All animal experiments adhered to the regulations stipulated in the Administration of Affairs Concerning Experimental Animals in China and received approval from the Animal Welfare and Research Ethics Committee of Inner Mongolia Agricultural University (Approval ID: NND2021013).

The E. coli O157:H7 strain used in this study was purchased from Bena Culture Collection (BNCC186579, Beijing, China). A 1 ml suspension of E. coli O157:H7 strain (at a concentration of 1 × 10⁷ CFU) preserved in the laboratory was inoculated into 100 ml of Luria-Bertani (LB) broth (Oxoid, Basingstoke, LTD, UK). Incubate the culture at 37°C with shaking at 200 rpm for 12 h, or until the OD₆₀₀ of the culture reaches 0.9. The bacterial suspension was serially diluted and plated onto LB agar. After incubation at 37°C for 18 h, colonies were counted, and the concentration was quantified as CFU/ml.

In this study, bovine rib samples were obtained from healthy adult Holstein cows at the Beiya Slaughterhouse in Hohhot, Inner Mongolia, China. All animals had passed veterinary health inspections and quarantine assessments, and were confirmed to be free of major diseases. Slaughter was conducted solely for commercial food production purposes. After slaughter at the abattoir, cow rib bones were collected immediately post-mortem, placed on ice, and transported to the laboratory for further processing. All bone marrow samples were processed within 1 hour of animal death to preserve cell viability and ensure reproducibility of the results. Bone marrow was extracted from cow ribs by flushing with phosphate-buffered saline (PBS; Hyclone, Logan, UT, USA). The collected cells were centrifuged at 2,900 g for 8 min, and the supernatant was discarded. The cells were subsequently resuspended in erythrocyte lysis buffer and incubated for 5 min to facilitate lysis. The cells were then centrifuged at 1,300 g for 8 min and subsequently cultured in RPMI 1,640 medium supplemented with 20% fetal bovine serum (Hyclone, Logan, UT, USA) and 20 ng/ml M-CSF (Kingfisher Biotech Inc., USA) at 37°C in a 5% CO₂ atmosphere. After 7 days, unattached cells were removed, and adherent cells were used for the experiments. To induce M1 macrophages, cells (2 × 10⁶ per well) were treated with 1 μg/ml of lipopolysaccharide for 24 h, followed by an 8-h resting period. M1 BMDMs were confirmed by immunofluorescence (Supplementary Figure 1).

The tissues used in this study were obtained from animals slaughtered at the Beiya Slaughterhouse in Hohhot, Inner Mongolia, China. Prior to slaughter and tissue collection, all animals underwent health screening, including tests for common reproductive diseases. To minimize the potential impact of underlying infections on the study results, only animals that met these health standards were selected. The experiment selected 30 healthy Holstein dairy cows aged 15–18 months, and fresh uterine horn samples were collected (~400 kg). All samples were collected from sexually mature cows in the proestrus phase of the estrous cycle. Endometrial tissue culture was established following the method of Li (5). After rinsing the uterine horns 3 times with PBS containing 100 IU/ml penicillin, 100 IU/ml streptomycin, and 2.5 mg/ml amphotericin B, the uterine horns were incubated at 4°C for 1 h. Under aseptic conditions, the uterine horns of cows were longitudinally incised, and small pieces of endometrial tissue measuring 2 × 2 mm were excised. These explants were then randomly allocated to six-well plates for culture.

BMDMs were randomly divided into four groups: Control, PGD₂-treated, E. coli-infected, and PGD₂ + E. coli co-treatment groups. In the PGD₂-treated and co-treatment groups, cells were pretreated with PGD₂ at a final concentration of 1 × 10⁻⁶ M (1 × 10⁻⁶ M; Cayman Chemical Company, Ann Arbor, MI, USA) for 24 h prior to infection (see Supplementary Figure 2 for concentration selection). Subsequently, E. coli was added at a multiplicity of infection (MOI) of 5:1 for further experimentation. After 1 h of infection with 100 μg/ml tobramycin to remove extracellular bacteria, the medium was changed to continue the culture (38).

The bovine endometrial tissues were randomly assigned to the following experimental groups: Control, PGD₂ (1 × 10⁻⁵ M), PGD₂ (4 × 10⁻⁹ M), E. coli, PGD₂ (1 × 10⁻⁵ M) + E. coli, and PGD₂ (4 × 10⁻⁹ M) + E. coli. Prior to infection, tissues in the PGD₂ treatment groups were pretreated with the corresponding concentration of PGD₂ for 24 h. Subsequently, infection was performed using E. coli at a concentration of 1 × 10⁶ CFU/ml. Each group was set up in three independent wells, with tissue samples obtained from different individual cows to enhance biological reproducibility. In addition, each well was analyzed in triplicate, and the average value was used for subsequent data analysis.

