Dilated thin-walled blood and lymphatic vessels in human endometrium: a potential role for VEGF-D in progestin-induced break-through bleeding

Abstract

Progestins provide safe, effective and cheap options for contraception as well as the treatment of a variety of gynaecological disorders. Episodes of irregular endometrial bleeding or breakthrough bleeding (BTB) are a major unwanted side effect of progestin treatment, such that BTB is the leading cause for discontinued use of an otherwise effective and popular medication. The cellular mechanisms leading to BTB are poorly understood. In this study, we make the novel finding that the large, dilated, thin walled vessels characteristic of human progestin-treated endometrium include both blood and lymphatic vessels. Increased blood and lymphatic vessel diameter are features of VEGF-D action in other tissues and we show by immunolocalisation and Western blotting that stromal cell decidualisation results in a significant increase in VEGF-D protein production, particularly of the proteolytically processed 21 kD form. Using a NOD/scid mouse model with xenografted human endometrium we were able to show that progestin treatment causes decidualisation, VEGF-D production and endometrial vessel dilation. Our results lead to a novel hypothesis to explain BTB, with stromal cell decidualisation rather than progestin treatment per se being the proposed causative event, and VEGF-D being the proposed effector agent.

Figure 1. Progestin-induced endometrial blood and lymph vessel dilatation.

Enlarged blood and lymphatic vessels are present in the endometrium of women treated with LNG-IUS prior to hysterectomy for heavy menstrual bleeding. (A–B) Untreated control samples. (C–D) LNG-IUS treated samples. Sections have been immunostained with CD31 (A, C), which labels both blood and lymphatic endothelial cells, or D2-40 (B, D), which labels lymphatic endothelial cells only. Black arrows: blood vessels.

Figure 2. Increased blood and lymphatic vessel area after progestin exposure.

Blood and lymphatic vessel cross sectional areas are increased in endometrium from women treated with LNG-IUS prior to hysterectomy for heavy menstrual bleeding. (A) The mean area of the five largest blood vessels in the functional layer and the basal layer in untreated control (white bars) and LNG-IUS treated (grey bars) endometrium. (B) The mean area of the five largest lymphatic vessels in the functional layer and the basal layer of untreated control (white bars) and LNG-IUS treated (grey bars) endometrium. Columns represent means ± SE. *, P<0.05.

Figure 3. Endometrial blood and lymphatic vessel density after progestin exposure.

Blood and lymphatic vessel densities do not change in endometrium from women treated with LNG-IUS prior to hysterectomy for heavy menstrual bleeding. (A) Vascular density of blood vessels in untreated control (white bars) and LNG-IUS treated (grey bars) samples of the functional layer and the basal layer. (B) Lymphatic vessel density in untreated and LNG-IUS treated endometrial samples. Columns represent means ± SE. *, P<0.05.

Figure 4. VEGF-D immunostaining in endometrial decidual cells.

Distinct VEGF-D immunostaining (brown) is present in pre-decidual cells of endometrium from women treated with LNG-IUS prior to hysterectomy. (A) Micrograph illustrating difference in VEGF-D immunostaining (brown) intensity in stromal pre-decidualised cells (pd) compared to non-decidualised cells (nd). (B) Higher power image of pre-decidualised stromal cells showing VEGF-D immunostaining in brown. (C) Higher power image of non-decidualised stromal cells showing reduced VEGF-D immunostaining. (D) Isotype matched negative control. Black arrow: blood vessel. (sections are lightly counterstained with haematoxylin to identify cell nuclei in blue).

Figure 5. VEGF-D protein in endometrial decidual cells.

Western analysis of vascular endothelial cell growth factor-C (VEGF-C) and VEGF-D peptide expression in primary cultures of human endometrial stromal cells (HESC). Panels A and B show representative data from 3 samples. Panels C and D show combined densitometry data from all 5 samples. (A) VEGF-C peptide expression (58, 41, 29/31 and 21 kD) in HESC (ND) and decidualised HESC (Dec). (B) VEGF-D peptide expression (53, 41, 29/31 and 21 kD) in HESC and decidualised HESC. (C–D) Densitometry of VEGF-C and VEGF-D peptide expression. White bars representing control non-decidualised HESC and black bars representing decidualised HESC. Columns represent means ± SE; * P<0.05.

Figure 6. Human endometrial xenografts in NOD/scid mice.

Representative micrographs of human endometrial xenografts taken from NOD/scid mice after 6 weeks growth at a subcutaneous site. (A) A mixture of human and mouse vessels at the edge (capsule) of the xenograft shown by double immunostaining with anti-human CD31 (brown, black arrow) and anti-mouse CD31 (blue, white arrow). Note the vessel of mixed human and mouse origin (grey arrow). (B) The central portion of the xenograft populated almost exclusively with vessels of human origin (double immunostaining with anti-human CD31 in brown and anti-mouse CD31 in blue). (C) Routine haematoxylin and eosin stained xenografts treated with estradiol valerate developed densely packed fibroblast-like stromal cells and cuboidal to columnar epithelial cells. (D) Routine haematoxylin and eosin stained xenografts grafts treated with MPA displayed pre-decidualised stromal cells with flattened epithelial cells. ep: epithelium, g: glands, m: myometrium, pd: pre-decidual stroma, s: stroma.

Figure 7. VEGF-D immunostaining of xenograft decidual cells.

Representative micrographs of human endometrial xenografts from NOD/scid mice treated with estradiol valerate (A,B) or medroxyprogesterone acetate (C,D). Serial sections were stained with haematoxylin and eosin (A,C) or immunostained with VEGF-D (B,D). Inset in (B): Isotype matched negative control. Note the pre-decidual cells in MPA-treated sections. ep: epithelium, g: glands, pd: pre-decidual stroma, s: stroma.

Figure 8. Progestin effects on xenograft vessel density and area.

Blood vessel density decreases and lymphatic vessel cross sectional vessel area increases in human endometrial xenografts treated with medroxyprogesterone acetate (MPA). (A) Blood vessel density and (B) lymphatic vessel density of untreated (E only) and treated (MPA) xenografts. (C) Blood vessel and (D) lymphatic vessel cross sectional area of E only and MPA treated xenografts. Bars represent means ± SE. * P<0.05.