invitrogen, USA) supplemented with Y-27632) was carefully added to each well, ensuring the Matrigel structure remained intact. The plate was incubated in a CO₂ incubator for 10–15 min to dissociate the organoids. The wells were gently scratched to facilitate release of the organoids from the Matrigel. The suspension was transferred to a 15 mL centrifuge tube, and digestion was terminated by adding DMEM/F-12 medium. After centrifugation at 300 × g for 5 min at room temperature, the organoid pellet was resuspended in 1–2 mL Advanced DMEM/F-12 medium to prepare a new organoid-Matrigel suspension. A 25 µL drop of this suspension was added to each well of a 48-well plate, followed by addition of organoid culture medium. Subsequent experiments were performed as required.

Organoid cryopreservation

Organoids at passage 3 or 4 were selected for cryopreservation. After removing residual medium, 400 µL of pre-cooled Cell Recovery Solution (Corning, USA) was added to each well and gently pipetted to disperse the Matrigel and collect released organoids. The organoid suspension was transferred to a 15 mL centrifuge tube and centrifuged at 300 × g for 5 min at 4 °C. The organoid pellet was resuspended in cryopreservation solution at a concentration of approximately 2–5 × 10⁴ organoids/mL. The cryotubes were first stored at − 80 °C for over 24 h, and then transferred to liquid nitrogen for long-term storage to maintain organoid viability and experimental stability.

Organoid hormone treatment

Experimental Grouping and Treatment: Organoids at passages 3–5 were used for hormone treatment assays. On day 4 after organoid passage, organoids were randomly divided into three groups: Blank Control Group: Standard organoid culture medium only. E2 Group: Organoid culture medium supplemented with 10 nM estradiol (E2). E2 + progesterone (P4) + cyclic adenosine monophosphate (cAMP) Group: Organoid culture medium supplemented with 10 nM E2, 1 µM P4, and 1 µM cAMP. Organoids in the E2 and E2 + P4 + cAMP groups were pretreated by culturing in medium containing 10 nM E2 for 48 h to allow adaptation to the hormonal microenvironment. After pretreatment, organoids were subjected to continuous hormone intervention according to their respective group protocols. The intervention lasted for 4 days, with the medium changed every 2 days to ensure stable culture conditions and effective hormone concentrations.

Organoid collection, fixation, and embedding

Organoids were collected by adding Cell Recovery Solution to disperse the Matrigel, and the organoid pellet was transferred into tubes. Organoids were fixed in 4% paraformaldehyde solution. Pre-warmed agarose solution was then added to the tubes and allowed to solidify at room temperature for 60 min to ensure complete agarose embedding. Samples were dehydrated in a graded ethanol series, cleared with xylene, and finally embedded in paraffin. Paraffin blocks were sectioned into 3 μm slices for subsequent experiments.

Hematoxylin and Eosin (HE) staining and periodic acid-schiff (PAS) staining

Sections were deparaffinized and rehydrated using xylene and a graded series of ethanol. Nuclei were stained with hematoxylin solution for 10 min, differentiated for 30 s in a hydrochloric acid differentiation solution, and the cytoplasm was stained with eosin for 2 min. Sections were then dehydrated and cleared with graded ethanol and xylene. A drop of periodic acid solution was added to the tissue section and incubated in a light-protected humid chamber for 10 min. Schiff’s reagent was then added, and the slides were incubated at 37 °C in the dark for 30 min. Nuclei were stained with hematoxylin for 30 s.

Immunofluorescence

After heat-induced antigen retrieval, endogenous peroxidase activity was quenched by incubation with 3% H₂O₂ at room temperature for 10 min. The sections were blocked with goat serum blocking solution (Solarbio, China) at 37 °C for 30 min. Subsequently, 70 µL of specific primary antibody was added and incubated overnight at 4 °C (Supplementary Table S2). After primary antibody incubation, 200 µL of secondary antibody was applied and incubated at room temperature for 1 h. Nuclei were counterstained with DAPI for 3 min at room temperature, and an antifade mounting medium was applied. Protein expression was visualized and analyzed using an inverted fluorescence microscope with appropriate excitation wavelengths.

