Jan Joris Brosens

SUMMARY: The inflammatory decidual reaction renders the cycling endometrium transiently permissive for embryo implantation before transforming it into the decidua, the maternal bed accommodating the fetal placenta during pregnancy. Disruptions in decidual tissue remodeling are linked to miscarriage and other pregnancy disorders. However, endometrial assessment is hampered by a lack of affordable technologies capable of mapping the spatiotemporal dysregulation of this dynamic and complex tissue. Employing a graph neural network on whole slide images of 493 CD56-immunostained endometrial samples, Endometronome was developed as a deep learning tool to spatially track the decidual reaction and provide accurate estimates of marker gene expression. When applied to 2,690 additional biopsies, this model consistently identified morphological correlates of prior miscarriage burden, a proxy for future risk. Further, a morphological signature indicative of metabolic glandular impairment discriminated between clinical miscarriage presentations. These findings illustrate how advanced imaging analysis of routine histology can transform miscarriage prevention strategies.

Histological images are composed of diverse local structures such as cells, glands, or tissue patches, whose organisation and relative abundance reflect underlying biological processes. While most computational approaches focus on analysing individual features, many clinically relevant signals arise from changes in the composition of these structures rather than isolated measurements. This motivates the need for explicitly modelling how groups of similar instances vary across samples and relate to outcomes. We present DCAFA (Differential Community Abundance and Feature Attribution Analysis), a regression-based framework for analysing hierarchical biomedical data through both compositional and feature-level perspectives. DCAFA groups instances into latent communities representing recurring morphological or phenotypic patterns, and then performs two complementary analyses: (i) community composition analysis, which identifies groups that are enriched or depleted across outcomes, and (ii) feature attribution analysis, which quantifies how instance-level features relate to outcomes directly or within specific communities. Both use generalised linear and mixed-effects models, enabling covariate adjustment and inference through effect sizes, confidence intervals, and false discovery rate control. We demonstrate the utility of DCAFA across multiple biomedical settings, including endometrial histopathology, spatial transcriptomics, multiplex immunofluorescence imaging, and predefined cell-type analyses in colorectal cancer. These examples identify interpretable compositional shifts and context-specific feature associations that are not captured by conventional feature-based approaches. By unifying differential abundance testing and feature attribution within a single statistical framework, DCAFA serves as an openly available toolbox that provides a practical and interpretable means of linking tissue composition with clinical or molecular outcomes in biomedical imaging data. Code available at: https://github.com/wgrgwrght/DCAFA

What is the relationship between constitutive CD138 expression in the endometrium and the reproductive tract microbiota composition? The presence of CD138⁺ cells in endometrial stroma is cycle-dependent and associated with impaired luteal phase endometrial timing but not altered vaginal or endometrial microbial composition. CD138-diagnosed chronic endometritis (CE) is associated with adverse reproductive outcomes including recurrent pregnancy loss (RPL) in uncontrolled studies. However, CD138 is constitutively expressed in the endometrium, potentially confounding the reported associations between CE, adverse endometrial function, and early pregnancy loss. Translational cohort study of a subset of 103 samples derived from 737 women embedded within the CERM trial, a double-blinded, randomized interventional trial evaluating the impact of pre-pregnancy antibiotic treatment for CE in RPL patients. Women aged ≥18 to 200 CD138-positive stromal cells/10 mm²) in 26 out of 27 proliferative endometrial samples. While CD138 immunoreactivity in the stroma declines markedly following ovulation (Mann–Whitney U-test; P < 0.005), gene expression analysis demonstrated a reduction in SDC1 expression encoding CD138/syndecan-1, across the menstrual cycle. When compared to CD138-negative samples, conspicuous diffuse staining in the stromal compartment was associated with significantly earlier endometrial histological dating (P < 0.01) and lower molecular timing ratios (P < 0.01). Poor correlation between CD138 and immunoreactivity was demonstrated. Sequencing of paired vaginal and ectocervical swabs and endometrial Tao brush samples collected from 114 patients demonstrated tightly interconnected microbial composition throughout the reproductive tract. No significant difference in vaginal, ectocervical, or endometrial community state type with CD138 expression was demonstrated. Analysis of supernatants of vaginal and ectocervical swabs and Tao Brush revealed an inverse correlation between the severity of stromal CD138 immunoreactivity in endometrial stroma and secreted levels of IL-10, TNF-α, and VEGF (q < 0.05). Microbial and Metataxonomic raw data are available in the European Nucleotide Archive (Projects PRJEB83331 and PRJEB83332). This study relied on patient-reported ovulation-based timing. This was, however, associated with the provision of validated ovulation tests. In addition, the study is limited by lack of collection of data on the underlying fertility-related co-morbidities due to exclusion of known contributory co-morbidities at the point of recruitment. This study challenges the purported relationship between CD138⁺ CE and the pathophysiology of CE-associated RPL. The findings indicate endometrial CD138 levels are non-immune and non-bacterial driven and are associated with endometrial immaturity. CD138-based CE testing and treatment should not be performed outside of a research context. Funding was provided by the Efficacy and Mechanism Evaluation (EME) Programme a National Institute for Health and Care Research and Medical Research Council partnership (17/60/22). Further funding was from Tommy’s National Centre for Miscarriage Research, and the Imperial National Institute for Health and Care Research Biomedical Research Centre Pregnancy and Prematurity Theme. G.D.S.C. is supported by the Genesis Research Trust. All authors report no direct conflict of interest. ISRCTN23947730.

