Lab-Grown Endometrium: New 3D Models Bring Human Reproduction Into Focus

For decades, the human endometrium—the dynamic lining of the uterus that governs embryo implantation, menstruation and the earliest moments of pregnancy—has remained one of the most difficult tissues in the human body to study. Now, a comprehensive review published in Nature Biomedical Engineering maps how a quiet revolution in tissue modelling, spanning organoids, assembloids and microfluidic ‘endometrium-on-a-chip’ platforms, is finally giving researchers access to living, hormone-responsive replicas of this remarkable tissue. The work, led by Adriana N. Vélez-Avilés and Ashley Abel of Yale School of Medicine together with Hugh S. Taylor and senior author Berna Sozen, traces the field’s evolution from flat, reductionist cell cultures to sophisticated three-dimensional systems that capture the cellular choreography of human reproduction.

The stakes could hardly be higher. The endometrium sits at the centre of embryo implantation, pregnancy maintenance and broader reproductive and systemic health, yet conditions affecting it—endometriosis, adenomyosis, recurrent implantation failure, recurrent pregnancy loss and chronic endometritis—remain poorly understood and notoriously difficult to treat. The authors argue that the root of this knowledge gap lies in the tools the field has historically depended upon: animal models and two-dimensional cultures that systematically fail to reproduce the human endometrium’s intrinsic biology.

The problem with animal models begins with basic evolutionary divergence. Mice, the workhorse of biomedical research, do not menstruate, and their endometrial architecture, hormone responsiveness and decidualization programmes—divided into the stem-cell-rich basalis that regenerates the tissue each cycle and the functionally active functionalis that is shed during menstruation—differ in fundamental ways from those of humans. Even the spiny mouse, the only rodent known to menstruate, offers only a partial bridge. Baboons show spontaneous endometriosis, making them useful but expensive and ethically constrained. The review details how comparative studies between human and rat endometrial co-cultures have repeatedly exposed these species-specific gaps, underscoring why findings from mice frequently fail to translate into human clinical practice.

Two-dimensional cell culture, meanwhile, has its own structural limitations. Landmark studies dating back to the early twentieth century established the foundations of tissue culture, and endometrial research made enormous strides with immortalized cell lines such as Ishikawa and HEC-1 adenocarcinoma-derived cells, alongside primary cultures of endometrial epithelial glands and stromal cells isolated as early as the 1970s. These systems enabled researchers to dissect steroid metabolism, integrin expression, 17 beta-hydroxysteroid dehydrogenase regulation and paracrine signalling between epithelium and stroma. Yet flat monolayers strip away the three-dimensional geometry, cell polarity, extracellular matrix interactions and multicellular composition that define how the endometrium actually behaves—particularly during the precisely timed window of implantation, when epithelial polarity shifts and stromal cells undergo decidualization under progesterone control.

The inflection point arrived in 2017, when two independent teams grew long-term, hormone-responsive organoids from human endometrium in chemically defined media. These miniature, self-organizing structures, built from endometrial epithelial stem and progenitor cells, recapitulated key features of endometrial physiology, expanded over successive passages and responded to oestrogen and progesterone much as the native tissue does. Suddenly, researchers could grow living endometrial epithelium indefinitely, derive it from patients with disease, and interrogate its biology with genetic and pharmacological tools. Follow-up work showed that patient-derived organoids could capture clinical heterogeneity in endometrial disease and were amenable to drug screening—opening a genuine pathway toward personalized medicine in gynaecology.

The next generation of models has pushed further toward physiological realism. Multi-lineage assembloids now combine endometrial epithelium with stromal fibroblasts, immune components, endothelial cells and extracellular matrix in engineered assemblies that mimic the tissue’s natural architecture. Recent protocols for generating mouse and human endometrial assembloids allow epithelial–stromal crosstalk to be studied within a single three-dimensional construct, while air–liquid interface methods have yielded assembloids possessing a luminal epithelium—the very surface a blastocyst must breach during implantation. Studies using assembloids have already modelled how decidual senescence impairs embryo implantation and how adenomyosis-related endometrial receptivity is compromised, delivering mechanistic insights that flat cultures simply could not provide.

Bioengineering has multiplied these capabilities. Microfluidic endometrium-on-a-chip devices compartmentalize perivascular stroma and endothelial cells to create vascularized tissue architectures, and experiments within them have shown that hemodynamic forces enhance decidualization via endothelial-derived prostaglandin E2 and prostacyclin—demonstrating that mechanical cues from blood flow, absent in static cultures, are genuine regulators of endometrial function. Fully synthetic hydrogels have replaced animal-derived Matrigel in several systems, enabling precise control of matrix composition and permitting organoid co-cultures of epithelium and stroma to be studied in defined extracellular environments. Bioprinted, hormone-responsive bilayer models now reproduce the tissue’s layered structure, and engineered platforms incorporating a 28-day hormonal cycle have simulated the human menstrual cycle in vitro, capturing epithelial cell transitions during menstruation and regeneration.

