Inhibition of erythropoietin‐producing hepatoma receptor B4 (EphB4) signalling suppresses the vascularisation and growth of endometriotic lesions

Abstract

Background and Purpose

The development of endometriotic lesions is crucially dependent on the formation of new blood vessels. In the present study, we analysed whether this process is regulated by erythropoietin‐producing hepatoma receptor B4 (EphB4) signalling.

Experimental Approach

We first assessed the anti‐angiogenic action of the EphB4 inhibitor NVP‐BHG712 in different in vitro angiogenesis assays. Then, endometriotic lesions were surgically induced in the dorsal skinfold chamber and peritoneal cavity of NVP‐BHG712‐ or vehicle‐treated BALB/c mice. This allowed to study the effect of EphB4 inhibition on their vascularisation and growth by means of intravital fluorescence microscopy, high‐resolution ultrasound imaging, histology and immunohistochemistry.

Key Results

Non‐cytotoxic doses of NVP‐BHG712 suppressed the migration, tube formation and sprouting activity of both human dermal microvascular endothelial cells (HDMEC) and mouse aortic rings. Accordingly, we also detected a lower blood vessel density in NVP‐BHG712‐treated endometriotic lesions. This was associated with a reduced lesion growth due to a significantly lower number of proliferating stromal cells when compared to vehicle‐treated controls.

Conclusions and Implications

Inhibition of EphB4 signalling suppresses the vascularisation and growth of endometriotic lesions. Hence, EphB4 represents a promising pharmacological target for the treatment of endometriosis.

Abbreviations

EPCs

endothelial progenitor cells

EphB4

erythropoietin‐producing hepatoma receptor B4

FITC

fluorescein isothiocyanate

FMD

functional microvessel density

HDMEC

human dermal microvascular endothelial cells

HE

haematoxylin and eosin

qRT‐PCR

quantitative real‐time PCR

SDF‐1

stromal cell‐derived factor‐1

α‐SMA

α‐smooth muscle actin

What is already known

• Endometriosis is crucially dependent on the formation of new blood vessels.

• EphB4 is overexpressed in endometriotic lesions.

What this study adds

• The selective EphB4 inhibitor NVP‐BHG712 suppresses multiple endothelial cell functions during angiogenesis.

• Inhibition of EphB4 signalling reduces the vascularisation and growth of endometriotic lesions.

What is the clinical significance

• EphB4 represents a promising pharmacological target for the treatment of endometriosis.

1. INTRODUCTION

Endometriosis is a benign, chronically progressing disease, which is associated with pelvic pain, dysmenorrhea and infertility (Giudice, 2010; Culley et al., 2013). It is caused by the presence of endometrium‐like tissue with stromal cells and glandular epithelial cells in the abdominal cavity (Burney & Giudice, 2012). The aetiology of endometriosis still gives rise to controversy. According to the implantation theory, endometriotic lesions develop from retrogradely menstruated and engrafted endometrial tissue (Sampson, 1927). The further progression of these lesions is driven by a variety of different hormones, growth factors and cytokines in the blood and peritoneal fluid (Koninckx, Kennedy, & Barlow, 1999).

During the last years, an increasing number of studies have shown that the survival of endometriotic lesions is crucially dependent on an adequate blood supply (Laschke & Menger, 2018). The formation of new blood vessels within endometriotic lesions is dependent not only on sprouting angiogenesis but also on the inosculation of pre‐existing endometrial microvessels and the recruitment of circulating endothelial progenitor cells (Laschke & Menger, 2018). Therefore, targeting these modes of vascularisation may represent a promising therapeutic strategy for the treatment of endometriosis (Laschke & Menger, 2012).

The establishment of functional microvascular networks within endometriotic lesions is tightly regulated by multiple pro‐ and anti‐angiogenic signalling pathways (Djokovic & Calhaz‐Jorge, 2014). Among these pathways, ephrin/erythropoietin‐producing hepatoma receptor B4 (EPH receptor B4; EphB4) may be particularly important for the development of endometriosis, because EphB4 is overexpressed in ectopic endometrium when compared to eutopic endometrium of women with endometriosis (Yerlikaya et al., 2016). Moreover, EphB4 has been shown to be an upstream regulator of oestrogen receptor‐α (NR3A1; Schmitt, Nguyen, Gupta, & Mayer, 2013). The activation of EphB4 is initiated by its membrane bound ligand ephrinB2 through direct cell–cell communication (Chen, Zhang, & Zhang, 2019) and results in the stimulation of angiogenesis (Gerety & Anderson, 2002; Gerety, Wang, Chen, & Anderson, 1999; Lv et al., 2016). Accordingly, application of the selective EphB4 inhibitor NVP‐BHG712 in an in vivo plug assay has been shown to suppress vascular endothelial cell growth factor (VEGF)‐driven blood vessel formation (Martiny‐Baron et al., 2010). In addition, You et al. (2017) reported that NVP‐BHG712 reduces the microvessel density of implanted endothelial cell spheroids in mice.

Based on these findings, we herein analysed whether disruption of EphB4 signalling inhibits vascularisation and growth of endometriotic lesions. For this purpose, we first tested the anti‐angiogenic action of NVP‐BHG712 in a panel of in vitro angiogenesis assays. Furthermore, we induced endometriotic lesions in the dorsal skinfold chamber and abdominal cavity of NVP‐BHG712‐ and vehicle‐treated mice to study their vascularisation and growth by means of intravital fluorescence microscopy, high‐resolution ultrasound imaging, histology and immunohistochemistry.