BMDMs were treated with PGD₂ (1 × 10⁻⁶ M) for 24 h, followed by infection with E. coli at a MOI of 5:1 for 4 h. Total RNA was isolated from the cells using TRIzol reagent. The RNA concentration and integrity were then evaluated using a NanoDrop spectrophotometer (Thermo Scientific, USA) to confirm that the RNA remained intact. The libraries were constructed following the manufacturer's instructions, after which transcriptome sequencing was performed and the data analyzed by OE Biotech Co (Shanghai, China). Gene counts for each sample were normalized using DESeq2 software, with expression levels estimated based on the baseMean value. The fold changes were calculated, and significance was assessed using a negative binomial distribution test (NB test). The differentially expressed genes (DEGs) selected met the criteria of |log2 Fold Change| > 1.5 and a significance P-value < 0.05. Significantly enriched genes and related pathways were identified through GO and KEGG enrichment analysis and their main biological functions were explored (with a significance threshold of P < 0.05). Finally, further enrichment analysis of GO and KEGG pathways was conducted using the OECloud tool (https://cloud.oebiotech.com/task/).

Cells were collected at 2, 4, and 6 h post-E. coli infection. Total RNA was extracted using the Axygen RNA Kit (Axygen Scientific, USA), and RNA was reverse-transcribed into cDNA using the PrimeScript RT kit (Vazyme, Nanjing, China). The mRNA levels of various genes including Prostaglandin-Endoperoxide Synthase 2 (PTGS2), HMGB-1, and HABP-2 were quantified using the SYBR Green Master Mix Kit (Roche Applied Science, Mannheim, Germany). Gene expression was normalized according to actin mRNA levels, and data were analyzed using the 2^−ΔΔCt method (39). All primers information is in Table 1.

We performed molecular docking analysis of the drug PGD₂ with the key target protein PTGDR. After removing all associated ligands, the three-dimensional structures of PGD₂ and PTGDR were obtained from PubChem and the Protein Data Bank (PDB), respectively. AutoDock v1.5.7 was then used for ligand preparation, including the removal of water molecules, processing of nonpolar hydrogen atoms, and identification of the active binding site. Docking conformations were calculated using AutoDock Vina, and the optimal docking model was selected. Finally, the key binding amino acids and optimal binding conformation were visualized using the PyMOL molecular graphics system v2.0 in the Python environment (40).

The concentrations of PGD₂, TNF-α, IL-1β, IL-6, IL-8 and IL-10 in the supernatants of BMDMs and endometrial tissue were measured using bovine PGD₂ (Cayman Chemical Company, Ann Arbor, MI, USA), TNF-α and IL-6 (R&D Systems, Minneapolis, MN, USA), and IL-1β, IL-8, and IL-10 (Kingfisher Biotech, St. Paul, MN, USA) ELISA kits according to the manufacturer's instructions.

Total protein was extracted from treated cells using the M-PER mammalian protein extraction reagent (Thermo Scientific, Waltham, MA, USA), and protein concentrations were determined using the BCA Protein Assay Kit (Thermo Scientific, Rockford, IL, USA). A total of 10 μg of protein from each sample was separated using a 12% SDS-PAGE gel, and the proteins were subsequently transferred to a polyvinylidene fluoride (PVDF) membrane for Western blot analysis. Subsequently, the membrane was blocked with StartingBlock™ (TBS) blocking buffer (Thermo Fisher, MA, USA) at room temperature for 1 h, followed by incubation with primary antibodies at 4°C for 14 h. The primary antibodies employed were antiphospho-ERK, anti-ERK, antiphospho-p38, anti-p38, antiphospho-NF-κB p65, anti-NF-κB p65 (Cell Signaling Technology, 1:1,000 dilution), and anti-GAPDH (1:10,000) monoclonal antibodies. Immunoreactive bands were visualized via chemiluminescence using horseradish peroxidase-conjugated secondary anti-rabbit and anti-mouse antibodies along with chemiluminescent substrate (Thermo Scientific). The band density on the blots was quantified using ImageJ software (National Institutes of Health, Bethesda, MD, USA).

Cell viability and intracellular E. coli survival were assessed via the MTT assay. BMDMs were seeded in 96-well plates at a density of 1 × 10⁴ cells per well, with 180 μl of culture medium added to each well, and cultured at 37°C in 5% CO₂. After 24 h of PGD₂ treatment, cells were infected with E. coli at an MOI of 5:1 for 2.5 and 6 h. Cell viability was evaluated using the MTT assay following the guidelines provided by the manufacturer (Solarbio, Beijing, China).

