Uterine epithelial cell proliferation and endometrial hyperplasia: evidence from a mouse model

Abstract

In the uterus, epithelial cell proliferation changes during the estrous cycle and pregnancy. Uncontrolled epithelial cell proliferation results in implantation failure and/or cancer development. Transforming growth factor-β (TGF-β) signaling is a fundamental regulator of diverse biological processes and is indispensable for multiple reproductive functions. However, the in vivo role of TGF-β signaling in uterine epithelial cells remains poorly defined. We have shown that in the uterus, conditional deletion of the Type 1 receptor for TGF-β (Tgfbr1) using anti-Müllerian hormone receptor type 2 (Amhr2) Cre leads to myometrial defects. Here, we describe enhanced epithelial cell proliferation by immunostaining of Ki67 in the uteri of these mice. The aberration culminated in endometrial hyperplasia in aged females. To exclude the potential influence of ovarian steroid hormones, the proliferative status of uterine epithelial cells was assessed following ovariectomy. Increased uterine epithelial cell proliferation was also revealed in ovariectomized Tgfbr1 Amhr2-Cre conditional knockout mice. We further demonstrated that transcript levels for fibroblast growth factor 10 (Fgf10) were markedly up-regulated in Tgfbr1 Amhr2-Cre conditional knockout uteri. Consistently, treatment of primary uterine stromal cells with TGF-β1 significantly reduced Fgf10 mRNA expression. Thus, our findings suggest a potential involvement of TGFBR1-mediated signaling in the regulation of uterine epithelial cell proliferation, and provide genetic evidence supporting the role of uterine epithelial cell proliferation in the pathogenesis of endometrial hyperplasia.

Keywords: endometrial hyperplasia; epithelial cell; proliferation; transforming growth factor β; uterus.

Figure 1

TGFBR1 localization in mouse uterus during estrous cycle. (A) Diestrus; (B) pro-estrus; (C) estrus and (D) metestrus. The uteri from 2- to 3-month-old mice containing a Tgfbr1^lacZ allele were processed for X-gal staining. The sections were counterstained with fast red to visualize the nuclei. Note the predominant X-gal staining in the smooth muscle layers and variable staining in the stromal compartment. SM, smooth muscle; LE, luminal epithelium; GE, glandular epithelium; St, stroma. Scale bar = 50 µm.

Figure 2

Uterine epithelial cell proliferation in control and Tgfbr1 Amhr2-Cre cKO uteri. Immunofluorescence of Ki67 (A; red) and Ki67 and ACTA2 (green) (B and C) using uterine sections from wild-type control (Ctrl) mice. Immunostaining of Ki67 (D and G; red) and Ki67 and ACTA2 (green) (E, F, H and I) using uterine sections from Tgfbr1 Amhr2-Cre cKO mice. Uterine samples were collected from adult mice (∼3 months of age) at the diestrous stage. Note the extensive Ki67 immunostaining in epithelial cells of Tgfbr1 Amhr2-Cre cKO mice compared with the controls. LE, luminal epithelium; GE, glandular epithelium; St, stroma; SM, smooth muscle. Scale bar (50 μm) is representatively shown in (A).

Figure 3

Increased epithelial cell proliferation in ovariectomized Tgfbr1 Amhr2-Cre cKO uteri. (A–F) Immunofluorescence of Ki67 (red) using uterine sections from ovariectomized control and Tgfbr1 Amhr2-Cre cKO mice. Uterine epithelia were labeled with anti-KRT8 antibody (green). Ovariectomy was performed on adult female mice from control and Tgfbr1 Amhr2-Cre cKO groups (∼3 month of age). Three independent samples were examined from each group and representative images were depicted. LE, luminal epithelium; GE, glandular epithelium; St, stroma. (G and H) Negative controls where primary antibodies were replaced with rabbit IgG. Scale bar (50 μm) is representatively shown in (A). (I) Quantification of Ki67-positive cells in glandular and luminal epithelia of control and Tgfbr1 Amhr2-Cre cKO uteri. Results are presented as the percentage of controls which is set to 100% (n = 3 per group). Data are mean ± SEM. *P < 0.05 versus corresponding controls.