2. METHODS

2.1. Materials

Human dermal microvascular endothelial cells (HDMEC; Cat# C‐12210) and ECGM‐MV medium were from PromoCell (Heidelberg, Germany). NVP‐BHG712 (4‐methyl‐3‐[[1‐methyl‐6‐(3‐pyridinyl)‐1H‐pyrazolo[3,4‐d]pyrimidin‐4‐yl]amino]‐N‐[3‐(trifluoromethyl)phenyl]‐benzamide), NMP (N‐methyl‐2‐pyrrolidone), DMSO, 10xM199, fluorescein isothiocyanate (FITC)‐labelled dextran, Hoechst 33342 and streptavidin‐peroxidase were from Sigma‐Aldrich (Taufkirchen, Germany). Water‐soluble tetrazolium (WST)‐1 assays and lactate dehydrogenase (LDH) assays (Cytotoxicity Detection Kit PLUS) were from Roche diagnostics (Mannheim, Germany). Triton X‐100 was from Carl Roth GmbH (Karlsruhe, Germany). Matrigel and polycarbonate membranes were from Corning Costar by VWR (Darmstadt, Germany). Fetal calf serum (FCS), penicillin and streptomycin were from Biochrom (Berlin, Germany). Dade Diff‐Quick was from Dade Diagnostika GmbH (München, Germany). Non‐adherent round‐bottom 96‐well plates were from Greiner Bio‐One GmbH (Frickenhausen, Germany). Methyl cellulose, the Alexa555 antibody (Cat# A‐21434, RRID:AB_2535855) and the Alexa488 antibody (Cat# A‐11008, RRID:AB_143165) were from Thermo Fisher Scientific (Dreieich, Germany). Acidic collagen extract of rat tails was from Serva Electrophoresis GmbH (Heidelberg, Germany). DMEM was from PAN‐Biotech (Aidenbach, Germany). Ketamine was from Serumwerke Bernburg (Bernburg, Germany). Xylazine was from Bayer AG (Leverkusen, Germany). Dorsal skinfold chambers were from Irola Industriekomponenten GmbH & Co. KG (Schonach, Germany). Dermal biopsy punches were from Stiefel Laboratorium GmbH (Offenbach am Main, Germany). Prolene sutures were from Ethicon Products (Norderstedt, Germany). The antibody against CD31 (Cat# DIA‐310, RRID:AB_2631039) was from Dianova (Hamburg, Germany). The antibody against cleaved caspase‐3 (Cat# 9661, RRID:AB_2341188) was from Cell Signaling Technology (Danvers, MA, USA). The antibody against α‐smooth muscle actin (α‐SMA; Cat# ab5694, RRID:AB_2223021), Ki67 (Cat# ab15580, RRID:AB_443209) and its biotinylated secondary antibody (Cat# ab64256, RRID:AB_2661852) as well as 3‐amino‐9‐ethylcarbazole (AEC Substrate System) were from Abcam (Cambridge, UK).

2.2. Cell culture

HDMEC were cultured in ECGM‐MV medium (0.05 ml·ml⁻¹ FCS, 0.004 ml·ml⁻¹ endothelial cell growth supplement, 10 ng·ml⁻¹ recombinant human epidermal growth factor, 90 μg·ml⁻¹ heparin and 1 μg·ml⁻¹ hydrocortisone) at 37°C in a humidified atmosphere and 5% CO₂ for the performance of in vitro assays. A stock solution of 10‐mM NVP‐BHG712 (dissolved in DMSO) was prepared and stored at −20°C. For the experiments, the stock solution was diluted to final concentrations of 1–50 μM in DMSO and cell culture medium, resulting in identical DMSO concentrations of 0.1% DMSO for NVP‐BHG712‐ and vehicle‐treated cells.

2.3. Viability and cytotoxicity assay

To test the viability of HDMEC exposed to different concentrations of NVP‐BHG712, a WST‐1 assay and a lLDH assay were performed. For this purpose, we seeded 1 × 10⁴ HDMEC in 100‐μl culture medium per well in 96‐well plates. The cells were exposed to vehicle (0.1% DMSO, control) and different concentrations of NVP‐BHG712 (1, 2, 5, 10 and 50 μM). For the lactate dehydrogenase assay, additional wells were incubated with 1% Triton X‐100 as cytotoxic high control. After 24 hr, the assays were performed according to the manufacturer's instructions. In brief, WST‐1 reagent or LDH reaction and stop solution were added to each well and incubated for the indicated times. Then the absorption at 450 nm (WST‐1 assay) or 492 nm (LDH assay) with 620 nm as reference was measured by means of a microplate reader (PhoMo; Anthos Mikrosystem GmbH, Krefeld, Germany). Each of the independent experiments (n = 5) was performed in quadruplicate.

2.4. Cell migration assay

To analyse the effect of NVP‐BHG712 on the migration of HDMEC, a transwell migration assay was performed using a 24‐well chemotaxis chamber with polycarbonate membranes (pore size of 8 μm). First, HDMEC were treated with 1‐ or 2‐μM NVP‐BHG712 or vehicle (DMSO; control) for 24 hr. Thereafter, 1 × 10⁵ pretreated HDMEC were transferred into the 24‐well inserts containing 250‐μl FCS‐free EGCM‐MV. The lower wells contained 750‐μl ECGM‐MV with 1% FCS. The chamber was incubated for 24 hr at 37°C in a humidified atmosphere with 5% CO₂. The remaining cells on the top of the transwell membrane were removed using a cotton swab and the migrated cells were fixed with methanol and stained with Dade Diff‐Quick. The migrated cells were counted in 20 regions of interest at 400× magnification under a phase contrast microscope (BZ‐8000; Keyence, Osaka, Japan). Each of the independent experiments (n = 5) was performed in quadruplicate and results are given as cells·mm⁻².

2.5. Tube formation assay

To analyse the effect of NVP‐BHG712 on the ability of HDMEC to form vessel‐like structures, a tube formation assay was performed. For this purpose, 50‐μl growth factor‐reduced Matrigel was transferred into each well of a 96‐well plate. After incubation for 30 min at 37°C, the polymerization of the Matrigel was completed and 1.5 × 10⁴ HDMEC in ECGM‐MV medium were added in each well. The cells were exposed to 1‐ and 2‐μM NVP‐BHG712 or vehicle at 37°C. After 24 hr, the tube formation was observed by means of a phase‐contrast microscope (BZ‐8000; Keyence). The number of meshes, that is, areas completely surrounded by tubes, was assessed by using the Image J software with an angiogenesis analyser plugin (National Institutes of Health [NIH], Bethesda, MD, USA). Each of the independent experiments (n = 5) was performed in sextuplicate and results are given as number of meshes per well.

2.6. Spheroid sprouting assay

The effect of NVP‐BHG712 on the sprouting activity of HDMEC was evaluated by means of a spheroid sprouting assay. For this purpose, 500 cells per well were seeded in non‐adherent round bottom 96‐well plates containing 100‐μl ECGM‐MV medium enriched with 20% (w/v) methyl cellulose. The plates were incubated for 24 hr at 37°C and 5% CO₂ to guarantee the formation of single spheroids. By mixing 8 volumes of acidic collagen extract of rat tails (2 mg·ml⁻¹, 4°C) with 1 volume of 10xM199, the collagen stock solution was adjusted to pH 7.4 with 1 volume of 0.2‐M NaOH. This stock solution was mixed in a 1:2 dilution with ECGM‐MV containing 20% FCS and 0.5% (w/v) methyl cellulose. Subsequently, 300 μl of the collagen mixture was transferred to each well of prewarmed 24‐well plates and allowed to polymerise for 45 min to prevent sedimentation of the spheroids. Additional 300 μl of the collagen mixture containing ~60 spheroids was pipetted on the top of the polymerised collagen gel. Subsequently, 0.4‐ml ECGM‐MV containing 1‐ and 2‐μM NVP‐BHG712 or vehicle (DMSO; control) was added to each well and incubated 24 hr at 37°C and 5% CO₂ in a humidified atmosphere. The average sprout length of 10 spheroids/group was quantitatively analysed in five independent experiments by means of the LAS V4.8 software (Leica DFC450 C; Leica Microsystems, Wetzlar, Germany).