To investigate the impact of PGD₂ on phagocytosis and bacterial killing, BMDMs were cultured at a density of 2 × 10⁶ cells per 35 mm glass-bottom Petri dish. After treatment, cells were labeled with 8 μm 11′-dioctadecyl-3,33′3′-tetramethylindocarbocyanine perchlorate (DiI, Thermo Scientific). After labeling E. coli with Hoechst 33258 dye for 30 mins, BMDMs were infected for 30 mins, 2.5 h, and 6 h at 37°C. The cells were then fixed with 4% paraformaldehyde and imaged using a confocal microscope (LSM 800; Carl Zeiss, Oberkochen, Germany) at × 400 magnification.

Immunofluorescence analysis of endometrial tissues from dairy cows was performed using established methods (41). After frozen sections, they were incubated overnight at 4°C using primary antibodies (1:100 dilution) to HMGB-1 and HABP-2. After blocking, tissue sections were incubated with donkey anti-rabbit IgG-Alexa Fluor 647 fluorescent secondary antibody (1:1000 dilution; Abcam) for 1 h in the dark. Imaging and quantification of fluorescence intensity were conducted using confocal microscopy (LSM 800, Zeiss, Oberkochen, Germany) at × 400 magnification.

Sections were made by dehydration through an alcohol gradient (70%, 80%, 90%, 100%) and paraffin embedding.These sections were then stained with hematoxylin and eosin (H&E) and imaged using an Axio Scan Z1 slide scanner (Zeiss, Thornwood, NY, USA). All images were captured under identical conditions.

Data were analyzed using GraphPad Prism 8 (GraphPad Software, CA, USA), and results are presented as mean ± standard deviation (SD). Statistical significance was evaluated using one-way analysis of variance (ANOVA) with Tukey's multiple comparisons test or two-way ANOVA with Bonferroni's post-hoc test, depending on the experimental design. Differences were considered statistically significant when the P value was < 0.05 (^*P < 0.05, ^**P < 0.01, ^***P < 0.001, and ^****P < 0.0001).

Transcriptome sequencing was performed to identify potential target genes associated with E. coli infection in BMDMs. The principal component analysis (PCA) identified distinct differences among various sample groups based on transcriptome data, with PCA1 and PCA2 accounting for 83.77 % and 12.49 % of the total variation, respectively (Figure 1A). Transcriptomic analysis of E. coli-infected BMDMs identified 4,522 significantly dysregulated genes, including 2,141 upregulated and 2,381 downregulated genes, based on a q-value < 0.05 and |log2 fold change (FC)| > 1 (Figures 1A, B). Gene Ontology (GO) analysis revealed that the most significant differential gene expression was observed in the inflammatory response (Figure 1C). The key genes involved were IL17A, IL17F, TNFAIP6, and PTGS2 (Figure 1D). Among these, we focused on PTGS2, which showed a significant increase in mRNA expression at 2, 4, and 6 h post-infection (Figure 1E, P < 0.01). In E. coli-infected macrophages, the secretion levels of PGD₂ were elevated at 6 and 24 h (Figure 1F, P < 0.0001). These results indicate that PGD₂ plays a critical role in the response of BMDMs to E. coli infection.

Molecular docking analysis of the interaction between PGD₂ and the PTGDR-associated protein (PDB ID: 7M8W) revealed that PGD₂ forms a stable binding with the PTGDR-associated protein, with a binding energy of −7.9 kcal/mol (Figure 2A). In addition, MTT assay results showed no significant difference in cell viability between PGD₂-treated BMDMs for 24 h and the control group, indicating that PGD₂ did not significantly impact macrophage viability (Figure 2B). These results indicate that PGD₂ binds effectively to PTGDR and does not exert a significant effect on macrophage viability.

Based on KEGG enrichment analysis, differentially expressed genes were significantly enriched in multiple inflammation-related signaling pathways, such as TNF signaling pathway, NF-κB pathway, cytokine-cytokine receptor interaction, and MAPK signaling pathway (Figure 3A). Therefore, we measured the secretion of cytokines and chemokine using ELISA and analyzed the activation of NF-κB and MAPK signaling pathways by Western blot. The ELISA results showed that PGD₂ pretreatment significantly increased the secretion of TNF-α, IL-1β, IL-6, IL-8, and IL-10 in BMDMs at 6 h post-E. coli infection. However, at 12 and 24 h post-infection, PGD₂ significantly downregulated the secretion of TNF-α, IL-1β, and IL-8 compared to the E. coli infection group, while the secretion of IL-6 and IL-10 remained elevated (Figures 3B–F, P < 0.01).

