Generation of epithelial-stromal assembloids as an advanced in vitro model of impaired adenomyosis-related endometrial receptivity.

Objective: To create a novel, more advanced in vitro model of human endometrium, using so-called assembloids, looking to explore endometrial receptivity in adenomyosis.

Design: Evaluation of assembloid responsiveness to hormonal stimulation by immunohistochemistry, enzyme-linked immunosorbent assay, and scanning electron microscopy.

Setting: University-based research unit in gynecology.

Patient(s): Twelve women, six of whom were affected by adenomyosis.

Intervention(s): Organoids (in the form of glandular fragments) and stromal fibroblasts were collected from endometrial biopsies. The two populations were combined inside an extracellular matrix to create 3D assembloids, which were then exposed to hormonal stimulation (β-estradiol for 48 hours, followed by β-estradiol/progesterone/cyclic adenosine monophosphate for 72 hours) to mimic the window of implantation.

Main Outcome Measure(s): Glycodelin, leukemia inhibitory factor (LIF), and homeobox A10 (HOXA10) expression, prolactin secretion, and pinopode development.

Result(s): Endometrial organoids and stromal cells were successfully isolated from women with and without adenomyosis and combined to generate the assembloid model. On stimulation, assembloids from both groups acquired a more secretory phase-like phenotype, as demonstrated by histology, and were shown to be positive for glycodelin, LIF, and HOXA10 by immunohistochemistry. Adenomyotic assembloids expressed significantly lower levels of LIF and HOXA10 within the stromal compartment after stimulation than did healthy assembloids in the same condition. Enzyme-linked immunosorbent assay revealed prolactin secretion in vitro, showing an upward trend in hormonally treated assembloids from both healthy and affected women. By scanning electron microscopy, fully formed pinopodes were discerned on the epithelial surface of healthy assembloids after stimulation, but they were absent in case of adenomyosis.

Conclusion(s): Primary assembloids can be generated from endometrial biopsies from both healthy subjects and women affected by adenomyosis. These assembloids are amenable to hormonal stimulation and mimic secretory phase-specific characteristics of endometrial tissue in vivo, including glycodelin, LIF, and HOXA10 expression, and pinopode formation. Assembloids from adenomyosis appear to be less sensitive to hormonal treatment, showing reduced expression of LIF and HOXA10 in the stromal compartment and failing to form pinopodes. All in all, endometrial assembloids may serve as an advanced preclinical model of adenomyosis-related impaired endometrial receptivity, opening up new horizons in understanding and treating the condition.