2.7. Ex vivo aortic ring assay

The anti‐angiogenic action of NVP‐BHG712 was further assessed in an ex vivo aortic ring assay as previously described (Gu et al., 2016). A total of 90 aortic rings from five BALB/c mice were used. The rings were embedded in 200‐μl Matrigel in 48‐well tissue culture grade plates. After polymerisation for 20 min at 37°C and 5% CO₂, the wells were overlaid with 800 μl of DMEM (10% FCS, 100 U·ml⁻¹ penicillin, 0.1 mg·ml⁻¹ streptomycin) containing vehicle (0.1% DMSO; n = 5; six replicates), 1‐ or 2‐μM NVP‐BHG712 (each n = 5; six replicates). After incubation for 6 days at 37°C and 5% CO₂ with medium change on Day 3, the vascular sprouting area (mm²) was analysed by phase‐contrast microscopy (BZ‐8000; Keyence).

2.8. Animals

This study was approved by the local governmental animal protection committee (Landesamt für Verbraucherschutz, Saarbrücken, Germany; permission number: 47/2016). All experiments were conducted in accordance with the Directive 2010/63/EU. Animal studies are reported in compliance with the ARRIVE guidelines (Kilkenny et al., 2010; McGrath & Lilley, 2015) and with the recommendations made by the British Journal of Pharmacology.

For the in vivo experiments, female BALB/c mice with an age of 12–16 weeks and a weight of 18–25 g (Institute for Clinical and Experimental Surgery, Saarland University, Homburg/Saar, Germany) served as donor mice for the generation of endometrial grafts and recipient mice for the implantation of dorsal skinfold chambers and the induction of intraperitoneal endometriotic lesions. The recipient mice were either housed one (dorsal skinfold chamber model) or four to six per cage (intraperitoneal model) on wood chips as bedding in the conventional animal facility of the Institute for Clinical and Experimental Surgery. They had free access to tap water and standard pellet food (Altromin, Lage, Germany) and were maintained under a 12‐hr day/night cycle. The cycle stage of the mice was determined by vaginal lavage as described previously (Rudzitis‐Auth, Nenicu, Nickels, Menger, & Laschke, 2016). Only mice in the stage of oestrus and dioestrus were included in the study.

2.9. Dorsal skinfold chamber model of endometriosis

The effect of NVP‐BHG712 on the vascularisation of developing endometriotic lesions was repetitively analysed by means of intravital fluorescence microscopy within the dorsal skinfold chamber model. This model is well established for the analysis of microvascular network formation in endometriotic lesions (Laschke, Elitzsch, Vollmar, & Menger, 2005; Nenicu, Körbel, Gu, Menger, & Laschke, 2014). For this purpose, dorsal skinfold chambers were implanted onto the back of recipient mice in the stage of dioestrus, as described previously in detail (Laschke, Vollmar, & Menger, 2011). After implantation, the animals were allowed to recover from anaesthesia and surgical trauma for 48 hr. In general, the animals tolerated the implantation of the dorsal skinfold chamber well, as indicated by normal feeding, cleaning and sleeping habits. Accordingly, it was not necessary to administer post‐surgical analgesia, which may have also affected the microcirculatory parameters assessed in this model. After the recovery period, donor mice were anaesthetised by an intraperitoneal injection of 75 mg·kg⁻¹ ketamine and 15 mg·kg⁻¹ xylazine. The uterine horns were removed, transferred to a petri dish containing DMEM (10% FCS, 100 U·ml⁻¹ penicillin and 0.1 mg·ml⁻¹ streptomycin) and opened longitudinally. The donor animals were then killed by cervical dislocation. Subsequently, the myometrium of the uterine horns was removed and endometrial fragments with a size of ~0.6 mm² were prepared. These fragments were stained with the fluorescent dye Hoechst 33342 (200 μg·ml⁻¹). For the transplantation of the fragments, the recipient animals were anaesthetised, as above. After removing the cover glass of the dorsal skinfold chamber, the striated muscle tissue of the chamber was carefully rinsed with prewarmed saline solution and two endometrial fragments were placed onto the tissue with a maximal distance to each other to exclude their mutual interference during the engraftment process (Laschke, Elitzsch, Vollmar, & Menger, 2005). Finally, the chamber was closed again with a new cover glass.

The vascularisation of the newly developing endometriotic lesions was assessed by means of repetitive intravital fluorescence microscopy, as described previously in detail (Körbel, Gerstner, Menger, & Laschke, 2018). For this purpose, the anaesthetised mice received an intravenous injection of the plasma marker 0.1‐ml 5% FITC‐labelled dextran (150,000 Da) into the retrobulbar venous plexus for contrast enhancement. The microscopic images were recorded and quantitatively analysed by means of the computer‐assisted off‐line analysis system CapImage (version 8.5; Zeintl, Heidelberg, Germany). The analyses included the planimetric determination of the size of the endometriotic lesions (mm²) and their functional microvessel density (FMD), that is, the length of red blood cell (RBC)‐perfused microvessels per observation area (cm·cm⁻²). Moreover, we measured the diameter (μm) of 20 randomly selected microvessels per lesion. Within these vessels, we assessed the centreline RBC velocity (μm·s⁻¹) according to the computer‐assisted line‐shift‐diagram method (De Vriese et al., 2000). At the end of the experiments, the animals were killed by cervical dislocation and the endometriotic lesions and the surrounding tissue were excised and further processed for immunohistochemical analyses.

2.10. Model of intraperitoneal endometriosis

For the induction of intraperitoneal endometriotic lesions, uterine tissue samples from donor mice in the stage of oestrus were prepared as previously described (Rudzitis‐Auth, Nenicu, Nickels, Menger, & Laschke, 2016). In brief, the uterine horns of anaesthetised mice were excised, transferred to a petri dish containing DMEM (10% FCS, 100 U·ml⁻¹ penicillin and 0.1 mg·ml⁻¹ streptomycin) and opened longitudinally. The donor animals were then killed by cervical dislocation. Thereafter, uterine tissue samples with a diameter of 2 mm were carefully removed by means of a dermal biopsy punch and two samples were fixed with a 6‐0 Prolene suture to the right and left abdominal wall of each anaesthetised recipient mouse through a midline incision. The laparotomy was then closed with running 6‐0 Prolene muscle and skin sutures. After the procedure, the mice received a subcutaneous injection of 5 mg·kg⁻¹ carprofen for post‐surgical analgesia and immediately underwent the first ultrasound imaging.