Immunohistochemistry

The procedure was identical to immunofluorescence up to the primary antibody incubation step (Supplementary Table S3). After that, 70 µL of HRP-conjugated secondary antibody prepared with 1% BSA was applied. DAB working solution was used for color development, and sections were counterstained with hematoxylin. Images were captured from five randomly selected fields per section at 400× magnification using a bright-field microscope.

CTL luminescence assay for assessing the effect of rhG-CSF on organoid viability

Organoids at passages 3–5 were dissociated into single cells using TrypLE Express and counted. Approximately 5,000 cells in 20 µL of Matrigel were seeded per well in a black clear-bottom 96-well plate. Each well received 100 µL of organoid culture medium. After 24 h, the medium was replaced with either organoid culture medium containing the indicated concentrations of rhG-CSF (MedChemExpress, USA) or blank control medium. The rhG-CSF concentration gradient was set as follows: 0, 10, 100, 200, and 500 ng/mL. Media were changed every 48 h. After 96 h of culture, 100 µL of CellTiter-Lumi™ 3D luminescence assay reagent (Beyotime, China) was added per well and incubated at room temperature for 25 min. Luminescence intensity was measured using a microplate reader (BioTek, USA).

Organoid sequencing and bioinformatics analysis

Following stimulation with the optimal concentration of rhG-CSF, organoids (passage 4) from three independent patient-derived lines (n = 3 for the control group and n = 3 for the rhG-CSF treated group) were washed, collected, and submitted to BGI (Shenzhen, China) for transcriptome sequencing to profile mRNA expression. To minimize technical variation, all samples were processed in a single batch. Raw sequencing data were quality controlled and filtered to obtain count expression profiles. Differential expression analysis was performed using the “DESeq2” package in R, with screening criteria set as |Log2FC| ≥ 1 and padj < 0.05. To further explore gene functions associated with phenotypic changes, “clusterProfiler,” “limma,” and “org.Hs.eg.db” packages were used for KEGG and gene set enrichment analysis (GSEA) [30, 31].

Western blotting

Organoids at passages 3–4 were grouped and treated as follows: the control group had the medium replaced with standard culture medium after 24 h; the G-CSF group received 100 ng/mL rhG-CSF; the AG490 group received 50 µM AG490 (MedChemExpress, USA); and the G-CSF + AG490 group was treated with both 100 ng/mL rhG-CSF and 50 µM AG490. After 5–7 days of continuous treatment, organoids were collected and lysed with RIPA buffer (Solarbio, China). Protein samples were loaded onto 10% SDS-PAGE gels, separated by electrophoresis, and transferred to PVDF membranes (Millipore, USA). Membranes were blocked with 5% BSA (Solarbio, China) for 1 h, then incubated overnight at 4 °C with primary antibodies (Supplementary Table S4). The following day, membranes were incubated with HRP-conjugated goat anti-rabbit IgG at room temperature for 1 h. Protein bands were visualized by chemiluminescence scanning (Thermo Fisher, USA). Densitometric quantification of the protein bands was performed using ImageJ software (NIH, Bethesda, MD, USA). The relative protein expression levels were normalized to the internal control, Tubulin. To evaluate apoptosis and the activation of the JAK2/STAT3 signaling pathway, the ratios of Bax/Bcl2, p-JAK2/JAK2, and p-STAT3/STAT3 were calculated based on the normalized densitometry values.

Statistical analysis

Image data from immunohistochemistry and Western blot protein bands were analyzed using ImageJ software (NIH, Bethesda, MD, USA). For immunohistochemistry, the expression levels of ER, PR, and Ki-67 were quantified as Mean Optical Density (Mean OD). All experiments were performed using three independent patient-derived organoid lines to ensure biological replication (n = 3). For each biological replicate, three technical repeats were performed. In statistical analyses, data from technical repeats were averaged to generate a single data point per patient line, strictly avoiding pseudoreplication. Statistical comparisons between groups were performed using GraphPad Prism 9.5.1. Measurement data are expressed as mean ± standard deviation (SD), while non-normally distributed data are presented as median (interquartile range). A P-value < 0.05 was considered statistically significant.

