The ischemic time window of ectopic endometrial tissue crucially determines its ability to develop into endometriotic lesions

Endometriosis develop from shed endometrial fragments via retrograde menstruation. This affects the survival, proliferation and vascularization of the tissue and its final ability to form endometriotic lesions. Within this study, uterine tissue samples from donor mice were precultivated for 24 h or 72 h to simulate avascular periods. Their morphology, microvessel density, apoptotic activity and expression of angiogenesis-related proteins were analyzed in vitro. The formation of endometriotic lesions in vivo was assessed after transplantation of precultivated uterine tissue samples to the abdominal wall and dorsal skinfold chambers by means of high-resolution ultrasound, intravital fluorescence microscopy, histology and immunohistochemistry. In vitro, 72-h-precultivated uterine tissue samples exhibit extensive areas of tissue necrosis and high numbers of apoptotic cells as well as a significantly reduced cell and microvessel density. These samples failed to develop into endometriotic lesions. In contrast, the 24-h-precultivated samples showed, that their early vascularization and growth in vivo was improved when compared to controls. This indicates that avascular periods have a strong impact on the survival of ectopic endometrial tissue and the chance for the development of endometriosis.

Protein expression of precultivated uterine tissue samples. We next assesed the expression of 53 angiogenesis-related proteins in freshly isolated as well as 24-h-and 72-h-precultivated pooled uterine tissue samples by means of a proteome profiler mouse angiogenesis array. Of note, we detected an unusually high expression of distinct proteins in the group of 72-h-precultivated uterine tissue samples, although our analyses on cell viability demonstrated that these samples mainly contained dying and non-viable cells or even cellular debris (Supplementary Table S1). Because non-specific binding of antibodies is a well-known problem in the analysis of such samples 17,18 , we therefore excluded the expression data of this group to prevent misleading interpretations due to false-positive artefacts. Precultivation for 24 h upregulated the expression of most of the analyzed proteins when compared to freshly isolated controls ( Table 1). The highest upregulation was found for the pro-angiogenic factors Cyr61 and keratinocyte chemoattractant (KC) ( Table 1).

Development of endometriotic lesions.
In a second set of experiments, we induced endometriotic lesions by suturing freshly isolated and precultivated uterine tissue samples from C57BL/6-TgN(ACTB-EGFP)1Osb/J donor mice to the abominal wall of C57BL/6J recipient mice. After 28 days, the grafts were analyzed by histology to determine the take rate ( Fig. 3A-G). This analysis revealed that the take rate progressively decreased with the duration of precultivation. While ~ 80% of freshly isolated uterine tissue samples developed into endometriotic lesions consisting of endometrial glands and stroma, the take rate of 24-h-precultivated samples was only ~ 40% (Fig. 3G). However, additional caliper measurements revealed that the overall size of the lesions orginating from 24-h-precultivated tissue samples was significantly higher when compared to freshly isolated controls (Fig. 3H). Moreover, we found that only one out of 22 72-h-precultivated uterine tissue samples developed into a typical lesion, which corresponds to a take rate of only ~ 5% (Fig. 3G). Hence, we excluded this group from all further quantitative analyses in the peritoneal model.
The growth and cyst formation of newly developing endometriotic lesions was repeatedly analyzed by means of high-resolution ultrasound imaging throughout the 28-days in vivo experiment (Fig. 4A,B). Directly after transplantation into the peritoneal cavity, freshly isolated and 24-h-precultivated uterine tissue samples exhibited a comparable initial volume, which increased over time (Fig. 4C). However, the growth rate of lesions in the group of 24-h-precultivated tissue samples was significantly higher between day 14 and 28 when compared to controls (Fig. 4D). Accordingly, these lesions also exhibited a markedly higher overall lesion volume, stromal tissue www.nature.com/scientificreports/ volume and stromal tissue growth on day 21 and 28 ( Fig. 4C,E,F). The cyst volumes of the lesions in both groups remained rather low during the entire observation period (Fig. 4G). However, on day 28 the number of detectable cyst-like dilated endometrial glands was significantly higher in lesions originating from 24-h-precultivated uterine tissue samples when compared to controls (Fig. 4H).