The development of endometriotic lesions was repetitively analysed using a Vevo 770™ high‐resolution ultrasound imaging system (VisualSonics, Toronto, ON, Canada), which was equipped with a real‐time microvisualisation (RMV™) 704 Scanhead (VisualSonics) with a centre frequency of 40 MHz and a focal depth of 6 mm, as previously described (Laschke et al., 2010). The ultrasound images were analysed with the three‐dimensional reconstruction and analysis software from VisualSonics (Vevo 770 V2.3.0). The analyses included the measurement of the overall volume of endometriotic lesions (mm³), their stromal tissue and cysts (mm³). For this purpose, the boundaries of endometriotic lesions and their cysts were manually outlined in parallel slices with a step size of 200 μm in the three‐dimensional ultrasound images. Based on the outlined areas, volumes were then computed by the VisualSonics software. Moreover, we calculated the growth of the lesions and their stromal tissue by dividing the measured lesion and stromal tissue volumes at individual time points by the initial lesion and stromal tissue volumes on Day 0 (given in %). In addition, we assessed the fraction of cyst‐containing lesions (in % of all analysed lesions).

At the end of the in vivo experiments, the largest diameter (D1) and perpendicularly aligned diameter (D2) of the endometriotic lesions were measured by means of a digital calliper and the lesion size (S) was calculated as S = D1 * D2 * π/4 (Becker et al., 2008). Finally, the animals were killed by cervical dislocation and the lesions were excised and further processed for histological, immunohistochemical and quantitative real‐time PCR (qRT‐PCR) analyses.

2.11. Experimental protocol

In a first set of experiments, a total of 24 endometrial fragments from five donor mice were isolated and transplanted into the dorsal skinfold chambers of 12 recipient animals. The mice were randomly divided in two groups and received either 10 mg·kg⁻¹ NVP‐BHG712 (n = 6) or vehicle (10% NMP; n = 6) by daily intragastric administration. Intravital fluorescence microscopy of the developing endometriotic lesions was performed directly after tissue transplantation (d0) as well as on Days 3, 6, 10 and 14. At the end of the experiments, the tissue samples were excised for histology and immunohistochemistry.

In a second set of experiments, a total of 80 uterine tissue samples from four donor mice were transplanted into the abdominal cavity of 20 recipient animals. Two transplants were fixed with a 6.0‐Prolene suture at the right and the left abdominal wall of the animals. The mice were randomly divided in two groups and received either 10 mg·kg⁻¹ NVP‐BHG712 (n = 10) or vehicle (10% NMP; n = 10) by daily intragastric administration. High‐resolution ultrasound imaging of the developing endometriotic lesions was performed directly after tissue transplantation (d0) as well as on Days 7, 14, 21 and 28. At the end of the experiments, the size of the endometriotic lesions was additionally assessed by means of a digital calliper. Subsequently, the tissue samples were excised and further processed for histology and immunohistochemistry. In addition, a total of 72 uterine tissue samples from three donor mice were transplanted into the abdominal cavity of 18 recipient animals, which received either 10 mg·kg⁻¹ NVP‐BHG712 (n = 9) or vehicle (10% NMP; n = 9) by daily intragastric administration. On Day 7 after transplantation, the tissue samples were excised and further processed for histology and immunohistochemistry. Finally, a total of 64 uterine tissue samples from five donor mice were transplanted into the abdominal cavity of 16 recipient animals, which received either 10 mg·kg⁻¹ NVP‐BHG712 (n = 8) or vehicle (10% NMP; n = 8) by daily intragastric administration. On Day 28 after transplantation, the tissue samples were excised and further processed for qRT‐PCR analyses.

2.12. Histology and immunohistochemistry

Formalin‐fixed specimens of the endometriotic lesions were embedded in paraffin. Three‐micrometre‐thick sections were cut and stained with haematoxylin and eosin (HE) according to standard protocols.

For the immunofluorescent detection of microvessels within the lesions, sections were stained with a monoclonal rat anti‐mouse antibody against the endothelial cell marker CD31 (1:100). A goat anti‐rat IgG Alexa555 antibody served as secondary antibody (1:200). Cell nuclei were stained with Hoechst 33342 (2 μg·ml⁻¹). The microvessel density (mm⁻²) was measured using a BZ‐8000 microscope (Keyence). For this purpose, the overall number of CD31⁺ microvessels was counted and divided by the area of the stromal tissue.

For the analysis of vessel maturation, additional sections were stained with a monoclonal rat anti‐mouse antibody against CD31 (1:100) and a polyclonal rabbit anti‐human antibody against α‐SMA (1:100). A goat anti‐rat IgG Alexa555 (1:200) and a goat anti‐rabbit IgG Alexa488 antibody (1:200) served as secondary antibodies. Cell nuclei were stained with Hoechst 33342 (2 μg·ml⁻¹). The sections were examined under a BX60 microscope (Olympus, Hamburg, Germany) and the fraction of CD31⁺/α‐SMA⁺ microvessels (given in %) was determined by dividing the number of CD31⁺/α‐SMA⁺ microvessels by the overall number of CD31⁺ microvessels.

For the immunohistochemical detection of proliferating and apoptotic cells in the stroma of endometriotic lesions, sections were stained with a rabbit polyclonal antibody against the proliferation marker Ki67 (1:100) and a rabbit polyclonal antibody against the apoptosis marker cleaved caspase‐3 (1:100). A biotinylated goat anti‐rabbit antibody (ready‐to‐use) served as secondary antibody followed by streptavidin‐peroxidase (ready‐to‐use). 3‐Amino‐9‐ethylcarbazole (AEC Substrate System) was used as chromogen and counterstaining was performed with hemalaun. The fraction of proliferating and apoptotic cells was assessed by counting the numbers of positive cells in 4 regions of interest at 400× magnification within the endometriotic lesions and is given in %.

The immuno‐related procedures used comply with the recommendations made by the British Journal of Pharmacology (Alexander et al., 2018).

2.13. qRT‐PCR

The mRNA expression level of EphB4 was assessed by qRT‐PCR within endometriotic lesions on Day 28. For this purpose, total RNA was isolated from the lesions by using the RNeasy Mini Kit (Qiagen, Hilden, Germany) according to the manual instructions. The RNA concentration was measured by a DeNovix DS‐11 spectrophotometer (Biozym, Hessisch Oldendorf, Germany). The QuantiNova Reverse Transcription Kit (Qiagen) was used for cDNA synthesis of 1‐μg extracted RNA. The analysis of the mRNA expression level was performed by using the QuantiNova SYBR Green PCR Kit (Qiagen) with a MiniOpticon RT‐PCR System (BioRad, München, Germany). GAPDH was used as an internal control. The specific sequences of the primers are listed in Table 1.

TABLE 1.

Specific forward and reverse primers for qRT‐PCR

Primer	Sequence
EphB4
Forward	GGA AAC GGC GGA TCT GAA ATG
Reverse	TGG ACG CTT CAT GTC GCA C
GAPDH
Forward	TGA CCT CAA CTA CAT GGT CTA CA
Reverse	CTT CCC ATT CTC GGC CTT G

2.14. Data and statistical analysis

The data and statistical analysis comply with the recommendations of the British Journal of Pharmacology on experimental design and analysis in pharmacology (Curtis et al., 2018). All experiments were designed to generate groups of equal size, using randomisation and blinded analysis. The statistical analysis was undertaken only for experiments where each group size was at least n = 5 of independent values and performed using these independent values and not technical replicates. Technical replicates were solely used to ensure the reliability of single values.