Results

Clinical information of patients

A total of 20 decidual tissue samples were collected from patients with RSA. Decidual organoids were successfully established from 11 samples using 3D organoid culture technology, yielding a success rate of 55%. The clinical characteristics of these 11 patients are summarized as follows: the mean age was 29.45 ± 2.30 years, the mean body mass index (BMI) was 22.89 ± 2.85 kg/m², and the mean gestational age at sampling was 60.00 ± 5.31 days. Obstetric history data, which were non-normally distributed, are presented as median (interquartile range [IQR]). The median number of pregnancies was 3 [2–5], the median number of deliveries was 0 (0–1), and the median number of previous miscarriages was 3 [2–5]. All included patients had confirmed normal embryonic karyotypes. Detailed clinical characteristics are summarized in Supplementary Table S5.

Growth, passage, and recovery of decidual organoids

Under bright-field microscopy, decidual organoids exhibited a three-dimensional, spherical morphology. At varying focal planes, different cellular layers were observable, with cross-sections showing circular or near-circular cystic structures. By day 9 of culture, the organoid diameter increased to approximately 200 μm (Fig. 1A). Decidual organoids demonstrated a remarkable capacity for long-term expansion; even after 10 passages, the organoids maintained robust growth and structural stability (Fig. 1B). Following cryopreservation in liquid nitrogen and subsequent recovery, the organoids retained their morphological integrity. Bright-field imaging revealed no significant morphological differences between pre- and post-cryopreservation organoids across different batches (Supplementary Figure S1).

Fig. 1.

Bright-field microscopy images of decidual organoids. A Representative images of organoid growth at different culture time points (Day 1 to Day 9). B Representative images of organoids at different passage numbers (Passage 1, 2, 4, 7, and 10). Scale bars: 100 μm A and 200 μm B. Magnification: 100×. Images are representative of n = 3 independent experiments

HE and PAS staining

H&E staining revealed that the cellular morphology and structure of decidual organoids closely resembled those of the original tissue. The organoids exhibited cystic structures with clear central lumens, orderly arranged peripheral cells, and round or oval nuclei—features consistent with their tissue of origin (Fig. 2A). As shown in Fig. 2B, a purplish-red signal was observed within the organoid lumen, indicating that the decidual organoids possessed secretory functions analogous to endometrial epithelial cells.

Fig. 2.

Characterization of decidual organoids. A H&E staining showing histological similarity with source tissues. B PAS staining demonstrating secretion functions. Scale bars: 100 μm A and 50 μm B. Original magnification: 100× A and 200× B. Data are representative of n = 3 biological replicates

Expression of key characteristic markers in decidual organoids

Immunofluorescence confirmed the expression of multiple key markers, including the mesenchymal marker vimentin; essential epithelial adhesion molecules E-cadherin and EpCAM; the epithelial-specific mucin Muc1; the critical transcription factor Foxa2; estrogen receptor (ER) and progesterone receptor (PR); as well as G-CSF and its receptor G-CSFR. The expression patterns of these markers in decidual organoids were highly consistent with those observed in the original decidual tissue (Fig. 3).

Fig. 3.