Vascularization, proliferation and apoptosis of endometriotic lesions. At the end of the in vivo
experiments, the endometriotic lesions were additionally analyzed by immunohistochemistry. The lesions developing from freshly isolated and 24-h-precultivated uterine tissue samples exhibited a comparable density of CD31 + microvessels ( Fig. 5A-C). Furthermore, the fraction of CD31 + /green fluorescent protein (GFP) + microvessels was comparably high (~ 70-75%) in both groups ( Fig. 5D-F). These findings indicate that the major part of the lesions' final microvasculature originated from those GFP + microvessels, which were originally present in the uterine tissue samples of the C57BL/6-TgN(ACTB-EGFP)1Osb/J donor mice. In contrast, only a few GFPmicrovessels progressively grew from the surrounding host tissue into the lesions and developed interconnections to the pre-existing GFP + microvessels during the 28-day observation period.

Early vascularization of endometriotic lesions in dorsal skinfold chambers.
In a final set of experiments, we transplanted freshly isolated as well as 24-h-and 72-h-precultivated endometrial fragments from C57BL/6-TgN(ACTB-EGFP)1Osb/J donor mice into the dorsal skinfold chambers of C57BL/6J recipient animals to study their early vascularization. Due to their GFP signal, the fragments could be easily detected within the GFPhost tissue by means of intravital fluorescence microscopy ( Fig. 7A,B). As performed in the peritoneal endometriosis model, we determined the take rate of the grafts by histology. This analysis revealed that ~ 60% of freshly isolated and ~ 80% of 24-h-precultivated endometrial fragments developed into endometriotic lesions after transplantation. In contrast, only ~ 13% 72-h-precultivated grafts finally exhibited the typical morphology of endometriotic lesions with stromal and glandular cells. Hence, this group was excluded from further quantitative analyses. Repetitive intravital fluorescence microscopy showed the progressive formation of blood-perfused microvascular networks within both freshly isolated and 24-h-precultivated endometrial grafts throughout the 14-day observation period (Fig. 7C-E). However, at the end of the experiments the functional capillary density of 24-h-precultivated endometrial fragments was significantly higher when compared to freshly isolated controls (Fig. 7E). There were no significant differences in lesion sizes between the two groups ( Fig. 7F).

Discussion
Retrograde menstruation represents a key process in the pathogenesis of endometriosis 20 . During the passage from the uterus into the peritoneal cavity shed endometrial fragments lack a blood supply and, thus, suffer from hypoxic stress. This may markedly affect their viability, tissue integrity as well as their angiogenic, inflammatory and proliferative activity. Our novel findings now indicate that this ischemic time window crucially determines the ability of ectopic endometrial tissue to develop into endometriotic lesions.