The in vitro data were first analysed for normal distribution and equal variance. Differences between the experimental groups were assessed by ANOVA (parametric data) or ANOVA on Ranks (non‐parametric data) followed by the Student–Newman–Keuls post hoc test (SigmaPlot 13.0; Jandel Corporation, San Rafael, CA, USA). The post hoc test was conducted only if F in ANOVA achieved the necessary level of statistical significance and there was no significant variance inhomogeneity. For the qRT‐PCR analyses, the measured values were corrected to control group values by data normalisation to reduce the variation between independent experiments.

For the in vivo experiments, the group sizes were chosen according to previous studies using the herein described endometriosis models (Feng, Menger, & Laschke, 2013; Rudzitis‐Auth, Nenicu, Nickels, Menger, & Laschke, 2016; Rudzitis‐Auth, Nickels, Menger, & Laschke, 2018). Data were first analysed for normal distribution and equal variance. In case of parametric data, differences between two experimental groups were assessed by the unpaired Student's t‐test. In case of non‐parametric data, differences between two experimental groups were assessed by the Mann–Whitney rank sum test. To test for time effects within each experimental group, ANOVA for repeated measurements was applied, as appropriate. This was followed by the Student–Newman–Keuls post hoc test (SigmaPlot 13.0; Jandel Corporation). The post hoc test was conducted only if F in ANOVA achieved the necessary level of statistical significance and there was no significant variance inhomogeneity. All data are given as mean ± SEM. Statistical significance was accepted for P < .05.

2.15. Nomenclature of targets and ligands

Key protein targets and ligands in this article are hyperlinked to corresponding entries in http://www.guidetopharmacology.org, the common portal for data from the IUPHAR/BPS Guide to PHARMACOLOGY (Harding et al., 2018) and are permanently archived in the Concise Guide to PHARMACOLOGY 2019/20 (Alexander et al., 2019).

3. RESULTS

3.1. Effect of NVP‐BHG712 on the viability of HDMEC

In a first step, we analysed the effect of different doses of the small molecule inhibitor NVP‐BHG712 on the viability of HDMEC by means of WST‐1 and lLDH assays (Figure 1a,b). We found that doses up to 2 μM have no effect on HDMEC, whereas higher doses of 5–50 μM reduce their viability (Figure 1a). Moreover, the LDH assay revealed that only a dose of 50‐μM NVP‐BHG712 exerts a cytotoxic effect on the exposed cells (Figure 1b). Based on these findings, we used doses of 1‐ and 2‐μM NVP‐BHG712 for all further in vitro analyses.

FIGURE 1.

Effect of NVP‐BHG712 on the viability of human dermal microvascular endothelial cells (HDMEC). (a, b) Viability (absorption_{450 nm}) of HDMEC (a) and cytotoxicity (absorption_{492 nm}) of NVP‐BHG712 (b). HDMEC were exposed for 24 hr to different doses (1–50 μM) of NVP‐BHG712, Triton X‐100 as high control (grey bar in b) or vehicle (control). Mean ± SEM (n = 5; four replicates); *P < .05 versus control

3.2. Effect of NVP‐BHG712 on the angiogenic activity of endothelial cells

To assess the effect of NVP‐BHG712 on the angiogenic activity of endothelial cells, we performed a panel of in vitro angiogenesis assays. By means of a transwell migration assay, we could demonstrate that 1‐ and 2‐μM NVP‐BHG712 suppress the migration of HDMEC when compared to vehicle‐treated controls (Figure 2a–d). These doses of the small molecule inhibitor also markedly reduced the number of meshes in the tube formation assay (Figure 2e–h). In line with these results, we found that NVP‐BHG712‐treated HDMEC spheroids exhibit a diminished sprouting ability, as indicated by a lower average sprout length when compared to spheroids of vehicle‐treated controls (Figure 2i–l).

FIGURE 2.

Effect of NVP‐BHG712 on the angiogenic activity of endothelial cells. (a–c) Light microscopic images of human dermal microvascular endothelial cells (HDMEC) in the transwell migration assay. The cells were exposed for 24 hr prior to the assay to vehicle (a, control), 1‐μM (b), or 2‐μM NVP‐BHG712 (c). Scale bars: 60 μm. (d) Migrated cells (mm⁻²), which were exposed for 24 hr to vehicle (control, white bar), 1‐μM or 2‐μM NVP‐BHG712 (black bars). Mean ± SEM (n = 5; four replicates); *P < .05 versus control. (e–g) Phase‐contrast microscopic images of tube‐forming HDMEC, which were exposed to vehicle (e, control), 1‐μM (f), or 2‐μM NVP‐BHG712 (g). Scale bars: 800 μm. (h) Number of meshes, which were exposed to vehicle (control, white bar), 1‐μM, or 2‐μM NVP‐BHG712 (black bars). Mean ± SEM (n = 5; six replicates); *P < .05 versus control. (i–k) Phase‐contrast microscopic images of sprouting HDMEC spheroids, which were exposed to vehicle (i, control), 1‐μM (j), or 2‐μM NVP‐BHG712 (k). Scale bars: 150 μm. (l) Average sprout length (μm) of HDMEC spheroids, which were exposed to vehicle (control, white bar), 1‐μM, or 2‐μM NVP‐BHG712 (black bars). Mean ± SEM (n = 5; 10 replicates); *P < .05 versus control. (m–o) Phase‐contrast microscopic images of mouse aortic rings, which were exposed for 6 days to vehicle (m, control), 1‐μM (n), or 2‐μM NVP‐BHG712 (o). Scale bars: 1 mm. (p) Sprout area (mm²) of the outer aortic vascular sprouting on Day 6 after incubation of aortic rings with vehicle (control, white bar), 1‐μM, or 2‐μM NVP‐BHG712 (black bars). Mean ± SEM (n = 5; six replicates); *P < .05 versus control

Finally, we performed an aortic ring assay to confirm the findings from our HDMEC‐based angiogenesis assays. Of interest, we also detected a strong anti‐angiogenic effect of NVP‐BHG712 on endothelial cells of mouse origin in this ex vivo angiogenesis assay (Figure 2m–p). In fact, the dose of 2‐μM NVP‐BHG712 almost completely suppressed the growth of vessel sprouts out of Matrigel‐embedded mouse aortic rings (Figure 2o,p).