Motivation: Many biological and clinical processes are dynamic, yet most datasets are cross-sectional, capturing populations at discrete states rather than tracking individuals over time. This makes it difficult to quantify how populations change across developmental, physiological, or disease-associated conditions. Existing trajectory and transport-based methods often rely on fixed feature spaces, assumptions tailored to transcriptomic time-course data, or approximately linear progression, limiting their ability to model heterogeneous and unbalanced transitions across diverse biomedical modalities. Flexible methods are needed that can infer directed population-level change from cross-sectional data while retaining biological interpretability. Results: We present GOTFlow, a framework for learning directed population transitions from cross-sectional biomedical data using graph-constrained optimal transport in a learned latent space. GOTFlow integrates representation learning with unbalanced optimal transport to jointly estimate embeddings and transport couplings between biological states. This enables hypothesis-driven modelling of progression structures while accommodating non-linear geometry, branching relationships, and changes in population mass. From the inferred transport plans, GOTFlow derives interpretable summaries of dynamics, including drift vectors quantifying transitions, and feature-level transported changes that highlight molecular drivers of progression. In synthetic data, GOTFlow recovered known transitions with strong agreement between inferred and ground-truth drifts. Across three biological applications, endometrial remodelling, breast cancer risk progression, and prion disease, GOTFlow identified state-to-state transitions and biologically meaningful feature shifts reflecting impaired decidualisation, increasing cancer risk, and neurodegenerative progression. These results establish GOTFlow as a general and interpretable framework for analysing directed population dynamics from cross-sectional data. Availability: Code available at: https://github.com/wgrgwrght/GOTFlow

In every menstrual cycle, progesterone acting on estrogen-primed endometrium elicits an inflammatory decidual reaction, rendering it poised for embryo implantation and transformation into the decidua of pregnancy. Here, we show that the sequential functions of the decidual reaction—implantation and decidualization—pivot on the time-sensitive loss of progesterone-resistant DIO2 ⁺ stromal cells that form a specialized implantation niche and reciprocal expansion of progesterone-dependent PLA2G2A ⁺ predecidual cells. Simultaneously, uterine natural killer (uNK) cell proliferation results in the accumulation of immunotolerant subsets. Examination of endometrial biopsies from 924 women revealed that the recurrence risk of miscarriage closely aligns with the incidence of a weakened or stalled decidual reaction, more so than poor uNK cell expansion. Analysis of paired biopsies obtained in different cycles and modeling in assembloids intimated that prior miscarriages disrupt intercycle endometrial homeostasis and calibration of the decidual reaction. Our findings show that erosion of the decidual reaction following a miscarriage drives the recurrence risk irrespective of maternal age.