Figure 4

Glandular defects observed in aged Tgfbr1 Amhr2-Cre cKO mice. (A and B) Tgfbr1 Amhr2-Cre cKO mice develop cystic glands (B; red arrows) compared with age-matched controls (A). (C–F) Loss of intervening stroma among uterine glands (red asterisks) in Tgfbr1 Amhr2-Cre cKO mice (E and F) in contrast to controls (C and D). Uterine samples were from 8-month- old random cycling control and Tgfbr1 Amhr2-Cre cKO mice (n = 3 per group). Anti-KRT8 (A and B) and anti-VIM antibodies (C–F) were used to label uterine epithelial cells and mesenchymal cells, respectively. (D) and (F) are higher power images for (C) and (E), respectively. LE, luminal epithelium; GE, glandular epithelium; St, stroma; Myo, myometrium. Scale bar = 20 µm (D and F) and 200 µm (A, B, C and E).

Figure 5

Epithelial cell lesions in Tgfbr1 Amhr2-Cre cKO mice. (A–D) Immunofluorescence staining of uterine sections using antibodies directed against KRT8 (green) and Ki67 (red) in aged control and Tgfbr1 Amhr2-Cre cKO mice. (E and F) Representative control uteri stained with KRT8 and Ki67. Scale bar = 25 µm (A–F). (G–T) Immunofluorescence of KRT8 (green) and FOXA2 (red; G and H) or vimentin (red; I–T) in the uterus of aged Tgfbr1 Amhr2-Cre cKO mice. Six 9- to 10.5-month-old Tgfbr1 Amhr2-Cre cKO mice were analyzed using immunofluorescence or immunohistochemistry. Epithelial cell defects were observed in all mice, and localization of vimentin in epithelial cells was observed in three mice. Representative images from a 10.5-month Tgfbr1 Amhr2-Cre cKO uterus are presented. (L–N) and (R–T) are higher power images for selected regions of (I–K) and (O–Q), respectively. Arrows (L–N) indicate localization of vimentin to KRT8-positive epithelial cells. UG, uterine gland; St, stroma; CG, cystic endometrial gland; GE, glandular epithelium. Scale bar is representatively shown in (G) and equals 10 µm (G and H, L–N and R–T) and 20 µm (I–K and O–Q).

Figure 6

Alteration of FOXA2 expression in endometrial glands of Tgfbr1 Amhr2-Cre cKO mice. (A–I) Immunofluorescence staining of KRT8 (green) and FOXA2 (red) in control (A–C; 8-month-old) and Tgfbr1 Amhr2-Cre cKO mice at the age of 8 (D–F) and 10.5 (G–I) months. Yellow arrows indicate glandular epithelium. Dotted lines in (D–I) outline cystic dilated glands. Note the absence of FOXA2 in the cystic glands (D–I) and endometrial glands marked with asterisks (F and I). GE, glandular epithelium; LE, luminal epithelium; St, stroma; CG, cystic endometrial gland. Scale bar (50 μm) is representatively shown in (A).

Figure 7

Elevated Fgf10 mRNA levels in the uterus of Tgfbr1 Amhr2-Cre cKO mice. (A) RT–PCR amplification of Fgf10 in uterine stromal cells and smooth muscle cells. Note that Fgf10 was readily detectable in uterine stromal cells. No target band was detected in the negative controls. Hprt was included as an internal control. SC, stromal cells; SMC, smooth muscle cells; Neg, negative control without reverse transcriptase. (B) Fgf10 mRNA expression was increased in ovariectomized Tgfbr1 Amhr2-Cre cKO uteri among the examined Fgfs compared with corresponding controls. n = 5 per group. OVX, ovariectomy. (C) Fgf10 mRNA abundance was reduced in uterine stromal cells treated with TGF-β1 (0.1–10 ng/ml) for 20 h versus controls. Three independent cell culture experiments were performed. Data are mean ± SEM. *P < 0.05.