3.3. Effect of NVP‐BHG712 on endometriotic lesions in dorsal skinfold chambers

To analyse the formation of new microvascular networks within endometriotic lesions and their microhaemodynamic characteristics in vivo, we used the mouse dorsal skinfold chamber model in combination with intravital fluorescence microscopy. This approach demonstrated that endometriotic lesions in NVP‐BHG712‐treated animals exhibit a significantly reduced FMD on Day 14 when compared to that of vehicle‐treated controls (Figure 3a–c). Individual microvessels within these lesions presented with significantly larger diameters over the 14 days observation time (Table 2). In contrast, the analysis of the centreline RBC velocity revealed no significant differences between NVP‐BHG712‐treated animals and controls (Table 2).

FIGURE 3.

Effect of NVP‐BHG712 on endometriotic lesions in dorsal skinfold chambers. (a, b) Intravital fluorescence microscopic images of endometriotic lesions (borders marked by broken line) on Day 14 after transplantation of endometrial fragments into the dorsal skinfold chamber of a vehicle‐treated control (a) and a NVP‐BHG712‐treated BALB/c mouse (b). Scale bars: 150 μm. (c) Functional microvessel density (FMD, cm·cm⁻²) of endometriotic lesions in dorsal skinfold chambers of vehicle‐treated controls (white circles; n = 6) and NVP‐BHG712‐treated BALB/c mice (black circles; n = 6). Mean ± SEM. *P < .05 versus control; **P < .05 versus d0. (d, e) Immunofluorescent detection of microvessels (arrows) within endometriotic lesions on Day 14 after transplantation of endometrial fragments into the dorsal skinfold chamber of a vehicle‐treated control (d) and a NVP‐BHG712‐treated BALB/c mouse (e). Scale bars: 20 μm. (f) Microvessel density (mm⁻²) of endometriotic lesions in vehicle‐treated controls (white bar; n = 6) and NVP‐BHG712‐treated BALB/c mice (black bar; n = 6). Mean ± SEM; *P < .05 versus control

TABLE 2.

Diameter (μm) and centreline RBC velocity (μm·s⁻¹) of individual microvessels within endometriotic lesions in dorsal skinfold chambers of vehicle‐treated controls (n = 6) and NVP‐BHG712‐treated BALB/c mice (n = 6), as assessed by intravital fluorescence microscopy and computer‐assisted image analysis throughout an observation period of 14 days

	d3	d6	d10	d14
Diameter (μm)
Control	18.6 ± 1.9*	15.0 ± 1.5**	11.7 ± 0.7**	10.1 ± 0.3**
NVP‐BHG712	25.2 ± 1.3*	18.5 ± 2.1**	14.4 ± 1.1*, **	12.2 ± 0.6*, **
Centreline RBC velocity (μm·s−1)
Control	60.9 ± 25.6	108.2 ± 20.5	150.7 ± 32.4	210.2 ± 44.2**
NVP‐BHG712	75.4 ± 35.5	92.2 ± 27.6	150.5 ± 25.5**	136.4 ± 24.1

Note. Mean ± SEM.

Abbreviation: RBC, red blood cell.

^* P < .05 versus control.

^** P < .05 versus Day 3.

The reduction of vascularisation of NVP‐BHG712‐treated lesions was confirmed by additional immunohistochemical stainings against the endothelial cell marker CD31 at the end of the in vivo experiments. These stainings showed a significantly reduced microvessel density of NVP‐BHG712‐treated lesions when compared to controls (Figure 3d–f).

3.4. Effect of NVP‐BHG712 on intraperitoneal endometriotic lesions

Finally, we used an endometriosis model, which allows evaluating the effect of NVP‐BHG712 on endometriotic lesions in a more disease‐related environment. For this purpose, we transplanted endometrial fragments from syngeneic donor mice into the peritoneal cavity of recipient animals and analysed their growth and cyst formation by repetitive high‐resolution ultrasound imaging. During the time course of 28 days, the transplanted tissue developed into endometriotic lesions, as proven by the presence of endometrial stroma and epithelial cells forming cyst‐like dilated glands (Figure 4a–d). In both NVP‐BHG712‐ and vehicle‐treated mice the lesions exhibited a comparable initial volume of ~1.2 mm³ (Figure 4e). Throughout the following weeks, the lesions in vehicle‐treated mice progressively grew until a final volume of ~3.3 mm³ on Day 28 (Figure 4e). This process was markedly suppressed in NVP‐BHG712‐treated mice, as indicated by a final lesion volume of ~1.7 mm³ (Figure 4e). The reduced lesion growth in NVP‐BHG712‐treated animals was caused by the inhibition of stromal tissue growth (Figure 4f–h). In contrast, the fraction of cyst‐like dilated endometrial glands and their volumes was not significantly different between the two groups (Figure 4i,j). Additional two‐dimensional measurements by means of a digital calliper also revealed a smaller lesion size in NVP‐BHG712‐treated animals on Day 28 when compared to controls (2.5 ± 0.3 vs. 1.9 ± 0.1 mm²).

FIGURE 4.

Effect of NVP‐BHG712 on the growth and cyst formation of intraperitoneal endometriotic lesions. (a–d) Haematoxilin‐eosin (HE)‐stained sections (a, b) and high‐resolution ultrasound imaging (c, d) of developing endometriotic lesions (borders marked by closed line, cyst‐like dilated endometrial glands marked by dotted line) 28 days after transplantation of uterine tissue samples to the abdominal wall of a vehicle‐treated control (a, c) and a NVP‐BHG712‐treated BALB/c mouse (b, d). Scale bars: a, b = 420 μm; c, d = 1 mm. (e–j) Overall lesion volume (e, mm³), lesion growth (f, %), stromal tissue volume (g, mm³), stromal tissue growth (h, %), cyst volume (i, mm³) and fraction of cyst‐containing lesions (j, %) of endometriotic lesions in vehicle‐treated controls (white bars; n = 10) and NVP‐BHG712‐treated BALB/c mice (black bars; n = 10). Mean ± SEM; **P < .05 versus Day 0; *P < .05 versus control

We further examined the lesions by means of immunohistochemistry (Figure 5a–o). The microvessel density and the number of Ki67⁺ proliferating stromal and glandular cells were comparable in NVP‐BHG712‐ and vehicle‐treated lesions on Day 28 (Figure 5c,d,g,h,m,n). In addition, both NVP‐BHG712‐ and vehicle‐treated lesions contained a comparably low number of cleaved caspase‐3⁺ apoptotic stromal cells at this late time point (Table 3), indicating that the observed effects of NVP‐BHG712 on angiogenesis and stromal cell proliferation were not induced by cytotoxicity. Although not proven to be significant, we detected less CD31⁺/α‐SMA⁺ mature microvessels in NVP‐BHG712‐treated lesions on Day 28 when compared to controls (Figure 5k,l,o).

FIGURE 5.

