ABSTRACTTechnical advances in whole tissue imaging and clearing have allowed 3D reconstruction of exocrine uterine glands deep seated in the endometrium. However, there are limited gland structure analysis platforms to analyze these imaging data sets. Here we present a pipeline for segmenting and analyzing uterine gland shape. Using this segmentation methodology, we derive individual metrics to describe gland length, shape, and branching patterns. These metrics are applied to quantify gland behavior with respect to organization around the embryo and proximity of each individual unit to the uterine lumen. Using this image analysis pipeline we evaluate uterine glands at the peri-implantation time points of a mouse pregnancy. Our analysis reveals that upon embryo entry into the uterus glands show changes in length, tortuosity, and proximity to the uterine lumen while gland branch number stays the same. These shape changes aid in reorganization of the glands around the site of embryo implantation. We further apply our analysis pipeline to human and guinea pig uterine glands, extending feasibility to other mammalian species. This work serves as a resource for researchers to extract quantitative, reproducible morphological features from three-dimensional uterine gland images in order to reveal insights about functional and structural patterns.

Uterine glands are branched, tubular structures whose secretions are essential for pregnancy success. It is known that pre-implantation glandular expression of leukemia inhibitory factor (LIF) is crucial for embryo implantation, however contribution of uterine gland structure to gland secretions such as LIF is not known. Here we use mice deficient in estrogen receptor 1 (ESR1) signaling to uncover the role of ESR1 signaling in gland branching and the role of a branched structure in LIF secretion and embryo implantation. We observed that deletion of ESR1 in neonatal uterine epithelium, stroma and muscle using the progesterone receptor PgrCre causes a block in uterine gland development at the gland bud stage. Embryonic epithelial deletion of ESR1 using a mullerian duct Cre line - Pax2Cre, displays gland bud elongation but a failure in gland branching. Surprisingly, adult uterine epithelial deletion of ESR1 using the lactoferrin-Cre (LtfCre) displays normally branched uterine glands. Intriguingly, unbranched glands from Pax2Cre Esr1flox/flox uteri fail to express glandular pre-implantation Lif, preventing implantation chamber formation and embryo alignment along the uterine mesometrial-antimesometrial axis. In contrast, branched glands from LtfCre Esr1flox/flox uteri display reduced expression of glandular Lif resulting in delayed implantation chamber formation and embryo-uterine axes alignment but deliver a normal number of pups. Finally, pre-pubertal unbranched glands in control mice express Lif in the luminal epithelium but fail to express Lif in the glandular epithelium even in the presence of estrogen. These data strongly suggest that branched glands are necessary for pre-implantation glandular Lif expression for implantation success. Our study is the first to identify a relationship between the branched structure and secretory function of uterine glands and provides a framework for understanding how uterine gland structure-function contributes to pregnancy success.

STUDY QUESTION:How does ovarian stimulation (OS), which is used to mature multiple oocytes for ART procedures, impact the principal cellular compartments and transcriptome of the human endometrium in the periovulatory and mid-secretory phases? SUMMARY ANSWER:During the mid-secretory window of implantation, OS alters the abundance of endometrial immune cells, whereas during the periovulatory period, OS substantially changes the endometrial transcriptome and impacts both endometrial glandular and immune cells. WHAT IS KNOWN ALREADY:Pregnancies conceived in an OS cycle are at risk of complications reflective of abnormal placentation and placental function. OS can alter endometrial gene expression and immune cell populations. How OS impacts the glandular, stromal, immune, and vascular compartments of the endometrium, in the periovulatory period as compared to the window of implantation, is unknown. STUDY DESIGN, SIZE, DURATION:This prospective cohort study carried out between 2020 and 2022 included 25 subjects undergoing OS and 25 subjects in natural menstrual cycles. Endometrial biopsies were performed in the proliferative, periovulatory, and mid-secretory phases. PARTICIPANTS/MATERIALS, SETTING, METHODS:Blood samples were processed to determine serum estradiol and progesterone levels. Both the endometrial transcriptome and the principal cellular compartments of the endometrium, including glands, stroma, immune, and vasculature, were evaluated by examining endometrial dating, differential gene expression, protein expression, cell populations, and the three-dimensional structure in endometrial tissue. Mann-Whitney U tests, unpaired t-tests or one-way ANOVA and pairwise multiple comparison tests were used to statistically evaluate differences. MAIN RESULTS AND THE ROLE OF CHANCE:In the periovulatory period, OS induced high levels of differential gene expression, glandular-stromal dyssynchrony, and an increase in both glandular epithelial volume and the frequency of endometrial monocytes/macrophages. In the window of implantation during the mid-secretory phase, OS induced changes in endometrial immune cells, with a greater frequency of B cells and a lower frequency of CD4 effector T cells. LARGE SCALE DATA:The data underlying this article have been uploaded to the Genome Expression Omnibus/National Center for Biotechnology Information with accession number GSE220044. LIMITATIONS, REASONS FOR CAUTION:A limited number of subjects were included in this study, although the subjects within each group, natural cycle or OS, were homogenous in their clinical characteristics. The number of subjects utilized was sufficient to identify significant differences; however, with a larger number of subjects and additional power, we may detect additional differences. Another limitation of the study is that proliferative phase biopsies were collected in natural cycles, but not in OS cycles. Given that the OS cycle subjects did not have known endometrial factor infertility, and the comparisons involved subjects who had a similar and robust response to stimulation, the findings are generalizable to women with a normal response to OS. WIDER IMPLICATIONS OF THE FINDINGS:OS substantially altered the periovulatory phase endometrium, with fewer transcriptomic and cell type-specific changes in the mid-secretory phase. Our findings show that after OS, the endometrial microenvironment in the window of implantation possesses many more similarities to that of a natural cycle than does the periovulatory endometrium. Further investigation of the immune compartment and the functional significance of this cellular compartment under OS conditions is warranted. STUDY FUNDING/COMPETING INTERESTS:Research reported in this publication was supported by the National Institute of Allergy and Infectious Diseases (R01AI148695 to A.M.B. and N.C.D.), Eunice Kennedy Shriver National Institute of Child Health and Human Development (R01HD109152 to R.A.), and the March of Dimes (5-FY20-209 to R.A.). The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health or March of Dimes. All authors declare no conflict of interest.

Ovarian stimulation (OS) exposes the endometrium to supraphysiologic levels of estradiol (E2) and an early rise in progesterone (P4). We examined the glandular, stromal, vascular and immune compartments of the endometrium in order to better understand the mechanisms underlying OS-mediated changes. We hypothesized that OS-mediated changes would be apparent as early as the periovulatory (PO) endometrium and that these changes would persist into the mid-secretory (MSE) phase leading to endometrial aberrations during the window of receptivity.