More than 60% of pregnancy losses occur during the first trimester, highlighting the need to understand the role of the oviduct in early pregnancy. In this study, we conditionally ablated the classical progesterone receptor (Pgr) in oviductal epithelial cells called Pgrd/d mouse model. We found that 40% of embryos collected from Pgrd/d females were non-viable or developmentally delayed, indicating that epithelial PGR expression is crucial for embryonic development. Single-cell RNA-sequencing revealed upregulation of proinflammatory genes, including interleukin 22 (IL-22), in the epithelial cells of Pgrd/d females. Pharmacological inhibition of inflammation using non-steroidal anti-inflammatory drugs significantly reduced IL-22 levels in the oviducts and rescued embryonic developmental rates in Pgrd/d females. Co-culture of wild-type zygotes with IL-22 significantly decreased the number of expanded blastocysts. Our findings suggest that progesterone signaling is vital for immunoregulation and normal preimplantation development, potentially providing insights for developing diagnostic tools and therapeutic strategies to address pregnancy failures. 

Oviducts were collected at 0.5 dpc from Pgr^f/f and Pgr^d/d (n = 8 mice/group). In another group of exogenous P₄ treatment, oviducts were collected from OVX females treated with Veh (n = 4 mice), P₄ for 2 h (n = 4 mice), or 24 h (n = 3 mice). Oviducts were pooled and collected in L15+1%FBS and placed on ice. The oviduct was dissected into distal (infundibulum with ampulla, referred to as InfAmp) or proximal (isthmus with UTJ, referred to as IsthUTJ) regions. Oviductal cell isolation was performed using 0.25% trypsin-EDTA (T4049, MilliporeSigma) enzymatic digestion as previously described. Cell clumps and tissue debris were strained twice through 40-μm cell strainers. Cells were then spun down and resuspended in 0.04% BSA (11020–021, AlbuMax, ThermoFisher Scientific) in PBS. The final cell concentration was targeted for 8,000 cells/run. scRNA-seq libraries were performed using the manufacturer’s protocol (10X Genomics Inc). Single Cell 3’ v2 chemistry was used. Briefly, individual cells (~8,000 cells/run) were separated into droplets by Gel bead in EMulsion (GEM) technology using 10X Chromium Controller. Emulsion beads were broken, and barcoded cDNAs were pooled for amplification. Libraries generated were sequenced using an Illumina NovoSeq 6000, targeting 1 billion reads for the pool in 1 lane, paired-end, and 100-bp read length.

 The computer was equipped with AMD Ryzen 9 5900X 12-Core Processor, 24 Threads, and used as a server using the Linux operating system. Raw data were processed using CellRanger-6.0.1 with mouse reference genome mm10-r-102 for sequence alignment. Web output for each dataset was included in Table S1. Generated loom files were then read in scanpy v1.9.1 using JupyterLab33 for analyses and saved in an h5ad format. For subsequent data analysis and visualization, a virtual machine (Docker) was used to create the analysis environment. Analysis packages including scanpy 1.9.1, anndata 0.7.5, umap 0.4.6, numpy 1.19.4, scipy 1.5.3, pandas 1.1.4, scikit-learn 0.23.2, statsmodels 0.12.1, python-igraph 0.8.3, Louvain 0.7.0, and leidenalg 0.8.2 were installed on the Docker. scRNA-seq data were then processed using similar pipeline, analysis packages, and cutoffs as our previous work, except “highly variable genes” were set to 4000 genes. Analysis was carried out using scanpy which was designed specifically for scRNA-seq data QC and analysis inspired by Seurat’s clustering tutorial. Principal component analysis (PCA) and Uniform Manifold Approximation and Projection (UMAP) clustering analyses were performed using 10 neighbors and PCs 1–37 based on variance ratio. Batch balanced K-Nearest Neighbor (BBKNN) was used to correct technical artifacts from multiple sample collections. The top 1000 differentially expressed genes were used for the determination of biological processes (BPs) enriched in each sample.