Effect of NVP‐BHG712 on the vascularisation and proliferation of intraperitoneal endometriotic lesions. (a–d) Immunofluorescent detection of microvessels (arrows) on Day 7 and Day 28 after transplantation of uterine tissue samples to the abdominal wall of vehicle‐treated controls and NVP‐BHG712‐treated BALB/c mice. Scale bars: 20 μm. (e–h) Immunohistochemical detection of Ki67⁺ proliferating stromal cells (arrows) and glandular epithelial cells (arrowheads) on Day 7 and Day 28 after transplantation of uterine tissue samples to the abdominal wall of vehicle‐treated controls and NVP‐BHG712‐treated BALB/c mice. Scale bars: 20 μm. (i–l) Immunohistochemical detection of CD31⁺/α‐SMA⁺ mature microvessels (arrows) within endometriotic lesions on Day 7 and Day 28 after transplantation of uterine tissue samples to the abdominal wall of vehicle‐treated controls and NVP‐BHG712‐treated BALB/c mice. Scale bars: 40 μm. (m) Microvessel density (mm⁻²) of endometriotic lesions in vehicle‐treated controls (white bars; n = 8–10) and NVP‐BHG712‐treated BALB/c mice (black bars; n = 9–10). Mean ± SEM. *P < .05 versus control. (n) Ki67⁺ cells (%) within the stroma and the glands of endometriotic lesions in vehicle‐treated controls (white bars; n = 8–10) and NVP‐BHG712‐treated BALB/c mice (black bars; n = 9–10). Mean ± SEM. *P < .05 versus control. (o) CD31⁺/α‐SMA⁺ microvessels (%) within endometriotic lesions of vehicle‐treated controls (white bars; n = 8–10) and NVP‐BHG712‐treated BALB/c mice (black bars; n = 9–10). Mean ± SEM. *P < .05 versus control

TABLE 3.

Relative mRNA expression level of EphB4 (% of control) within endometriotic lesions of vehicle‐treated controls (n = 8) and NVP‐BHG712‐treated BALB/c mice (n = 8) after 28 days

mRNA expression level of EphB4 (% of control)
Control a	100.0 ± 9.5
NVP‐BHG712	78.8 ± 8.6

Note. Mean ± SEM.

Furthermore, we analysed additional endometriotic lesions on Day 7. This analysis included nine NVP‐BHG712‐treated mice and eight vehicle‐treated animals, because one animal died in the control group during anaesthesia. At this early time point, we could show that the density of CD31⁺ microvessels was significantly lower in NVP‐BHG712‐treated lesions when compared to vehicle‐treated controls (Figure 5a,b,m). Of interest, these lesions also presented with a significantly reduced fraction of proliferating Ki67⁺ stromal cells (Figure 5e,f,n) and a significantly higher fraction of CD31⁺/α‐SMA⁺ mature microvessels (Figure 5i,j,o).

Finally, we measured the mRNA expression levels of EphB4 within NVP‐BHG712‐ and vehicle‐treated lesions on Day 28 by means of qRT‐PCR. NVP‐BHG712‐treated lesions exhibited a tendency towards lower EphB4 mRNA expression levels when compared to vehicle‐treated controls (Table 4).

TABLE 4.

Apoptotic cleaved caspase‐3⁺ cells (%) in the stroma of endometriotic lesions on Day 28 after transplantation of uterine tissue samples to the abdominal wall of vehicle‐treated control (n = 10) and NVP‐BHG712‐treated BALB/c mice (n = 10), as assessed by immunohistochemical analysis

Cleaved caspase‐3+ cells (%)
Control a	1.2 ± 0.2
NVP‐BHG712	1.1 ± 0.2

Note. Mean ± SEM.

4. DISCUSSION

Recent studies suggest that EphB4 is involved in the pathogenesis of endometriosis. EphB4 is overexpressed in ectopic lesions when compared to eutopic endometrium. This is associated with an increased expression of hypoxia‐inducible factor‐1α, VEGF receptor 2 and PDGF‐B (Yerlikaya et al., 2016). Moreover, EphB4 signalling is a strong promotor of blood vessel formation (Füller, Korff, Kilian, Dandekar, & Augustin, 2003; Martiny‐Baron et al., 2010). An adequate vascularisation is, in turn, a major prerequisite for the development and survival of endometriotic lesions inside the peritoneal cavity (Laschke & Menger, 2018). In line with these findings, we herein show that the disruption of EphB4 signalling suppresses angiogenesis in endometriotic lesions. This diminishes the proliferation of their stromal cells, resulting in a reduced overall lesion size.

In a first set of in vitro experiments, we assessed the anti‐angiogenic effects of the selective EphB4 inhibitor NVP‐BHG712. We found that non‐cytotoxic doses of the compound reduce the migration, tube formation and sprouting activity of human endothelial cells. Moreover, they suppress the growth of vessel sprouts out of mouse aortic rings. This demonstrates that the inhibition of EphB4 signalling targets multiple endothelial functions, which are necessary for the process of blood vessel formation. Previous studies indicate that this is most probably mediated by the suppression of the phosphatidylinositol‐3 kinase/Akt pathway (Maekawa et al., 2003; Steinle et al., 2002). This pathway also crucially regulates cell proliferation and angiogenesis in endometriosis (Barra, Ferro Desideri, & Ferrero, 2018; Rudzitis‐Auth, Körbel, Scheuer, Menger, & Laschke, 2012).

We next analysed the effects of NVP‐BHG712 on the vascularisation of newly developing endometriotic lesions. For this purpose, we transplanted murine endometrial fragments into dorsal skinfold chambers and examined them by means of intravital fluorescence microscopy. This approach bears the major advantage that the dynamic process of angiogenesis can be repetitively analysed in vivo, including the measurement of microhemodynamic parameters, such as the diameter and centreline RBC velocity of individual microvessels but also microvascular network perfusion (Laschke & Menger, 2016). In line with our in vitro results, we detected a significantly reduced FMD of NVP‐BHG712‐treated lesions when compared to vehicle‐treated controls. However, this was only the case at the final observation time point on Day 14. This finding can be explained by the different mechanisms that are involved in the vascularisation of ectopic endometrial fragments (Laschke & Menger, 2018). These fragments already contain preformed microvessels (Machado et al., 2014), which can be reperfused by developing interconnections to the host microvasculature, also referred to as inosculation (Laschke & Menger, 2018). Cheng et al. (2011) reported that this occurs via a “wrapping and tapping” process, which is initiated by the arrangement of endothelial cells from a preformed microvessel around a blood‐perfused host vessel. There, they release MMP‐9 and MMP‐14, resulting in the degradation of the underlying host endothelium and an inflow of blood into the preformed vessel segment (Cheng et al., 2011). It may be assumed that this process was not completely suppressed by NVP‐BHG712 treatment in our experimental setting. In fact, only one successful vessel interconnection could have already enabled the blood perfusion of major parts of the preformed microvascular networks within the endometrial fragments of both groups. This could explain the comparable FMD in NVP‐BHG712‐ and vehicle‐treated lesions during the first days after transplantation of endometrial fragments. On the other hand, preformed microvessels can serve as the origin of new blood vessels, which progressively grow out of their wall into the surrounding tissue (Laschke et al., 2008). The latter mechanism of sprouting angiogenesis may have been inhibited by NVP‐BHG712 throughout our in vivo experiment, finally resulting in a significantly lower FMD when compared to vehicle‐treated controls during the late phase of the 14‐day observation period.

