{"paper_id":"0dce3be1-56ea-489c-9cd7-ac3e28bcd91c","body_text":"1Scientific  RepoRts  |          (2019) 9:7037  | https://doi.org/10.1038/s41598-019-43185-8\nwww.nature.com/scientificreports\nVEGF Receptor 1-Expressing \nMacrophages Recruited from Bone \nMarrow Enhances Angiogenesis in \nEndometrial Tissues\nKazuki Sekiguchi1,2,3, Yoshiya Ito1,2, Kyoko Hattori1,2,3, Tomoyoshi Inoue1,2, Kanako Hosono1,2, \nMasako Honda1,2,3, Akiko Numao3, Hideki Amano1,2, Masabumi Shibuya4, Nobuya Unno3 & \nMasataka Majima1,2\nAngiogenesis is critical in maintenance of endometrial tissues. Here, we examined the role of \nVEGF receptor 1 (VEGFR1) signaling in angiogenesis and tissue growth in an endometriosis model. \nEndometrial fragments were implanted into the peritoneal wall of mice, and endometrial tissue \ngrowth and microvessel density (MVD) were determined. Endometrial fragments from wild-type (WT) \nmice grew slowly with increased angiogenesis determined by CD31\n+ MVD, peaking on Day 14. When \ntissues from WT mice were transplanted into VEGFR1 tyrosine kinase-knockout mice, implant growth \nand angiogenesis were suppressed on Day 14 compared with growth of WT implants in a WT host. \nThe blood vessels in the implants were not derived from the host peritoneum. Immunostaining for \nVEGFR1 suggested that high numbers of VEGFR1\n+ cells such as macrophages were infiltrated into the \nendometrial tissues. When macrophages were deleted with Clophosome N, both endometrial tissue \ngrowth and angiogenesis were significantly suppressed. Bone marrow chimera experiments revealed \nthat growth and angiogenesis in endometrial implants were promoted by host bone marrow-derived \nVEGFR1\n+/CD11b+ macrophages that accumulated in the implants, and secreted basic fibroblast growth \nfactor (bFGF). A FGF receptor kinase inhibitor, PD173047 significantly reduced size of endometrial \ntissues and angiogenesis. VEGFR1 signaling in host-derived cells is crucial for growth and angiogenesis \nin endometrial tissue. Thus, VEGFR1 blockade is a potential treatment for endometriosis.\nEndometriosis, characterized by extra-uterine growth of endometrial tissue, is a common gynecological dis-\nease in women of reproductive age\n1. The most common symptoms are pelvic pain and infertility, both of which \nhave an adverse effect on quality of life2,3. Endometriosis is highly dependent on angiogenesis4 (the formation of \nnew blood vessels from pre-existing vessels), which is critical for both normal development and homeostasis as \nwell as for certain pathological conditions\n5,6. There are numerous endogenous factors that regulate angiogenesis; \nhowever, vascular endothelial growth factor (VEGF) and its receptors are the prime regulators of both physio-\nlogical and pathological angiogenesis\n7,8. The VEGF pathway plays a critical role in ischemic angiogenesis and \ntumor growth via receptor signaling-dependent mechanisms9–11. The most common isoform of VEGF , VEGF-A, \nbinds to two receptor tyrosine kinases: VEGF receptor 1 (VEGFR1) and VEGF receptor 2 (VEGFR2). VEGFR2 is \nexpressed mainly by endothelial cells, whereas VEGFR1 is also expressed by hematopoietic stem cells and inflam-\nmatory cells, such as monocytes and macrophages, and regulates chemotaxis\n12–14. VEGFR1 binds VEGF-A with \nan affinity approximately ten times that of VEGFR2; however, the underlying biological mechanism is not fully \nunderstood. VEGFR2-null mice fail to develop blood vessels and die in utero, indicating that VEGFR2 signaling \nis essential for development of the vascular system\n15. By contrast, VEGFR1-null mice exhibit overgrowth and \n1Department of Pharmacology, Kitasato University School of Medicine, Sagamihara, Kanagawa, Japan. 2Department \nof Molecular Pharmacology, Graduate School of Medical Sciences, Kitasato University, Sagamihara, Kanagawa, \nJapan. 3Department of Obstetrics and Gynecology, Kitasato University School of Medicine, Sagamihara, Kanagawa, \nJapan. 