To simulate in vitro different periods, in which the tissue lacks a blood supply, we herein precultivated uterine tissue samples from donor mice at 37 °C for 24 h or 72 h in standard cell culture medium in a humidified atmosphere and 5% CO 2 . We are aware that this experimental setting does not exactly reflect the conditions to which endometrial fragments are exposed to during retrograde menstruation through the fallopian tubes and their subsequent distribution inside the peritoneal cavity. In fact, the physiological oxygen concentration in the female reproductive tract and in the peritoneal cavity is much lower than under atmospheric conditions 21,22 . Hence, in vivo endometrial fragments may suffer more from hypoxia when compared to the fragments in our in vitro setting. On the other hand, it is well known that cells grown in a standard normoxic oxygen concentration may actually be exposed to hypoxia to near-anoxic conditions at the cellular level dependent on the density of the cell layer or the depth of the overlying medium 23 . In line with this finding, we herein detected significantly higher fractions of HIF-1α + cells within 24-h-and 72-h-precultivated uterine tissue samples when compared to freshly isolated, non-cultivated controls. In addition, it should be considered that the extent of oxygen diffusion from surrounding tissues may markedly vary between individual endometrial fragments and over time dependent on their in situ localization. Moreover, they may exhibit different shapes and sizes, which crucially determine their survival by oxygen diffusion. Furthermore, the time that is required for the successful engraftment of shed endometrial fragments inside the peritoneal cavity of endometriosis patients is completely unknown. Therefore, our approach should not be interpreted as an attempt to exactly mimic the in vivo situation during retrograde menstruation, but to provide a standardized experimental setting for our proof-of-principle study. For these reasons, it could be of interest to further confirm our results in future studies by means of alternative models   In a first set of experiments, we could demonstrate that the physiological morphology of uterine tissue samples is progressively lost during precultivation. However, we still detected endometrial glands within 24-h-precultivated samples. Moreover, the cell density, viability, microvessel density and fraction of apoptotic cells within these samples did not markedly differ from that of freshly isolated controls. This is an important finding considering the fact that an intact tissue integrity has been reported to be essential for the establishment of endometriotic lesions in different endometriosis models 15,24 . In contrast, 72-h-precultivated uterine tissue samples exhibited extensive areas of tissue necrosis, cellular debris and high numbers of apoptotic cells as well as a significantly reduced cell and microvessel density. Accordingly, these samples also failed to develop into endometriotic lesions after in vivo implantation. This indicates that the onset of endometriosis is only possible in a critical time window during which shed endometrial tissue must successfully reach the peritoneal cavity and engraft at ectopic sites. It may be speculated that pathological changes of the uterus function, such as the known hyper-and dysperistalsis in endometriosis patients 25,26 , provide ideal conditions to fulfil this requirement.
In our mouse model of surgically induced endometriosis we found a significantly lower take rate of 24-h-precultivated uterine tissue samples when compared to freshly isolated controls. On the other hand, newly developing endometriotic lesions originating from the precultivated samples exhibited a much more aggressive growth throughout the 28-days observation period, as indicated by markedly increased lesion and stromal tissue volumes in our ultrasound analyses. This interesting observation may be explained by the results of our proteome profiler mouse angiogenesis array. In this array, most of the analyzed proteins were strongly upregulated in the precultivated samples, most probably due to hypoxia-induced pathways, e.g. HIF-1α/Cyr61 signaling 27,28 . Accordingly, the upregulation of pro-angiogenic factors, such as Cyr61 and KC, may have markedly increased the angiogenic activity of the tissue. Noteworthy, we also detected an upregulation of MMP-3. MMPs are proteolytic enzymes that are secreted by endometrial fragments for the breakdown and the remodeling of the extracellular matrix, which is required for the invasion into the adjacent tissue 29 . Several studies suggest that an altered expression of  Furthermore, MMP-3 has been shown to be upregulated by IL-1, which was also overexpressed in our precultivated uterine tissue samples 32 . Additional immunohistochemical analyses of the endometriotic lesions on day 28 revealed a reduced number of MPO + neutrophilic granulocytes and CD68 + macrophages within lesions originating from 24-h-precultivated uterine tissue samples when compared to controls. This is a surprising finding considering the fact that these immune cells have been shown to promote the growth, development, vascularization and innervation of endometriotic lesions 33,34 . Hence, based on our ultrasound results we would have expected higher immune cell numbers in lesions of 24-h-precultivated tissue samples. However, the contribution of the immune system to the pathogenesis of endometriosis is highly complex and crucially dependent on the recruitment and activation of specific immune cell subtypes 33,35 . Accordingly, we further analyzed M1 and M2 macrophage polarization within the endometriotic lesions. Of interest, we found that the number of pro-inflammatory CD86 + M1 macrophages was significantly reduced within endometriotic lesions originating from 24-h-precultivated uterine tissue samples when compared to controls, whereas the number of pro-angiogenic and regenerative CD163 + M2 macrophages did not differ between the two groups. This indicates a shift of the M1:M2 ratio towards the M2 phenotype within lesions originating from 24-h-precultivated uterine tissue samples, which may have stimulated lesion development despite low overall macrophage numbers within the ectopic endometrial tissue.