Of interest, we also detected significantly larger diameters of individual microvessels within NVP‐BHG712‐treated lesions when compared to vehicle‐treated controls. This is a typical sign for an impaired vessel stabilisation and maturation during the development of new microvascular networks (Laschke et al., 2014). In line with this observation, EphB4 signalling has been shown to be crucially involved in the regulation of vascular remodelling (Yang et al., 2016). In fact, the disruption of this signalling pathway leads to detachment of pericytes from the microvascular endothelium (Dimova et al., 2013).

Moreover, we analysed endometriotic lesions in the peritoneal cavity, which represents a more disease‐related environment when compared to the striated muscle and subcutaneous tissue within the dorsal skinfold chamber (Laschke & Menger, 2007). By means of high‐resolution ultrasound imaging, we could demonstrate that intraperitoneally transplanted uterine tissue samples of the control group progressively increase in size during the 28‐day observation period, whereas NVP‐BHG712‐treated grafts exhibit a significantly reduced growth rate. Of interest, we further found that NVP‐BHG712‐treated lesions exhibit lower EphB4 mRNA expression levels when compared to vehicle‐treated controls. This is in line with a study of You et al. (2017), which reported similar effects on EphB4 mRNA expression levels using even a lower dosage of the inhibitor (8 mg·kg⁻¹, intragastric administration, every second day) when compared to our treatment regime (10 mg·kg⁻¹, intragastric administration, every day). Hence, it may be speculated that the herein observed effects of NVP‐BHG712 on endometriotic lesions are not only mediated by its inhibitory action on EphB4 kinase activity (Martiny‐Baron et al., 2010) but also on its inhibitory action on EphB4 mRNA expression.

Immunohistochemical analyses on Day 28 revealed that vehicle‐ and NVP‐BHG712‐treated endometriotic lesions do not differ in terms of their microvessel density and proliferation of their stromal and glandular cells. This observation can be explained by the fact that the engraftment of the transplanted uterine tissue is already completed at this late time point. Accordingly, this phase of lesion development is rather characterised by vessel maturation and remodelling than by extensive angiogenesis and cell turnover (Rudzitis‐Auth, Nenicu, Nickels, Menger, & Laschke, 2016; Rudzitis‐Auth, Nickels, Menger, & Laschke, 2018). Therefore, we performed additional analyses of peritoneally transplanted uterine tissue on Day 7 and detected a reduced microvessel density and stromal cell proliferation within NVP‐BHG712‐treated lesions when compared to controls. This indicates the importance of EphB4 signalling for the early vascularisation and growth of newly developing endometriotic lesions.

More detailed immunohistochemical analyses focusing on the maturation stage of individual microvessels revealed that NVP‐BHG712‐treated lesions contained a significantly higher fraction of CD31⁺/α‐SMA⁺ mature microvessels on Day 7 when compared to controls. It may be speculated that this observation is due to the strong anti‐angiogenic effect of the inhibitor in this early phase of lesion formation, resulting in the inhibition of newly ingrowing CD31⁺/α‐SMA⁻ immature microvessels from the surrounding host tissue into the lesions and, thus, in a relative increase of the fraction of pre‐existing mature microvessels within the transplanted uterine tissue. In contrast, on Day 28, we detected less CD31⁺/α‐SMA⁺ mature microvessels in NVP‐BHG712‐treated lesions when compared to controls. This further confirms the inhibitory effect of NVP‐BHG712 on vessel maturation and remodelling.

Finally, we and others reported that the vascularisation of endometriotic lesions is also dependent on vasculogenesis, that is, the incorporation of bone marrow‐derived circulating endothelial progenitor cells into the microvascular endothelium (Laschke, Giebels, et al., 2011; Becker et al., 2011). This is mediated by the interaction of stromal cell‐derived factor (SDF‐1), also known as C‐X‐C motif chemokine 12 (CXCL12ß), and its receptor chemokine receptor type 4 (CXCR4) on endothelial progenitor cells (Virani, Edwards, Thomas, Childs, & Tayade, 2013). Accordingly, the inhibition of this interaction by AMD3100 (plerixafor) reduces the number of homing endothelial progenitor cells in endometriotic lesions, which results in an impaired lesion vascularisation (Laschke, Giebels, et al., 2011). Of interest, it has been shown that EphB2 and EphB4 activation enhances SDF‐1‐induced endothelial cell chemotaxis (Salvucci, de la Luz, Martina, McCormick, & Tosato, 2006). Moreover, Dimova et al. (2013) found that exposure of the area vasculosa of chicken embryos to NVP‐BHG712 leads to the down‐regulation of SDF‐1 and CXCR4 expression. Although not further analysed in the present study, it may be therefore speculated that the inhibition of EphB4 signalling also contributes to a reduced lesion vascularisation by the suppression of SDF‐1/CXCR4‐driven vasculogenesis.

In summary, we have demonstrated that inhibition of EphB4 by NVP‐BHG712 suppresses the vascularisation and growth of endometriotic lesions. Accordingly, EphB4 represents a promising target for the treatment of endometriosis. Currently, small molecule inhibitors targeting EphB4, such as tesevatinib (XL‐647), are under development and clinical evaluation for the therapy of different malignancies (Chen, Zhang, & Zhang, 2019). Hence, it would be interesting to analyse in future clinical studies the effects of these novel inhibitors on endometriotic lesions.

AUTHOR CONTRIBUTIONS

J.R.‐A., M.D.M. and M.W.L. participated in research design. J.R.‐A., S.A.F. and V.B. conducted the experiments. J.R.‐A. and S.A.F. performed data analysis. J.R.‐A. and M.W.L. drafted the manuscript. J.R.‐A., S.A.F., V.B., M.D.M. and M.W.L. critically revised the manuscript and gave approval to the final manuscript version.

CONFLICT OF INTEREST

The authors declare no conflicts of interest.

DECLARATION OF TRANSPARENCY AND SCIENTIFIC RIGOUR

This Declaration acknowledges that this paper adheres to the principles for transparent reporting and scientific rigour of preclinical research as stated in the BJP guidelines for Design & Analysis, Immunoblotting and Immunochemistry and Animal Experimentation, and as recommended by funding agencies, publishers and other organisations engaged with supporting research.

Rudzitis‐Auth J, Fuß SA, Becker V, Menger MD, Laschke MW. Inhibition of erythropoietin‐producing hepatoma receptor B4 (EphB4) signalling suppresses the vascularisation and growth of endometriotic lesions. Br J Pharmacol. 2020;177:3225–3239. 10.1111/bph.15044