4Gakubunkan i nstitute of Physiology and Medicine, Jobu University, t akasaki, Gunma, Japan. Kazuki \nSekiguchi and Yoshiya ito contributed equally. correspondence and requests for materials should be addressed to \nM.M. (email: mmajima@med.kitasato-u.ac.jp)\nReceived: 30 November 2018\nAccepted: 8 April 2019\nPublished: xx xx xxxx\nopeN\n\n\n2Scientific  RepoRts  |          (2019) 9:7037  | https://doi.org/10.1038/s41598-019-43185-8\nwww.nature.com/scientificreportswww.nature.com/scientificreports/\ndisorganization of blood vessels, suggesting that VEGFR1 is a negative regulator of angiogenesis during embry-\nonic development.\nHowever, when we generated transgenic mice expressing a variant of VEGFR1 that lacks the tyrosine kinase \ndomain (VEGFR1TK− /− ), the mice appeared healthy and showed normal blood vessel formation16. Expression \nof VEGF and VEGF receptors (VEGFRs) increases during the healing of wounds and gastric ulcers and during \nrecovery from ischemia\n17,18. Indeed, we recently showed that VEGFR1 signaling facilitates angiogenesis during \nrecovery from ischemia and gastric ulcers 17,18. Under these pathological conditions, VEGFR1-expressing cells \nincrease angiogenesis, leading to speedier recovery from ischemia and tissue damage.\nWe previously reported that endogenous prostaglandin (PG) E 2 plays a role in the growth of endometrial \ntissue and angiogenesis in a mouse transplantation model by inducing VEGF production 19. We also found that \nsignaling via PGE 2 increased VEGF-dependent angiogenesis during chronic inflammation, and in the tumor \nmicroenvironment10,20–22. PG receptor signaling-mediated increases in cAMP levels facilitate both angiogenesis \nand VEGF production23. Further, we found that VEGF neutralizing antibody treatment reduced the growth of \nendometrial tissue and angiogenesis in a mouse transplantation model24. The implantation of endometrial tissues \nisolated from wild type (WT) mice to the peritoneal cavity in VEGFR1TK−/−  mice showed reduced angiogenesis \nand growth of implants, suggesting that host VEGFR1 signaling is critical for the maintenance of implants 24. \nHowever, it remains unknown how VEGFR1 signaling regulates angiogenesis and development of endometriosis.\nIn the present study, we clarified that VEGFR1 signaling in host-derived cells, especially CD11b+ macrophages \nplays a role in growth and angiogenesis in endometrial tissues. This study suggests that blocking VEGFR1 with \nantibodies or small molecule kinase inhibitors may become a promising option to the treatment of endometriosis.\nMaterials and Methods\nAnimals. Eight-week-old female C57BL/6 wild-type (WT) mice were purchased from CLEA Japan (Tokyo) \nand used as controls in experiments involving 8-week-old female VEGFR1TK − /− , which were developed pre-\nviously (Recombinant DNA Experiment Approve Number 3937) 16. The knockout mice were backcrossed to a \nC57BL/6 background for more than ten generations. Green fluorescent protein transgenic C57BL/6 (GFP+TG) \nmice were also generated in-house (Recombinant DNA Experiment Approve Number 3937). TK − /−  mice and \nGFP+TG mice were crossed to obtain GFP +/+TK− /−  mice (GFP+TK− /−  TG)17. All mice were housed in a lim-\nited access animal facility with a temperature maintained at 25 ± 1 °C and relative humidity at 60 ± 5%. A 14 h \nlight/10 h dark (6 AM to 8 PM) cycle was established using artificial lighting. All experimental procedures were \napproved by the Animal Experimentation and Ethics Committee of the Kitasato University School of Medicine \n(1114, 2015–022), and were performed in accordance with the guidelines for animal experiments set down by \nthe Kitasato University School of Medicine, which are in accordance with the “Guidelines for Proper Conduct of \nAnimal Experiments” published by the Science Council of Japan. Mice used for survival studies were examined \nby animal care takers and the overall health status was checked by trained professionals. Mice were euthanized \nby pentobarbital sodium when they were found in a moribund state as identified by inability to maintain upright \nposition and/or labored breathing. The mice for in vivo experiments were constantly checked daily throughout \nthe experiment periods. Drugs were given under inhalation anesthesia with isoflurane. Tissue collection proce-\ndures were performed under anesthesia with pentobarbital sodium. At the end of the experiments, the animals \nwere euthanized by exsanguination under anesthesia with pentobarbital sodium followed by cervical dislocation.