Immunofluorescence analysis comparing marker expression between decidual organoids and source tissues. A–D Representative images showing the expression of specific markers in organoids (top rows) and corresponding source tissues (bottom rows). Panels display staining for: A Vimentin (green) and ER (red); B Epcam (green) and PR (red); C Muc1 (green) and Foxa2 (red); and D G-CSF (green) and G-CSFR (red). Nuclei were counterstained with DAPI (blue). Scale bars: 100 μm for organoids and 50 μm for tissues. Original magnification: 100× (organoids) and 200× (tissues). Images are representative of n = 3 independent biological replicates

Hormone responsiveness of decidual organoids

As shown in Fig. 4, treatment with estradiol (E2) resulted in mild expansion of the organoid lumen, uniform cytoplasmic staining, and regular nuclear arrangement. In contrast, combined treatment with E2, progesterone (P4), and cAMP induced significant luminal expansion and vacuolation, resembling a decidualization-like response. ER expression was significantly increased in the E2 group, while in the E2 + P4 + cAMP group, ER levels were slightly reduced compared to the E2 group but remained higher than in the control group (p > 0.05). PR expression was minimally detected in the control group but markedly increased in the E2 + P4 + cAMP group. Ki-67 positivity was observed in all groups; the proportion of Ki-67-positive cells was significantly higher in the E2 group compared with controls, while in the E2 + P4 + cAMP group, this proportion decreased relative to the E2 group but remained elevated compared to controls.

Fig. 4.

Responsiveness of decidual organoids to reproductive hormone treatments assessed by immunohistochemistry. A–C Representative immunohistochemical images (left panels) and quantification of relative expression levels (right graphs) for: A Estrogen Receptor (ER); B Progesterone Receptor (PR); and C Ki-67. Brown staining indicates positive expression; nuclei were counterstained with hematoxylin (blue). Scale bars: 50 μm. Original magnification: 200×. Images are representative of n = 3 independent biological replicates. Statistical significance was determined using one-way ANOVA followed by Tukey’s multiple comparison test (ns, not significant; *, P < 0.05; **, P < 0.01)

Determination of rhG-CSF intervention concentration on decidual organoid viability

The effect of rhG-CSF on decidual organoid viability was dose-dependent. A significant increase in viability was observed at 100 ng/mL rhG-CSF compared to lower concentrations, while further increases to 200 ng/mL and 500 ng/mL did not yield significant additional changes (Supplementary Figure S2).

Bioinformatics analysis results

To evaluate overall transcriptomic differences, we performed unsupervised principal component analysis (PCA) on the global RNA-seq data. The PCA plot demonstrated a distinct separation between the Control and RSA groups along the first principal component (PC1, 33.2%), indicating fundamental differences in their global transcriptomic profiles despite individual heterogeneity (Supplementary Figure S3A). Subsequent differential expression analysis identified 407 differentially expressed genes following rhG-CSF stimulation (286 upregulated and 121 downregulated), as visualized in the volcano plot and heatmap (Supplementary Figure S3B, C). Further functional characterization via KEGG pathway enrichment analysis suggested significant enrichment in the MAPK, IL-17, JAK-STAT, and NF-kappa B signaling pathways. Consistent with these findings, GSEA demonstrated enrichment trends in the JAK-STAT, MAPK, NOD-like receptor, and TOLL-like receptor signaling pathways (Fig. 5).

Fig. 5.

Functional enrichment analysis reveals activation of the JAK-STAT signaling pathway in decidual organoids. A Bubble plot of the top 20 significantly enriched KEGG pathways for differentially expressed genes. The x-axis represents the GeneRatio. The size of the dots indicates the number of enriched genes (Count), and the color gradient represents the significance level (q-value). B GSEA plot showing signaling pathways significantly enriched in the rhG-CSF treatment group

Western blot

Western blot analysis showed that, following rhG-CSF stimulation, the protein levels of PCNA, VEGFA, IL-1β, and IL-6 were increased. The expression of the anti-apoptotic protein Bcl-2 was upregulated, while the pro-apoptotic protein Bax was downregulated, resulting in a decreased Bax/Bcl-2 ratio (Fig. 6). No significant differences in G-CSF and G-CSFR protein expression were detected between the RSA and control groups in primary decidual tissue (Supplementary Figure S4). In decidual organoids, G-CSF protein levels did not significantly change among the rhG-CSF, AG490, and rhG-CSF + AG490 groups, while G-CSFR expression was significantly elevated in the rhG-CSF and rhG-CSF + AG490 groups. Compared with controls, the p-JAK2/JAK2 and p-STAT3/STAT3 ratios were significantly increased in the rhG-CSF group. In the AG490 group, both ratios were significantly decreased compared to controls. In the rhG-CSF + AG490 group, the p-JAK2/JAK2 and p-STAT3/STAT3 ratios were significantly lower than in the rhG-CSF group (Fig. 7).

