Interaction of human trophoblast cells with gland-like endometrial spheroids: a model system for trophoblast invasion.

STUDY QUESTION: Do maternal endometrial epithelial cell (EEC) differentiation and polarity impact the invasive capacity of extravillous trophoblast (EVT) cells during early human implantation? SUMMARY ANSWER: In a three dimensional (3D) confrontation co-culture the invasiveness of the human trophoblast cell line AC-1M88 was inversely correlated with the degree of differentiation and polarization of human endometrial adenocarcinoma cell spheroids. WHAT IS KNOWN ALREADY: In a previous study desmosomal and adherens junction proteins were shown to spread from a subapically restricted lateral position to the entire lateral membrane in human glandular EECs during the implantation window of the menstrual cycle. Whether this change in EEC junction localization has an impact on the interaction of EVT cells with glandular EECs during early human implantation is not known. STUDY DESIGN, SIZE, DURATION: A new 3D cell culture system was developed in order to mimic early implantation events in humans. As a model for the invasion of endometrial glands by EVT cells, spheroids of three differently differentiated and polarized endometrial adenocarcinoma cell lines were confronted with an EVT cell line in co-culture experiments. PARTICIPANTS/MATERIALS, SETTING,
METHODS: Three human adenocarcinoma EEC lines were chosen for this study because of their differences in differentiation and polarization: HEC-1-A, which is well differentiated and highly polarized, Ishikawa, which is well differentiated and moderately polarized, and RL95-2, which is moderately differentiated and poorly polarized. When the cell lines were grown in reconstituted basement membrane, they formed gland-like, multicellular spheroids. The degree of polarization within the different EEC spheroids was assessed by 3D confocal immunofluorescence microscopy detecting the basal membrane protein integrin α6, the apical tight junction-associated protein ZO-1 and the desmosomal plaque protein desmoplakin 1/2 (Dsp). Cells of the human EVT cell line AC-1M88, which is a fusion cell line of primary EVT cells and choriocarcinoma-derived JEG-3 cells, were added to the different EEC spheroids to examine their interaction. For the analyses of trophoblast-endometrial confrontation sites, HLA-G was used as a specific EVT cell marker. MAIN RESULTS AND THE ROLE OF CHANCE: The endometrial HEC-1-A and Ishikawa cells formed gland-like structures in reconstituted basement membrane with apicobasal polarization towards their well-developed internal lumina, while most of the RL95-2 spheroids showed no lumen formation at all. The three EEC lines strongly differed in their apicobasal distribution pattern of Dsp. Ishikawa and HEC-1-A spheroids showed a subapical concentration of Dsp. In contrast, an equal distribution of Dsp was discerned along the entire lateral membranes in RL95-2 spheroids. In 3D confrontation co-cultures the highest invasiveness of AC-1M88 was observed in the poorly polarized RL95-2 spheroids. LIMITATIONS, REASONS FOR CAUTION: Human endometrial and trophoblast cell lines were used for this study because of ethical and legal restrictions for implantation studies with human blastocysts and because of limited access to primary human endometrial cells. WIDER IMPLICATIONS OF THE
FINDINGS: The presented 3D cell culture system can be used to investigate the contribution of epithelial junctions to trophoblast-endometrial interactions. The identified impact of endometrial differentiation and polarity on the invasiveness of EVT cells improves our understanding of the relevance of endometrial receptivity for early implantation and may contribute to higher success rates in assisted reproductive technology. STUDY FUNDING/COMPETING INTERESTS: This work was supported by Grant 146/14, 'START-Program', Medical Faculty, RWTH Aachen University, to V.U.B., by Grant Lec_16_12, 'RWTH Lecturer Award', RWTH Aachen University to I.C.-L. and by the German Research Council (Grant LE 566-20-1). The authors declare no conflict of interest.