\nBone marrow transplantation. Bone marrow transplantation was performed as previously described 25. \nBriefly, donor bone marrow cells were harvested from GFP+TG or GFP+TK− /−  TG mice, bone marrow mono-\nnuclear cells were isolated by filtration through nylon mesh filter, and the mononuclear cells were transplanted \ninto irradiated WT mice via the tail vein. GFP\n+TG bone marrow-transplanted mice were named GFP+WT BM \nchimeric (BMC) mice (n = 12). GFP+TK−/−  TG bone marrow-transplanted mice were named GFP+TK−/−  BMC \nmice (n = 12). After 6–8 weeks of bone marrow transplantation, peripheral blood from mice was collected via tail \nvein. Mononuclear cells were obtained from whole blood by Lymphosepar II (Immuno-Biological Laboratories, \nFujioka). FACS analysis for the peripheral leukocytes was performed on FACS Calibur (BD Biosciences, Franklin \nLakes, NJ, USA). Mice in which more than approximately 90% of the peripheral leukocytes were GFP-positive \nwere used for the experiments.\nEndometrial transplantation model.  Endometrial transplantation was performed as previously \ndescribed (Fig.  1)24,26. Briefly, donor and recipient mice were bilaterally ovariectomized through paravertebral \nincisions to exclude endogenous estrogen and menstrual cycle. All donor and recipient mice received subcutane-\nous (s.c.) injections of estradiol dipropionate (100 mg/kg) in sesame oil (Obahormone depot; Aska, Tokyo) every \nweek from the time of ovariectomy 24,27. Seven days after ovariectomy, the uterine horns from the donor were \nremoved, trimmed of connective tissue, and opened longitudinally in a tissue culture dish containing Dulbecco’s \nmodified Eagle’s medium F-10 (Gibco, Grand Island, NY) at 37 °C, supplemented with 100 U/mL penicillin and \n100 mg/mL streptomycin (Gibco, Grand Island, NY). Four round endometrial fragments (3  mm in diameter), \nwhich include the myometrium, were collected using a biopsy punch (Kai medical, Japan). The endometrial tis-\nsues were transplanted to the peritoneal wall of recipient mice with a 7-0 polypropylene suture (Ethicon, Johnson \n& Johnson, Japan), as described previously (Fig. 1)\n24,26; this location was chosen because it is in contact with the \nendometrial surface epithelium of the implants and peritoneum. Endometrial fragments from WT or TK− /−  mice \nwere implanted ectopically into the peritoneum of either WT or TK− /−  mice. The wound was closed with a 3-0 \nsuture and mice were placed on a warming carpet to prevent hypothermia. The day of implantation was defined \nas Day 0, and mice were euthanized under anesthesia on Days 7, 14, 21, or 28 post-implantation. The endometrial \nimplants were removed and captured by taking digital photographs.\n\n3Scientific  RepoRts  |          (2019) 9:7037  | https://doi.org/10.1038/s41598-019-43185-8\nwww.nature.com/scientificreportswww.nature.com/scientificreports/\nThe captured digital images were uploaded to a computer and opened with ImageJ image analysis software. \nThe implant outline was defined from the photographic image. Following tracing, the areas of the implants were \ncalculated by ImageJ image analysis software. The results were expressed as the size of the implants per mm2. The \nfour implants obtained from an individual recipient mouse were randomly assigned to experimental analyses; \nOne of them was prepared for gene expression which examined by real-time reverse transcription-polymerase \nchain reaction (RT-PCR). The other one was used for immunohistochemistry. The rest of the two were prepared \nfor immunofluorescence. All histological samples were first fixed in 4% formaldehyde in 0.1 M sodium phosphate \nbuffer (pH 7.4) at 4 °C for 24 h for analyses. When implants from WT mice were transplanted into host WT mice, \nwe expressed the transplants of WT; Implant →  WT; Host combination as WT →  WT. Using WT mice and TK\n− /−  \nmice, we created four different cross transplantation experimental groups; WT  →  WT (n = 13), TK− /−  →  WT \n(n = 12), WT → TK −/−  (n = 12), and TK−/−  → TK −/−  (n = 12).