In addition, PGD₂ treatment significantly enhanced the phosphorylation levels of MAPK (ERK, P38) and NF-κB (P65) in E. coli-infected BMDMs at 15 and 30 min (Figure 4, P < 0.01). However, at 60 min post-infection, PGD₂ attenuated the phosphorylation levels of MAPK in BMDMs (Figure 4, P < 0.05). These findings suggest that PGD₂ may influence the activation of macrophage MAPK and NF-κB signaling pathways during E. coli infection, thereby affecting the secretion of pro-inflammatory cytokines and chemokines.

The effect of PGD₂ on the ability of BMDMs to phagocytose and kill E. coli was assessed using Dil-labeled BMDMs and Hoechst-stained E. coli. After 0.5 h post-infection, no significant differences were observed, suggesting that PGD₂ does not affect the phagocytic ability of BMDMs (Figures 5A, C). After 2.5 h of infection, the fluorescence intensity of bacteria in the PGD₂ was higher in the treatment group than the E. coli infection group, indicating that PGD₂ reduced the killing ability of BMDMs. However, after 6 h of infection, the fluorescence intensity of bacteria in the PGD₂ treatment group was lower, suggesting that PGD₂ enhanced the killing ability of BMDMs against E. coli (Figures 5B, D, P < 0.05). The effect of PGD₂ on E. coli survival within macrophages was evaluated through the MTT assay. Consistent with these observations, PGD₂ demonstrated a significant increase in the survival of internalized E. coli in macrophages (Figure 5E, P < 0.05). Together, these results suggest that PGD₂ decreased intracellular killing in macrophages.

To determine the effect of PGD₂ on the damage caused by E. coli infection in BMDMs and bovine endometrial tissues, we measured the expression of DAMPs (HMGB-1, HABP-2) using qPCR and immunofluorescence. Compared to the E. coli infection group, PGD₂ significantly increased HMGB-1 and HABP-2 mRNA expression levels in BMDMs (Figure 6A, P < 0.05). Immunofluorescence staining of E. coli-infected endometrial explants showed that PGD₂ treatment led to an increased expression of HMGB-1 and HABP-2 at 9 h post-infection (Figures 6B–D, P < 0.0001). However, this expression was decreased at 24 h (Figures 6B–6D, P < 0.0001). In conclusion, our data demonstrate that PGD₂ modulates the release of HMGB-1 and HABP-2 in E. coli-infected endometritis in dairy cows, exhibiting distinct roles at different times.

To gain more insight into the effects of PGD₂ on uterine injury, we assessed histologic changes in the endometrium of dairy cows. Hematoxylin and Eosin (H&E) staining results demonstrate that both the control and PGD₂-treated groups maintained intact endometrial structures, characterized by tightly arranged epithelial cells and well-defined glands and blood vessels. At 9 h post-infection, E. coli-infected tissues exhibited complete shedding of endometrial epithelial cells although glandular epithelial cells remained largely unaffected. PGD₂-treated groups (1 × 10⁻⁵ M and 4 × 10⁻⁹ M) also showed endometrial epithelial cells shedding and loosening of glandular epithelial cells, some of which exhibited partial disintegration (Figure 7A). At 24 h post-infection, PGD₂ treatment reduced endometrial damage, with the E. coli-infected group showing severe epithelial and glandular cell loss and necrosis, whereas the PGD₂-treated groups retained relatively intact glandular and vascular structures (Figure 7B). Additionally, measurements of endometrial epithelial thickness and gland count (42) revealed no significant differences in epithelial thickness between the PGD₂-treated and E. coli-infected groups at both 9 and 24 h. Notably, at 9 h post-infection, the PGD₂-treated group showed a reduction in gland count compared to the E. coli-infected group, a trend that reversed by 24 h (Figures 7C, D, P < 0.05). These findings provide further evidence that PGD₂ may play a dual role in E. coli-induced endometritis, aggravating tissue damage during the early phase of infection while contributing to tissue protection at later stages.

We assessed cytokine and chemokine secretion in E. coli-infected endometrial tissue. After 9 h, PGD₂ treatment significantly increased TNF-α, IL-1β, IL-6, and IL-8 levels compared to the E. coli infection group, while IL-10 levels decreased. However, at 24 h post-infection, the expression levels of TNF-α, IL-1β, and IL-6 were markedly reduced, while the secretion of the chemokine IL-8 continued to increase, and the anti-inflammatory cytokine IL-10 was significantly upregulated (Figure 8, P < 0.05). These findings suggest that PGD₂ exerts a time-dependent immunoregulatory effect during E. coli infection, promoting inflammatory responses at the early stage while potentially exerting anti-inflammatory effects at the later stage.