Fig. 6.

rhG-CSF promotes inflammation and proliferation while inhibiting apoptosis in decidual organoids. A Representative Western blot images showing the protein expression levels of IL-6, IL-1β, VEGFA, PCNA, Bcl2, and Bax. Tubulin was used as the internal loading control. B Quantitative analysis of the relative protein expression levels. The band intensities were normalized to Tubulin. The ratio of Bax to Bcl2 was calculated to assess apoptosis. Data are presented as mean ± SD from three independent experiments (n = 3). Statistical significance was determined using Student’s t-test. (*, P < 0.05; ****, P < 0.0001)

Fig. 7.

rhG-CSF activates the JAK2/STAT3 signaling pathway in decidual organoids. A Representative Western blot images showing the protein expression levels of G-CSF, G-CSFR, p-JAK2, JAK2, p-STAT3, and STAT3. Tubulin was used as the internal loading control. B Quantitative analysis of the relative protein expression levels. Band intensities of G-CSF and G-CSFR were normalized to Tubulin. The ratios of p-JAK2 to JAK2 and p-STAT3 to STAT3 were calculated to assess pathway activation. Data are presented as mean ± SD from three independent experiments (n = 3). Statistical significance was determined using one-way ANOVA followed by Tukey’s multiple comparison test. (ns, not significant; *, P < 0.05; **, P < 0.01)

Discussion

Research on RSA has long been limited by the scarcity of suitable experimental models. Traditional two-dimensional cell cultures are inadequate for recapitulating the critical functions of decidual tissue during early pregnancy. The pioneering report of endometrial epithelial organoids in 2017 demonstrated that these models can closely mimic the tissue morphology and physiological functions of in vivo organs, thereby significantly advancing the field [27]. Organoid-based in vitro models not only circumvent the ethical challenges inherent in clinical trials but also facilitate the development and validation of diverse potential therapies [32, 33]. Patient-derived organoid models accurately recapitulate individual pathological phenotypes, making them ideal tools for studying the etiology of RSA and for drug screening.

One of the major causes of RSA is endometrial dysfunction [34, 35]. However, direct sampling of endometrial tissue may carry the risk of compromising reproductive potential. In this study, we selected decidual tissue from RSA patients as the source for organoid culture, thereby avoiding ethical concerns and preserving patients’ reproductive capacity. Since RSA occurs during pregnancy, the decidual tissue at this stage more accurately reflects the maternal-fetal interface and the physiological state post-implantation.

Here, we successfully established uterine decidual organoids and demonstrated the feasibility of culturing these organoids in vitro. H&E staining revealed that the decidual organoids exhibited a typical glandular architecture resembling the original decidual tissue, characterized by multilayered cellular organization and cystic structures. PAS staining indicated strong positive signals within the organoid lumen, suggesting the presence of glycoproteins and glycolipids, which are indicative of secretory function analogous to native decidual tissue. Immunohistochemistry confirmed that the organoids respond to reproductive hormones: E2 treatment significantly increased ER and Ki-67 expression, suggesting that E2 promotes ER-mediated proliferation. Furthermore, combined E2 + P4 + cAMP treatment induced features consistent with the secretory phase of the endometrium, demonstrating that these organoids can recapitulate cyclical changes in response to hormonal stimuli, effectively modeling the complex microenvironment of the human endometrium. The establishment rate of RSA organoids in our study was approximately 55%, a figure that reflects the inherent challenges associated with culturing pathological tissues. Unlike healthy proliferative endometrium, which typically yields high establishment rates, RSA decidual tissues are often compromised by extensive inflammation, hemorrhage, and apoptosis, limiting the availability of viable epithelial progenitors. This phenomenon is consistent with other disease-derived organoid models, such as those for endometrial cancer or liver diseases, where success rates generally range from 60% to 70% [36–38]. Additionally, the reported rate represents the average across the entire study period, encompassing the initial phase of method standardization. With the optimized protocol now established, we aim to refine these parameters to enhance the robustness of RSA organoid generation in future studies.