Our immunohistochemical analyses further showed that lesions developing from freshly isolated and 24-h-precultivated uterine tissue samples exhibited a comparable microvessel density on day 28. In both groups,  www.nature.com/scientificreports/ 24-h precultivation period. Moreover, it demonstrates that the primary vascularization mode in the present endometriosis model is inosculation, i.e. the interconnection of pre-existing microvessels inside the grafted tissue with blood vessels of the surrounding host tissue 36,37 . However, particularly in the early phase after transplantation, sprouting angiogenesis is a major prerequisite for this process 38 . Hence, the hypoxia-induced overexpression of multiple pro-angiogenic factors may have markedly accelerated inosculation within the grafted 24-h-precultivated uterine tissue samples. Accordingly, we also found an improved early vascularization of 24-h-precultivated endometrial fragments in the dorsal skinfold chamber model, as indicated by a higher functional capillary density when compared to freshly isolated controls. This improved vascularization can explain the increased growth rate of the grafts within the peritoneal cavity. Besides, it is known that hypoxia does not only induce angiogenesis, but also directly promotes cell proliferation 39 , which may have further contributed to the higher overall volume and stromal tissue volume of lesions originating from 24-h-precultivated uterine tissue samples. Finally, hypoxia stimulates the production of reactive oxygen species in endometriotic cells, which enhances the growth of endometriotic lesions 40,41 . This is mediated by the activation of different pro-inflammatory intracellular pathways, including nuclear factor-κB and cyclooxygenase-2/prostaglandin E2 41 .
In summary, the present study demonstrates that a lack of blood supply has a strong impact on the integrity, survival as well as angiogenic and proliferative activity of ectopic endometrial tissue. If such ischemic periods are too long, the tissue completely regresses and is not anymore able to develop into endometriotic lesions. In contrast, shorter ischemic periods promote the overexpression of multiple growth factors inside the tissue, resulting in the formation of endometriotic lesions of aggressive growth. Hence, the duration of the ischemic time window for shed endometrial tissue may represent a key selection factor determining the chance for the development of endometriotic lesions inside the peritoneal cavity and, thus, the risk to suffer from endometriosis.

Materials and methods
Animals. For this study, 12-16-week-old female C57BL/6J wild-type mice and transgenic C57BL/6-TgN(ACTB-EGFP)1Osb/J mice (Institute for Clinical and Experimental Surgery, Saarland University, Homburg/Saar, Germany) were used. The transgenic mice expressed an enhanced green fluorescent protein (EGFP) cDNA under the control of a chicken β-actin promoter and cytomegalovirus enhancer in all tissues except of hair and erythrocytes 42 . Hence, tissue transplantation from these mice into wild-type animals allowed the differentiation of pre-existing GFP + microvessels originating from the tissue grafts and ingrowing GFP − microvessels from the surrounding host tissue 43 .
The mice were either housed four to six (peritoneal endometriosis model) or one per cage (dorsal skinfold chamber model) on wood chips as bedding in the conventional animal facility of the Institute for Clinical and Experimental Surgery. The animals had free access to tap water and standard pellet food (Altromin, Lage, Germany) and were maintained under a 12-h day/night cycle.
