Cells expressing PAX8 are the main source of homeostatic regeneration of adult mouse endometrial epithelium and give rise to serous endometrial carcinoma

Abstract

Humans and mice have cyclical regeneration of the endometrial epithelium. It is expected that such regeneration is ensured by tissue stem cells, but their location and hierarchy remain debatable. A number of recent studies have suggested the presence of stem cells in the mouse endometrial epithelium. At the same time, it has been reported that this tissue can be regenerated by stem cells of stromal/mesenchymal or bone marrow cell origin. Here, we describe a single-cell transcriptomic atlas of the main cell types of the mouse uterus and epithelial subset transcriptome and evaluate the contribution of epithelial cells expressing the transcription factor PAX8 to the homeostatic regeneration and malignant transformation of adult endometrial epithelium. According to lineage tracing, PAX8⁺ epithelial cells are responsible for long-term maintenance of both luminal and glandular epithelium. Furthermore, multicolor tracing shows that individual glands and contiguous areas of luminal epithelium are formed by clonal cell expansion. Inactivation of the tumor suppressor genes Trp53 and Rb1 in PAX8⁺ cells, but not in FOXJ1⁺ cells, leads to the formation of neoplasms with features of serous endometrial carcinoma, one of the most aggressive types of human endometrial malignancies. Taken together, our results show that the progeny of single PAX8⁺ cells represents the main source of regeneration of the adult endometrial epithelium. They also provide direct experimental genetic evidence for the key roles of the P53 and RB pathways in the pathogenesis of serous endometrial carcinoma and suggest that PAX8⁺ cells represent the cell of origin of this neoplasm.

Keywords: Endometrial cancer; Mouse models; Single-cell transcriptome; Stem cells.

Fig. 1.

Single-cell transcriptome analysis of the mouse uterus. (A) UMAP projection of single-cell transcriptomic data from the mouse uterus, highlighting the main cell types. NK, natural killer. (B) UMAP projection of luminal (LE) and glandular (GE) epithelial populations labeled by group as identified by unsupervised SNN clustering. (C) Heatmap representing the top 55 differentially expressed genes across the three subpopulations identified in B and independent from prior labels (for the list of genes, see Table S1). (D,E) Preferential expression of genes in luminal (D) and glandular (E) clusters. Right panels, immunostaining for TROP2, encoded by Tacstd2, and FOXA2. ABC Elite, with Hematoxylin counterstaining. Scale bar: 50 µm (applicable to both images).

Fig. 2.

Characterization of PAX8⁺ cells. (A) Percentage of cells within the MCA with detected Pax8 mRNA expression across uterine cell types. (B) Normalized Pax8 mRNA expression in luminal (LE) and glandular (GE) epithelial cells. (C) Immunohistochemical detection of PAX8 expression in both luminal and glandular epithelial cells. ABC Elite, with Hematoxylin counterstaining. Scale bar: 50 µm. (D) Design for lineage tracing of PAX8⁺ cells. Dox, doxycycline; PI, days post-induction after a single doxycycline pulse. (E) Lineage tracing of PAX8⁺ cells in mouse endometrium collected at PI 2, 90 and 300. tdTomato expression, magenta; DAPI counterstaining, blue. Confocal microscopy. Scale bar: 75 µm. Quantification of labeled epithelial cells: PI 2, n=3; PI 90, n=3; PI 300, n=4. In all columns, data are the mean±s.d.; P-values are from Student's two-tailed unpaired t-test: P>0.05.

Fig. 3.

Clonal expansion of PAX8-expressing cells in Pax8-rtTA Tre-Cre Confetti mice. (A) Experimental design. Dox, doxycycline; PI, days post-induction after a single doxycycline pulse. (B) Detection of GFP (green) and RFP (red) at PI 3, 50 and 90. Note that a subset of PAX8⁺ cells has the ability to proliferate, resulting in large monochromatic clones. Counterstaining is with DAPI, blue. Confocal microscopy. Scale bar: 25 µm. (C,D) Quantification of clone size and distribution at PI 3, 50 and 90 in luminal (LE) and glandular (GE) epithelium. PI 3, n=3; PI 50, n=4; PI 300, n=3. All columns, data are the mean±s.d.; P-values are from Student's two-tailed unpaired t-test.

Fig. 4.

Mouse model of SEC. (A) Survival of Pax8-rtTA TRE-Cre Trp53^loxP/loxP Rb1^loxP/loxP Ai9 mice that were treated with doxycycline (Dox, n=21) or vehicle (n=10); log-rank P=0.0004. (B-D) Cross-section of the uterus (B), with areas of neoplastic endometrial epithelium with marked cytological atypia and glandular pattern (C, arrow), and invasion into the myometrium and perimetrium (D, arrow). Arrow and arrowhead in B indicate the location of C and D, respectively. (E) tdTomato expression in the endometrial epithelial neoplasm developed 273 days after intraperitoneal doxycycline administration. (F,G) Dysplastic epithelial lesions (arrows) in luminal (F) and glandular (G) endometrial epithelium. (B-D,F,G) Hematoxylin and Eosin. (E) ABC Elite method, with Hematoxylin counterstaining. Scale bars: 600 µm in B; 100 µm in E (applicable to C and D); 50 µm in G (applicable to F).

Fig. 5.

Comparative immunophenotyping of SEC. (A) Human SEC. Glandular pattern and marked cytological atypia (inset, arrow); intense, diffuse staining of p53, Ki67⁺ cells, lack of expression of estrogen receptor (ER); strong, diffuse staining of p16 and lack of progesterone receptor (PR) expression. Arrowheads, normal cells with expression of ER and PR near the tumor. HE, Hematoxylin and Eosin; other images, ABC Elite method, with Hematoxylin counterstaining. (B) Detection of the proliferation markers Ki67, p16, ER and PR in mouse endometrial tumor (top row) and normal tissue (bottom row). Arrowheads, Ki67⁺ cell and cytoplasmic staining of p16. ABC Elite method, with Hematoxylin counterstaining. Scale bars: 200 µm in A; 60 µm in B.

Fig. 6.

FOXJ1 expression. (A) Normalized Foxj1 mRNA expression in epithelial cell subsets within the MCA. FoxJ1 is detected in some endometrial glandular (GE) but not luminal (LE) epithelial cells. (B,C) tdTomato expression in GE (B) and uterine tube tubal epithelium (C) of FoxJ1^CreERT2::GFP Ai9 mouse at 1 day after intraperitoneal tamoxifen administration in the diestrous phase of the estrous cycle. Counterstaining with DAPI, blue. Scale bar: 50 μm.