Perhaps the most consequential frontier is implantation itself. Because direct observation of human embryo implantation is ethically impossible, the field has had to infer its mechanics from indirect evidence. That barrier is now falling. Stem-cell-derived blastoids and trophoblast organoids can be co-cultured with engineered endometrial models to recreate the first contact between embryo and maternal tissue. Recent studies have modelled the embryo–endometrial interface in three dimensions, produced human receptive endometrial assembloids designed to decode the implantation window, and developed microfluidic platforms that measure embryo adhesion in real time. The review highlights 2026 studies reporting three-dimensional post-implantation co-culture of human embryo and endometrium models, alongside paired investigations of implantation and implantation failure—work that together promises to reveal the molecular dialogue between embryo and mother at unprecedented resolution. Single-cell atlases of the endometrium and the maternal–fetal interface, including a recently published spatiotemporal dissection of the human maternal–fetal interface, are providing the reference maps against which these engineered systems can be validated.

The clinical implications radiate across women’s health. Endometriosis, a chronic systemic disease affecting an estimated one in ten women of reproductive age, has seen its cellular basis illuminated by single-cell transcriptomic studies of eutopic and ectopic tissue, multi-ancestry genome-wide association analyses and investigations of microRNA signatures in tissue, serum and extracellular vesicles—several of which show promise as non-invasive diagnostic biomarkers. Patient-derived organoids from endometriosis lesions and early peritoneal endometriosis models now allow drug testing on a patient’s own disease cells. Endometrium-on-chip platforms built from patient cells are being evaluated for assessing endometrial receptivity and guiding personalized fertility treatment, while recent organoid work has even traced the donor-derived cellular origin of endometrium after uterus transplantation. Insulin- and glucose-induced alterations in endometrial transcriptomes observed on-chip hint at mechanisms linking metabolic health to fertility, and scaffold-free organoids have been shown to respond to the excess androgens characteristic of polycystic ovarian syndrome.

The authors are candid that substantial challenges remain. Matrigel, the tumor-derived basement membrane matrix underpinning many organoid protocols, is chemically undefined and variable, motivating ongoing efforts to develop fully synthetic alternatives. Vascularization, immune cell incorporation and the biophysical environment of the uterus—contractions, blood flow, cyclic mechanical remodeling—are only partially captured by existing systems. Ethical governance is advancing alongside the science: revised international stem cell guidelines now specifically address stem-cell-based embryo models, and frameworks for embedded ethics and dynamic consent are being proposed for research that grows ever closer to reconstructing human development in a dish. Competition-of-interest disclosures and peer oversight remain part of the field’s infrastructure as its translational potential grows.

What emerges from this sweeping synthesis is a portrait of a field in the midst of a genuine paradigm shift. By uniting developmental biology, stem-cell science, tissue engineering and single-cell genomics, researchers are converging on engineered endometria that are hormone-responsive, multicellular, vascularized and patient-specific. Such systems could transform how infertility is diagnosed, how endometriosis is treated and how the safety of new therapeutics is tested—without recourse to animal models that so often mislead. More profoundly, they are rewriting what it means to understand human reproduction: replacing inference from mice with direct observation in human-like tissue, and giving clinicians, for the first time, a living laboratory in which the earliest events of human life can be watched, perturbed and, ultimately, protected.

Beyond implantation, the cyclical nature of the endometrium itself presents a modelling challenge that these new platforms are beginning to address. Unlike most human tissues, the endometrium undergoes scarless repair after each menstrual shedding, a process driven by adult stem and progenitor cells residing in the basalis. Organoid systems that sustain long-term expansion now allow this regenerative capacity to be examined directly, complementing single-cell reference atlases that have catalogued the tissue’s cellular diversity across the menstrual cycle.

The immune dimension is equally critical. Uterine natural killer cells and other maternal immune populations orchestrate fetal–maternal tolerance and guide the differentiation of invading trophoblast, yet most current models lack immune components entirely. Integrating these lineages into assembloids and chip platforms remains an active area of development, informed by single-cell reconstructions of the maternal–fetal interface.

For patients, the promise is tangible: organoids derived from diseased tissue preserve the molecular signatures of the individual they came from, enabling drug responses to be tested outside the body. As these models mature, they may reduce reliance on both animal experimentation and empirical trial-and-error in fertility clinics, offering a mechanistic bridge between a patient’s cellular biology and her clinical care.

Subject of Research: Bioengineered three-dimensional models of the human endometrium, including organoids, assembloids and microfluidic devices, for studying implantation, endometrial disease and personalized reproductive medicine.

Article Title: Engineering the human endometrium at the intersection of development and reproduction

Article References: Vélez-Avilés, A. N., Abel, A., Taylor, H. S., & Sozen, B. (2026). Engineering the human endometrium at the intersection of development and reproduction. Nature Biomedical Engineering. https://doi.org/10.1038/s41551-026-01789-2

Image Credits: AI Generated

DOI: 10.1038/s41551-026-01789-2

Keywords: human endometrium, organoids, assembloids, endometrium-on-a-chip, embryo implantation, endometriosis, tissue engineering, decidualization, stem cells, women's health, single-cell transcriptomics, personalized medicine