All Isolation of uterine tissue samples. To generate uterine tissue samples for in vitro analyses as well as for the induction of peritoneal endometriotic lesions, C57BL/6-TgN(ACTB-EGFP)1Osb/J donor mice in the stage of estrus were anesthetized by an intraperitoneal injection of 75 mg/kg ketamine (Pharmacia GmbH, Erlangen, Germany) and 15 mg/kg xylazine (Rompun ® ; Bayer, Leverkusen, Germany). After midline laparotomy, the uterine horns of the donor mice were isolated and placed into a petri dish containing Dulbecco's modified Eagle medium (DMEM (PAN Biotech, Aidenbach, Germany); 10% fetal calf serum, 100 U/mL penicillin, 0.1 mg/mL streptomycin; Thermo Fisher Scientific, Dreieich, Germany). The uterine horns were openend longitudinally and tissue samples with a diameter of 2 mm were carefully removed by means of a dermal biopsy punch (Stiefel Laboratorium GmbH, Offenbach am Main, Germany).
Endometrial fragments for the dorsal skinfold chamber model were also harvested from the uterine horns of anesthetized C57BL/6-TgN(ACTB-EGFP)1Osb/J mice in the stage of estrus. For this purpose, the two uterine horns were excised, placed in a petri dish containing DMEM (10% fetal calf serum, 100 U/mL penicillin, 0.1 mg/ mL streptomycin; PAN Biotech, Thermo Fisher Scientific) and opened longitudinally. Subsequently, the tissue was fixed and the perimetrium was carefully removed by means of microsurgical instruments under a stereomicroscope (M651; Leica Microsystems, Wetzlar, Germany). Circular fragments with a diameter of ~ 1.3 mm were then excised from the underlying endometrium.
Precultivation of uterine tissue samples. To  Protein expression of uterine tissue samples. To analyze the expression of angiogenesis-related proteins in uterine tissue samples, a proteome profiler mouse angiogenesis array kit was used according to the manufacturer´s instruction (R&D Systems, Wiesbaden, Germany). In brief, pooled uterine tissue samples were stored in lysis buffer (10 mM Tris pH 7.5, 10 mM NaCl, 0.1 mM ethylenediaminetetraacetic acid (EDTA), 0.5% Triton-X 100, 0.02% NaN 3 , 0.2 mM phenylmethylsulphonyl fluoride (PMSF), 1:75 v/v protease inhibitor cocktail and 1:100 v/v phosphatase inhibitor cocktail (all from Sigma-Aldrich; Taufkirchen, Germany)) and homogenized. The tissue lysate was then incubated for 30 min on ice and afterwards centrifuged at 4 °C for 5 min at 16,000×g. The supernatants were used for whole protein extracts. A total of 250 µg protein per group was used for the array. The samples were mixed with the biotinylated detection antibody cocktail and incubated for 1 h at room temperature. Subsequently, the mixture was exposed over night at 4 °C to the capture antibodies-spotted array membrane. The visualization of the labeled specific target proteins was achieved with streptavidin-horseradish peroxidase and chemiluminescent detection reagents using an Intas ECL Chemocam Imager (Intas Science Imaging Instruments GmbH, Göttingen, Germany).

Induction of peritoneal endometriotic lesions.
To investigate the effects of precultivation on the development of peritoneal endometriotic lesions, a 24-h-and 72-h-precultivated uterine tissue sample as well as a freshly isolated control sample were transplanted into the abdominal cavity of recipient C57BL/6J mice in the stage of estrus, as previously described 44 . For this purpose, the mice were anesthetized by an intraperitoneal injection of ketamine (75 mg/kg body weight; Pharmacia) and xylazine (15 mg/kg body weight; Rompun ® , Bayer). After midline laparotomy, the three tissue samples were randomly fixed with a 6-0 Prolene suture (Ethicon Products, Norderstedt, Germany) to the left and right abdominal wall. The laparotomy was then closed again with running 6-0 Prolene muscle and skin sutures. Dorsal skinfold chamber model. The mouse dorsal skinfold chamber model was used to further investigate the effect of precultivation on the vascularization of developing endometriotic lesions, as previously described 48,49 .