We further validated the organoid model by confirming the expression of multiple molecular markers via immunofluorescence. Vimentin expression in decidual organoids was consistent with that in primary tissue, indicating the preservation of key mesenchymal components critical for maintaining structural integrity [39, 40]. The presence of E-cadherin and EpCAM demonstrated robust cell–cell adhesion properties, mimicking the structure and function of endometrial epithelial cells [41, 42]. High Muc1 expression indicated secretory activity characteristic of endometrial epithelial cells [43, 44], while Foxa2 expression supported a role in regulating cellular differentiation and tissue development [45]. The expression of ER and PR confirmed the organoids’ responsiveness to hormonal stimulation, providing an ideal experimental system for studying hormone signaling pathways [46–48]. Additionally, the concurrent expression of G-CSF and its receptor G-CSFR in the cultured decidual organoids suggests the potential of this model for investigating how decidual cells contribute to the local inflammatory microenvironment in RSA, providing a foundational platform for studying the actions of rhG-CSF.

Our Western blot analyses revealed that G-CSF and G-CSFR protein levels in decidual tissue from RSA patients did not differ significantly from controls. Similarly, rhG-CSF stimulation did not alter G-CSF protein levels in organoids but did lead to an increase in G-CSFR expression. As the primary receptor mediating G-CSF function, G-CSFR expression is a key determinant of the efficiency and strength of G-CSF signaling [49, 50]. Therefore, the upregulation of G-CSFR suggests a potential increase in cellular sensitivity to G-CSF, which may enhance its biological effects [51, 52]. This regulatory mechanism could be beneficial for maintaining decidual health and supporting embryo implantation and development.

During the peri-implantation phase, the endometrium undergoes proliferation, differentiation, and decidualization to create an optimal environment for embryo attachment and growth [53]. PCNA is a well-established marker of cellular proliferation [54, 55], and we observed that rhG-CSF significantly increased PCNA expression in decidual organoids, suggesting a pro-proliferative effect. By enhancing proliferation, rhG-CSF has the potential to increase endometrial thickness and receptivity, thereby supporting implantation and early embryonic development. Additionally, rhG-CSF treatment increased the expression of the anti-apoptotic protein Bcl-2 and decreased the expression of the pro-apoptotic protein Bax, resulting in a reduced Bax/Bcl-2 ratio. This indicates an enhanced anti-apoptotic capacity in organoids following rhG-CSF exposure. Increased VEGFA expression suggests a pro-angiogenic effect, potentially enhancing nutrient and oxygen delivery within the endometrium to create a favorable environment for early pregnancy. IL-1β and IL-6 play important roles in inflammatory regulation; IL-1β facilitates the establishment of immune tolerance during implantation [56, 57], while IL-6, potentially through STAT3 activation, is involved in extracellular matrix remodeling and angiogenesis [58]. Mild, non-bacterial inflammation during early pregnancy is essential for normal implantation [59]. Thus, the increase in IL-1β and IL-6 levels observed in decidual organoids after rhG-CSF treatment suggests that rhG-CSF induces an in vitro cytokine profile conducive to implantation, though further clinical validation is required to confirm its role in promoting pregnancy success in RSA patients.

KEGG and GSEA analyses indicated a significant enrichment and activation of the JAK-STAT signaling pathway following rhG-CSF stimulation. Upon binding to G-CSFR, G-CSF is known to induce phosphorylation of JAK2 at the cytoplasmic domain of the receptor, leading to downstream activation of STAT proteins (mainly STAT3 and STAT5), which mediate cell proliferation, differentiation, and inflammatory responses [60, 61]. The JAK2/STAT3 pathway, in particular, regulates leukemia inhibitory factor (LIF) and downstream STAT3 activation, contributing to making the endometrium receptive to embryo implantation [62]. Inhibition of STAT3 activation significantly reduces implantation rates, underscoring the important role of the JAK2/STAT3 pathway in implantation [63].

