PSII-5 - Characterization of the endometrial epithelium of postpartum dairy cows using a multimodal transcriptomic approach to investigate disease programming of fertility

Abstract: The productive lifespan of cattle destined to milk production is currently suboptimal in most dairy systems across the US due to the high incidence of health disorders and reproductive failure affecting postpartum lactating cows. Bacterial access to the underlying endometrial stroma, early postpartum, can create a chronic inflammatory state. Chronic endometrial inflammation is linked to metritis, an early postpartum uterine disease affecting approximately 40% of lactating dairy cows and delays the reestablishment of the uterine histoarchitechture. We investigated decreased fertility in lactating dairy cows that experience uterine disease and determined that metritis delays the reestablishment of the endometrial epithelium within the first 30 days postpartum (dpp), with greater impact observed within the epithelium located at the endometrial stratum basalis (closest to myometrium). These findings led us to hypothesize that early postpartum uterine disease prevents the adequate regeneration of the GE through its effect on a population of stem/progenitor cells that may reside within the stratum basalis and potentially give rise to the GE during uterine involution. To test this hypothesis, we used a combined transcriptomic approach to 1) better comprehend the heterogeneity of transcriptional profiles within the regenerating endometrial epithelium using high-resolution single nuclei RNA-seq, 2) to further characterize the spatial distribution of distinct and/or unique epithelial populations within the bovine endometrium using spatial transcriptomics. To test our hypothesis, we utilized the whole tissue Visium technology (10X Genomics) of healthy (n = 3) and snRNA-seq (Chromium Single Cell 3’) of healthy (n =1) and diseased (n = 1) cows slaughtered at 30 dpp. Fresh dissociated endometrial tissue (snRNA-seq) and OCT-fixed cross-sections were processed and sequenced in an Illumina NovaFlowSeq 6000. FASTQ files and respective images (Visium) were processed for each sample and aligned to the Bos taurus reference genome using SpaceRanger or CellRanger countpipelines. Subsequent analyses were conducted through the Seurat package of RStudio. Count matrices were individually normalized using SCTransform, and subsequently integrated (CCAIntegration) separately for Visium and Chromium experiments. Chromium samples were submitted to dimensionality reduction and clustering, and subsequently mapped to Visium objects through anchor-based integration for prediction of cell type spatial distribution. We identified a population of LGR5+ epithelia mapped to the upper stratum basalis, and two populations of N-cadherin (CDH2+) at the lower basalis. At single-nuclei resolution, N-cadherin+ clusters were marked by unique signatures of Wnt signaling (CDH2+LEF1+) and pluripotency (CDH2+KIT+), while the LGR5+ cluster was marked by differential expression of SOX9, suggesting the presence of an epithelial stem/progenitor pool. When stratified by disease status, LGR5 expression levels were lower in the disease when compared to the healthy (Chromium) while CDH2+ populations were scarcer in the metritis dataset, collectively suggesting a putative role of disease in modulating the transcriptional programs of various epithelial phenotypes.