High
For the implantation of the dorsal skinfold chamber, C57BL/6J mice in the stage of diestrus were anesthetized by an intraperitoneal injection of ketamine (75 mg/kg body weight; Pharmacia) and xylazine (15 mg/kg body weight; Rompun ® , Bayer). Two titanum chamber frames (Irola Industriekomponenten, Schonach, Germany) were fixed on the dorsal skinfold of the back of the animals.The skin and muscle layers within the circular area of the observation window were removed. Thereafter, a cover glass was fixed by means of a snap ring in the observation window and the animals were allowed to recover from the anesthesia and surgical trauma for 48 h. Subsequently, the mice were anesthetized again and the cover glass of the observation window was removed. After rinsing the chamber tissue thoroughly, a 24-h-and 72-h-precultivated endometrial fragment as well as a freshly isolated control fragment were randomly placed on the striated muscle tissue within the dorsal skinfold chamber of each animal with a maximal distance to each other. Subsequently, the observation window was closed again with a new cover glass.
Intravital fluorescence microscopy. The vascularization and size of endometrial fragments was analyzed by means of intravital fluorescence microscopy directly after transplantation (d0) into the dorsal skinfold chamber as well as on day 3, 6, 10 and 14. For this purpose, the anesthetized mice received a single intravenous injection of 0.1 mL 5% fluorescein isothiocyanate (FITC)-labeled dextran (150,000 Da; Sigma-Aldrich) into the retrobulbar venous plexus. This enhanced the imaging contrast of blood vessels by staining of intravascular blood plasma. The mice were placed on a Plexiglas stage and the vascularization of the developing endometriotic lesions was visualized by means of a Zeiss Axiotech microscope (Zeiss, Oberkochen, Germany) equipped with a 100-W mercury lamp attached to a filter block for blue, green and ultraviolet light. The microscopic images were recorded by a charge-coupled device video camera (FK6990; Pieper, Schwerte, Germany) and transferred to a DVD system for off-line evaluation. The functional capillary density, i.e. the length of red blood cell (RBC)- www.nature.com/scientificreports/ perfused microvessels per observation area (cm/cm 2 ), and the size of the lesions (mm 2 ) were assessed by means of the software package CapImage (version 8.5; Zeintl, Heidelberg, Germany). At the end of the experiment, the transplanted tissue was harvested and fixed in formalin for further histological and immunohistochemical analyses.
Histology and immunohistochemistry. Formalin-fixed specimens of freshly isolated and precultivated uterine tissue samples as well as endometriotic lesions were embedded in paraffin. Three-µm-thick sections were cut and stained with hematoxylin and eosin (HE) according to standard procedures. In addition, sections of uterine tissue samples, which had been exposed to 0.2% trypan blue for 30 min before fixation, were counterstained with 0.1% nuclear fast red (Sigma-Aldrich). The sections were examined under a BX60 microscope (Olympus). The fraction of transplanted uterine tissue samples, which finally developed into endometriotic lesions containing endometrial stroma and glands, also referred to as take rate (%), was determined. These lesions were included in all further quantitative analyses.