Recurrent implantation failure (RIF) is defined after three or more good quality embryo transfers following in vitro fertilisation without a successful pregnancy outcome. Many factors contribute to RIF however, the endometrial contribution remains unclear. Previous work by our group has identified a micro-RNA (miRNA) miR-185-5p as conserved across placental mammal irrespective of implantation strategies. We tested the hypothesis that miR-185-5p and the pathways it regulates, may be disrupted in the endometria of women with RIF. A human endometrial epithelial cells line (Ishikawa cells) was transfected with mimics or inhibitors for miR-185-5p for 24 (for implantation assay) or 48 hr (for proteomic analysis) along with non-targeting controls. There was a significant different in percentage attachment of BeWoW spheroids to cells transfected with miR-185-5p mimic compared to inhibitor (P<0.05). Transfection of epithelial cells with miR-185-5p altered expression of 1450 (mimic alone) and 509 (inhibitor alone) proteins respective of which, 146 were modified by both. Comparison of predicted targets of miR-185-5p, proteins modified by this study, and key endometrial genes from the literature determined genes and proteins associated with CNP family were further investigated in biopsies from individuals with (n=10) and without RIF (n=9) with CSNK1D expression significantly lower (p<0.05) in individuals with RIF. Collectively these data demonstrate that miR-185-5p modifies pathways that are important for successful implantation in humans.

In each menstrual cycle, progesterone acting on estrogen-primed endometrium elicits an inflammatory decidual reaction, rendering it poised for embryo implantation and transformation into the decidua of pregnancy. Here, we show that the sequential functions of the decidual reaction - implantation and decidualization - pivot on the time-sensitive loss of progesterone-resistant stromal cells that form a transient implantation niche and reciprocal expansion of progesterone-responsive pre-decidual cells. In parallel, proliferation and differentiation increase the abundance of immunotolerant uterine natural killer (uNK) cells. Examination of pre-pregnancy endometrial biopsies from 924 women revealed that the frequency of cycles culminating in a blunted or stalled decidual reaction closely aligns with the age-independent recurrence risk of miscarriage. Further, analysis of 632 biopsies obtained in different cycles from 316 women indicated that prior miscarriages disrupt intercycle endometrial homeostasis, an observation supported by modelling the impacts of prolonged decidual inflammation in three-dimensional endometrial assembloids. Although stalling of the decidual reaction is often accompanied by a poor expansion of immunotolerant uNK cells, miscarriages do not impact intercycle uNK cell dynamics. Our findings indicate that intrinsic uterine mechanisms hardwire the recurrence risk of miscarriage, underscoring the need for pre-pregnancy diagnostics and therapeutics. One Sentence Summary: The frequency of menstrual cycles culminating in a suboptimal decidual reaction determines the recurrence risk of miscarriage.

Understanding how glucose regulates endometrial function will support preconception guidance and/or the development of targeted interventions for individuals living with diabetes wishing to embark on pregnancy. We found that glucose can influence endometrial epithelial cell receptivity to embryo implantation by regulating posttranslational modification of proteins involved in the maintenance of cell polarity. Impaired or inappropriate endometrial receptivity could contribute to fertility and/or early pregnancy complications caused by poor glucose control.

The human endometrium is a dynamic entity that plays a pivotal role in mediating the complex interplay between the mother and developing embryo. Endometrial disruption can lead to pregnancy loss, impacting both maternal physical and psychological health. Recent research suggests that the endometrial microbiota may play a role in this, although the exact mechanisms are still being explored, aided by recent technological advancements and our growing understanding of host immune responses. Suboptimal or dysbiotic vaginal microbiota, characterized by increased microbial diversity and reduced Lactobacillus dominance, has been associated with various adverse reproductive events, including miscarriage. However, the mechanisms linking the lower reproductive tract microbiota with pregnancy loss remain unclear. Recent observational studies implicate a potential microbial continuum between the vaginal and endometrial niche in patients with pregnancy loss; however, transcervical sampling of the low biomass endometrium is highly prone to cross-contamination, which is often not controlled for. In this review, we explore emerging evidence supporting the theory that a dysbiotic endometrial microbiota may modulate key inflammatory pathways required for successful embryo implantation and pregnancy development. We also highlight that a greater understanding of the endometrial microbiota, its relationship with the local endometrial microenvironment, and potential interventions remain a focus for future research.