Consistent with these findings, we observed that rhG-CSF markedly increased the phosphorylation of JAK2 and STAT3, as evidenced by elevated p-JAK2/JAK2 and p-STAT3/STAT3 ratios in the rhG-CSF group. AG490, a JAK2 inhibitor, significantly suppressed rhG-CSF-induced phosphorylation of JAK2/STAT3, and the addition of rhG-CSF could not reverse this effect. These observations suggest that the protective cellular responses induced by rhG-CSF in this organoid model are closely associated with the activation of the JAK2/STAT3 signaling pathway.

Interestingly, we observed that the upregulation of G-CSFR induced by rhG-CSF was maintained even in the rhG-CSF + AG490 group. This JAK2-independent maintenance of G-CSFR upregulation suggests a complex mechanism of receptor regulation. It is plausible that G-CSFR expression is additionally modulated by parallel signaling cascades, such as the MAPK/ERK or PI3K/AKT pathways, or through specific receptor internalization and recycling dynamics. Such complex receptor regulation warrants further investigation to fully elucidate the network governing G-CSF responsiveness in the decidua.

Several limitations of this study must be acknowledged. Clinically, rhG-CSF is often administered in the non-pregnant period to promote endometrial proliferation and improve the microenvironment, whereas early pregnancy decidua differs from non-pregnant endometrium in cellular composition, signaling activity, and microenvironmental features. Thus, using early pregnancy decidua as the source for organoid culture may not fully reflect the effects of rhG-CSF on the non-pregnant endometrium. Nonetheless, decidual tissue more accurately models the physiological characteristics of the maternal–fetal interface during pregnancy and avoids ethical constraints, making decidual organoid models valuable for studying the mechanisms of pregnancy maintenance.

While our study provides valuable insights into the transcriptomic landscape of decidual organoids following rhG-CSF treatment, we must acknowledge the potential influence of passage effects. It is well-documented that prolonged in vitro culture of organoids can lead to transcriptomic drift, epigenetic modifications, or clonal selection [64, 65]. To strictly mitigate these passage-induced artifacts and ensure the reproducibility of our RNA-seq data, we uniformly utilized early-passage organoids (passage 4) for all sequencing analyses and restricted our functional assays to passages 3 to 5. This strategy effectively preserves the primary characteristics of the source tissue. However, the lack of longitudinal transcriptomic profiling to assess the genomic and phenotypic stability of these organoids at later passages (e.g., passage 10) remains a limitation. Future investigations incorporating single-cell RNA sequencing and long-term stability tracking will be essential to fully characterize the dynamic passage effects in decidual organoid models.

In addition, while we validated the basic establishment of organoids derived from normal pregnancies to ensure system stability, our functional intervention assays were performed specifically in RSA-derived organoids. Therefore, we did not directly compare rhG-CSF responsiveness between RSA and normal organoids in the present study. We hypothesize that normal tissues, which may already operate near a physiological optimum, could display a distinct response profile compared with the “rescue” effect observed in the pathological RSA model. Future studies incorporating larger sample sizes of normal controls are warranted to clarify potential differential pharmacological responses.