\nDeletion of macrophages with Clophosome. Recipient mice were injected intraperitoneally (i.p.) with \n0.7 mg of Clophosome N (F70101C-N; FormuMax Scientific, Palo Alto, CA, USA) per mouse (n = 4) or control \nliposomes (F70101-N) (n = 4) every four days starting at the Day 0 implantation (Fig. 1).\nAdministration of an inhibitor of FGF.  Recipient mice received an intraperitoneal (i.p.) injection of \nPD173047 (25 mg/kg/day, Selleck Chemicals, Houston, TX) every day for 2 weeks starting at the Day 0 implan-\ntation (n = 8 per group) (Fig. 1). Control mice received PBS (n = 8 per group). PD173047 is a selective inhibitor \nfor FGF receptor 1 (FGFR1), and PD173047 also inhibits bFGF (FGF-2) induced cell growth and proliferation28,29\nImmunohistochemical analysis.  After fixation in 4% formaldehyde, tissues were embedded in paraffin. \nSections (3 μm thick) were cut using a sliding microtome and dewaxed in xylene, and endogenous peroxidases \nwere quenched by incubation in 3% H2O2 buffer. Antigen retrieval was performed by heating sections in 0.01 M \nsodium citrate buffer (pH 6.0) in a microwave oven. The sections were then incubated at 4 °C overnight with a \npolyclonal rabbit anti-CD31 antibody (1:800; Ab28364; Abcam, Cambridge, MA, USA). After washing in phos-\nphate buffer solution (PBS), sections were stained with conjugated secondary antibody (Histofine Simple Stain \nMAX PO; Nichirei Bioscience, Tokyo), washed again, and stained with DAB (dimethylaminoazobenzene) for \napproximately 2 minutes. Finally, sections were counterstained with Mayer’s hematoxylin. Control sections were \ntreated with isotype-matched control IgG.\nImmunofluorescence analysis. Fixed samples of endometriotic lesions were then embedded in OCT com-\npound (Sakura Finetek U.S.A., Inc., Torrance CA) and frozen at − 80 °C before 8 μm sections were cut using a \ncryostat. The OCT compound was removed by washing in PBS, and the sections were incubated in 1% bovine \nserum albumin (BSA)/PBS at room temperature for 1 h overnight at 4 °C to block non-specific binding. Next, \nsections were incubated with the following primary antibodies at 4 °C overnight: polyclonal rabbit anti-VEGFR1 \n(1:200; abcam2350; Abcam, Cambridge, MA, USA), polyclonal goat anti-VEGFR1 (1:200; sc-316-g; Santa \nCruz Biotechnology, Santa Cruz, CA, USA), polyclonal rabbit VEGF-A (1:100, ab46154; Abcam), monoclo-\nnal rat anti-CD31 (1:200; BD550274; BD Biosciences, Franklin Lakes, NJ, USA), monoclonal rat anti-CD11b \nFigure 1. Experimental protocols for experimental endometriosis. Both donor and recipient mice were treated \nwith estradiol (E). In some experiment, recipient mice were treated with Clophosome N (C) or PD173074. \nTissue samples for analyses were collected at the indicated time.\n\n4Scientific  RepoRts  |          (2019) 9:7037  | https://doi.org/10.1038/s41598-019-43185-8\nwww.nature.com/scientificreportswww.nature.com/scientificreports/\n(1:200; BD550282; BD Biosciences), polyclonal goat anti-S100A4 (1:200; TA318024; OriGene Technologies, \nRockville, MD, USA), or polyclonal rabbit anti-bFGF (1:200; ab72316; Abcam). After washing in PBS, the sec-\ntions were incubated with the following secondary antibodies (all at 1:200) for 1 h at room temperature: Alexa \nFluor 488-conjugated donkey anti-rabbit IgG, Alexa Fluor 594-conjugated donkey anti-rabbit IgG, Alexa Fluor \n594-conjugated donkey anti-rat IgG, Alexa Fluor 594-conjugated donkey anti-goat IgG, and/or Alexa Fluor \n647-conjugated donkey anti-rabbit IgG. Control sections were incubated in isotype-matched controls for mon-\noclonal antibodies. Images were observed and captured under a confocal scanning laser microscope (LSM710; \nCarl Zeiss, Jena, Germany; ×400 magnification) or a fluorescence microscope (Biozero BZ-9000; Keyence, Osaka; \n×400 magnification)\n30. Positive cells were quantified randomly from 4 fields at ×400 magnification per mouse.