For the immunohistochemical detection of proliferating and apoptotic cells in uterine tissue samples and endometriotic lesions, sections were stained with a rabbit polyclonal antibody against the proliferation marker Ki67 (1:500; Abcam, Cambridge, UK) and a rabbit polyclonal antibody against the apoptosis marker Casp-3 (1:100; New England Biolabs GmbH, Frankfurt, Deutschland). Moreover, sections of uterine tissue samples were stained with a rabbit polyclonal antibody against HIF-1α (1:50; Abcam). Additional sections of endometriotic lesions were stained with a rabbit polyclonal antibody against the lymphocyte marker CD3 (1:100; Abcam), a rabbit polyclonal antibody against the neutrophilic granulocyte marker myeloperoxidase (MPO) (1:100; Abcam), a rabbit polyclonal antibody against the pan-macrophage marker CD68 (1:100; Abcam), a rabbit polyclonal antibody against the M1 macrophage marker CD86 (1:100; New England Biolabs GmbH) and a rabbit polyclonal antibody against the M2 macrophage marker CD163 (1:200; Abcam). A goat anti-rabbit IgG biotinylated antibody (ready-to-use; Abcam) followed by avidin-peroxidase (1:50; Sigma-Aldrich) or a goat anti-rabbit peroxidaselabeled antibody (1:100; Abcam) served as secondary antibodies. 3-Amino-9-ethylcarbazole (AEC Substrate System; Abcam) was used as chromogen and counterstaining was performed with hemalaun. The fraction of proliferating, apoptotic and hypoxic cells (%) as well as the number of CD3 + lymphocytes, MPO + neutrophilic granulocytes, CD68 + macrophages, CD86 + M1 macrophages and CD163 + M2 macrophages (mm −2 ) was assessed by counting the numbers of positive cells in four regions of interest within the uterine tissue samples or endometriotic lesions. Additionally, the cell density was measured by counting all cell nuclei within the stroma of uterine tissue samples (mm −2 ).
For the immunofluorescent detection of microvessels, sections were stained with a monoclonal rat anti-mouse antibody against the endothelial cell marker CD31 (1:100; Dianova GmbH, Hamburg, Germany). A goat anti-rat IgG Alexa555 antibody (1:100; Invitrogen, Darmstadt, Germany) served as secondary antibody. Cell nuclei were stained with Hoechst 33342 (2 µg/mL; Sigma-Aldrich). The microvessel density (mm −2 ) was measured using the BZ-8000 microscope (Keyence). For this purpose, the overall number of CD31 + microvessels was counted and divided by the area of stromal tissue.
Experimental protocol. In a first set of experiments, a total of 143 uterine tissue samples from 12 C57BL/6-TgN(ACTB-EGFP)1Osb/J mice were isolated. One third of the samples was either precultivated for 24 h or 72 h, whereas the last third served as freshly isolated control. After the cultivation period, the tissue samples were directly processed for histology, immunohistochemistry and protein expression analyses.
In a second set of experiments, a total of 66 freshly isolated as well as 24-h-and 72-h-precultivated uterine tissue samples from 4 C57BL/6-TgN(ACTB-EGFP)1Osb/J mice were transplanted into 11 C57BL/6J mice, whereby one graft of each group was randomly fixed at either the right or left abdominal wall of each recipient animal. Ultrasound image analyses of the newly developing endometriotic lesions were repeatedly performed directly after transplantation (d0) as well as on days 7, 14, 21 and 28. At the end of the in vivo experiments, the size of the lesions was assessed by means of a digital caliper. Thereafter, the lesions were excised and further processed for histology and immunohistochemistry.
In a third set of experiments, a total of 27 freshly isolated as well as 24-h-and 72-h-precultivated endometrial fragments from 4 C57BL/6-TgN(ACTB-EGFP)1Osb/J mice were transplanted into 9 dorsal skinfold chambers of C57BL/6J mice. Intravital fluorescence microscopy was performed directly after transplantation (d0) as well as on days 3, 6, 10 and 14. At the end of the experiment, the grafts were excised and further processed for histology. Statistical analysis. Data were first analyzed 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. Differences between three experimental groups were assessed by one-way ANOVA followed by the Scientific Reports | (2022) 12:5625 | https://doi.org/10.1038/s41598-022-09577-z www.nature.com/scientificreports/ Student-Newman-Keuls post hoc test (SigmaPlot 13.0; Jandel Corporation, San Rafael, CA, USA). All data are given as mean ± standard error of the mean (SEM). Statistical significance was accepted for P < 0.05.

Data availability
The datasets generated and analysed during the current study are available from the corresponding author on request.