Furthermore, there are several limitations regarding our mechanistic exploration that should be noted. First, while we demonstrated that AG490 blocks the rhG-CSF-induced phosphorylation of JAK2 and STAT3, we did not perform additional Western blot analyses to verify the reversal of downstream functional markers (PCNA, Bcl-2, Bax, VEGFA, IL-1β, and IL-6) in the inhibitor-treated groups. Therefore, our current conclusion regarding these downstream effects remains suggestive and relies partially on existing research consensus rather than direct experimental evidence. Second, our mechanistic validation relied solely on a single pharmacological inhibitor without genetic validation, such as siRNA knockdown. Third, due to the inherent technical challenges of extracting transient phosphorylated proteins from 3D organoid models compared to traditional 2D cell lines, the p-JAK2 signal in our Western blot analysis was relatively weak. Consequently, the definitive mechanistic link between rhG-CSF and JAK2 activation is not yet fully established, and our interpretation is informed by the canonical nature of this signaling axis [66–68]. In future investigations, we are committed to employing rigorous genetic approaches and comprehensive validation of downstream effectors to fully elucidate the specific mechanistic details of the G-CSF/JAK2/STAT3 axis in this model.

Finally, organoids lack a complete immune microenvironment and vascular system, limiting their ability to fully recapitulate complex cell–cell interactions and dynamic in vivo changes, thereby restricting detailed exploration of implantation and early developmental mechanisms. Our research group aims to address these limitations in future studies to further advance the understanding of RSA.

Conclusion

In this study, we successfully established and characterized a patient-derived decidual organoid model that effectively recapitulates the structural and functional characteristics of decidual tissue in RSA. Our in vitro findings demonstrate that rhG-CSF promotes cell proliferation, inhibits apoptosis, modulates inflammatory signaling, and enhances angiogenic potential in decidual organoids—cellular responses that are potentially associated with the activation of the JAK2/STAT3 signaling pathway. These observed effects likely represent a targeted rescue of RSA-associated pathological deficits, although concurrent generic trophic effects applicable to normal decidua cannot be ruled out. Notably, while we have established a correlation, direct experimental evidence confirming that the inhibition of this pathway specifically reverses the observed downstream functional effects is currently lacking. These in vitro data provide valuable mechanistic insights that may partially explain the potential therapeutic benefits of rhG-CSF, though further clinical studies and rigorous mechanistic validations are required to fully validate its efficacy and mode of action in RSA management.

Ethics statement.

The studies involving human participants were reviewed and approved by the Medical Ethics Committee of the First Affiliated Hospital of Guangxi Medical University (Project Title: Mechanisms and clinical related research of recurrent miscarriage; Approval Number: 2023-K057-01; Date of Approval: March 14, 2023). The patients/participants provided their written informed consent to participate in this study.

Supplementary Information

Below is the link to the electronic supplementary material.

.

Author contributions

C.W. and X.T. contributed equally to this work and are considered co-first authors.L.P. was responsible for Conceptualization, Design, Project Administration, Funding Acquisition.C.W. and X.T. performed the main Investigation. C.W. conducted the primary Data Curation and Formal Analysis. X.T. focused on Visualization and supporting Data Analysis. Both C.W. and X.T. contributed significantly to the Writing – Original Draft. Y.W. and J.C. assisted with the Investigation and Methodology validation. Z.Z. and W.Y. participated in data collection and provided Resources. Y.Z., W.J., and X.Z. contributed to the final stages of Formal Analysis and Investigation.All authors read and approved the final manuscript.

Funding

This work was supported by grants from the National Natural Science Foundation of China (Nos. 82260306 and 81960281), the Guangxi Key R&D Program (Guike AB20159031 and AB24010080), and the Innovative Clinical Research Climbing Program of the First Affiliated Hospital of Guangxi Medical University (YYZS2022006).

Data availability

RNA-seq data were deposited into the Gene Expression Omnibus database under accession number GSE313966 and are available at the following URL: https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi? acc=GSE313966. All the other raw data supporting the results of this study are available within the paper or archived in Supplementary material files.

Declarations

Ethics approval and consent to participate

The studies involving human participants were reviewed and approved by the Medical Ethics Committee of the First Affiliated Hospital of Guangxi Medical University (Project Title: Mechanisms and clinical related research of recurrent miscarriage; Approval Number: 2023K057-01; Date of Approval: March 14, 2023). The patients/participants provided their written informed consent to participate in this study.

Competing interests

The authors declare no competing interests.