\nDetermination of vessel density. Microvessel density (MVD) in areas showing the most intense neovascular-\nization (hot spots) within the endometrial implants was used as a measure of angiogenesis, as previously described24,31. \nBriefly, blood vessels in the ectopic endometrium were stained with an anti-CD31 antibody and areas showing the \nhighest levels of neovascularization were identified by scanning the endometrial tissues at low power (×40 and ×100 \nmagnification). Individual microvessels within the area of maximum neovascularization were counted in one ×400 \nfield. We determined MVD in the peritoneum to muscle layer, which lies just below the endometrial implant, and in \nthe distant peritoneum 5 mm from the peritoneum at which implants were transplanted. CD31\n+ endothelial cells were \nclearly differentiated from the adjacent microvessels, stromal cells, and other connective tissue elements. MVD was \nexpressed as the mean of blood vessels in three high-power-fields (150 μm × 150 μm).\nIsolation of cells from implants. In another set of experiment, mice in the WT →  WT (n = 4) and \nTK− /−  →  TK− /− (n = 4) were anesthetized with pentobarbital sodium solution (60 mg/kg, i.p.), and the excised implants \nwere placed immediately at room temperature in RPMI, minced into small pieces using scissors, and incubated in \nRPMI containing 0.05% collagenase (Type IV; Sigma Chemical Co., St. Louis, MO, USA) at 37 °C for 20 min. The tissue \nwas then pressed through a 70 μm cell strainer. The cells were centrifuged at 2600 rpm for 10 min at 4 °C, and pelleted \ncells were resuspended in PBS. Leukocytes were isolated from the homogenates by density-gradient centrifugation on \n33% Percoll\n™ (GE Healthcare Life Sciences, Piscataway, NJ, USA), as previously reported32. Non-parenchymal cells \nwere collected from the interface between the 33% and 66% Percoll™ density cushions and centrifuged at 2700 rpm \nfor 30 min at 4 °C. Viable, nucleated cells were counted by trypan blue exclusion and diluted to a uniform cell density.\nFlow cytometry analysis. Cells were incubated with the 2.4G2 mAb (anti-cγRIII/II) to block non-specific \nbinding of the primary mAb. Then, cells were stained with a combination of the following fluorochrome-conjugated \nantibodies: anti-CD11b (clone M1/70, BioLegend, San Diego, CA, USA), anti-CD34 (clone MEC14.7, BioLegend) \nand anti-CD133 (clone 315-2C11, BioLegend). Samples were measured on a FACSVerse\n™ (BD, Franklin Lakes, NJ, \nUSA). The data were analyzed using Kaluza software v1.3 (Beckman Coulter, Brea, CA, USA)33.\nQuantitative real-time RT-PCR analysis. Total RNA was isolated from endometriotic tissues using \nTRIzol reagent (Life Technologies, Grand Island, NY , USA), according to the manufacturer’s instructions. \nRT-PCR and real-time PCR were performed to measure CD31, VEGF-A, bFGF , cTGF , EGF , TGF-ß, Ang-1, \nAng-2, and human glyceraldehyde-3-phosphate dehydrogenase (GAPDH) mRNA expression, as previously \ndescribed\n34.\nThe following primer sequences were used:\n CD31, 5′-CAGAGCCAGCAGTATGAGGAC-3′ (forward) and 5′ -GCAACTATTAAGGTGGCGATG-3′ \n(reverse);\n VEGF-A, 5′-ACGACAGAAGGAGAGCAGAAG-3′ (forward) and 5′-ATGTCCACCAGGGTCTCAATC-3′ \n(reverse);\n bFGF , 5′-GGCTGCTGGCTTCTAAGTGTG-3′  (forward) and 5′ -TTCCGTGACCGGTAAGTATTG-3′ \n(reverse);\n CTGF , 5′-AACCGGGGAGGGAAATTATAG-3′ (forward) and 5′ -TGGAATCAGAATGGTCAGAGG-3′ \n(reverse);\n EGF , 5′-ATGGGAAACAATGTCACGAAC-3′  (forward) and 5′ -CATCTCTCCCAAGCACTGAAC-3′ \n(reverse);\n TGF-ß, 5′-TGTATTCCGTCTCCTTGGTTC-3′  (forward) and 5′ -AACAATTCCTGGCGTTACCTT-3′ \n(reverse); \n Ang-1, 5′-TGAAGGAGGAGAAAGAAAACC-3′ (forward) and 5′-GGATGCTGTTGTTGTTGGTAG-3′ \n(reverse);\n Ang-2, 5′-TACACACTGACCTTCCCCAAC-3 ′ (forward) and 5 ′-AGTCCACACTGCCATCTTCTC-3 ′ \n(reverse); and\n GAPDH, 5′-ACATCAAGAAGGTGGTGAAGC-3′ (forward) and 5′-AAGGTGGAAGAGTGGGAGTTG-3′ \n(reverse).\nCell culture. Bone marrow-cells were isolated from the femur and tibia of 8-week-old WT mice (n = 4) and \nTK− /−  mice (n = 4)32. Femurs and tibias of mice were flushed with PBS, and erythrocytes were lysed by treatment \nwith RBC lysis buffer (BioLegend). For the generation of bone marrow-derived macrophages, bone marrow cells \nwere cultured in RPMI 1640 medium containing 10% fetal calf serum and macrophage colony stimulating factor \n(M-CSF) (20 ng/ml, BioLegend) plated in 6-well plates (1.0 × 10\n6 cells per well). At day 7, cells were either left  \nuntreated or treated with recombinant murine placental growth factor (PlGF) (BioVision, Inc., CA, USA) in \nRPMI 1640 medium for 6 hours. Bone marrow-derived macrophages were then harvested and homogenized in \nTRIzol (Life Technologies), and mRNA levels were measured by real-time RT-PCR.\n\n5Scientific  RepoRts  |          (2019) 9:7037  | https://doi.org/10.1038/s41598-019-43185-8\nwww.nature.com/scientificreportswww.nature.com/scientificreports/\nData analyses. Data were expressed as the mean ± standard error of the mean (SEM). All statistical analyses were \nperformed using JMP 10 (SAS Institute, Cary, NC, USA). For statistical evaluations a normality test and a variance test \nwere done. Data that were normally distributed were analyzed using the parametric tests. Comparisons between the \ntwo groups were performed using Student’s t-test. One-way analysis of variance (ANOV A) followed by Tukey-Kramer \npost-hoc test was used to compare data among multiple groups. Student’s t-tests were applied to the analyses for the \norigin of the vasculature, pro-angiogenic factors, treatment with chlophosome, PD173047, cultured cells response, \nand flow cytometry. The changes in size and angiogenesis in the endometrial implants were compared with one-way \nANOV A with Tukey-Kramer post-hoc test. A P value < 0.05 was considered statistically significant. The data for PCR \nand histology were collected from one implant per mice. The sizes of multiple implants obtained from an individual \nmouse were averaged, and the averaged value per mice was compared for analysis. For analysis of flow cytometry, we \ncollected four endometrial implants from an individual mouse, and combined into a single sample.\nResults\nHost VEGFR1 signaling is critical in maintenance of endometrial tissues and angiogenesis.   \nWhen WT endometrial fragments were implanted into estrogen-stimulated WT mice, the implanted endome-\ntrial tissues grew gradually. Growth peaked at Day 14 post-implantation (Day 0: 6.56 ± 0.18 mm\n2 vs. Day 14: \n10.07 ± 0.51 mm2, P = 0.0029), and the size of the implants decreased thereafter (Day 21: 9.21 ± 0.54 mm2, Day \n28: 7.92 ± 0.72 mm2) (Supplementary Fig. S1a,b). When we stained the endometrial tissues with an anti-CD31 \nantibody, we found that the density of neovascularized blood vessels in transplanted tissues at Day 14 was higher \nthan that in naïve endometrial tissues (Day 0: 5.39 ± 0.31/2.25 × 10\n4 μm2, Day 14: 6.95 ± 0.34/2.25 × 104 μm2, \nP = 0.025) (Supplementary Fig. S1c,d). These results were essentially the same as those in our previous report24.\nTo estimate the role of host VEGFR1 signaling, we implanted WT or TK − /−  endometrial tissues into the \nperitoneal cavities of WT or TK − /−  mice (Fig. 2a,b). When TK− /−  endometrial fragments were implanted into \nthe WT peritoneal cavity (TK − /−  →  WT), the growth of the implants at Day 14 was not different from that of \nWT →  WT (TK− /−  →  WT, 9.43 ± 0.75 mm 2 vs. WT →  WT, 10.16 ± 0.55 mm 2, P = 0.80) (Fig.  2a). By con-\ntrast, the WT →  TK− /−  led to significant growth suppression at Day 14 when compared with the WT →  WT \n(WT →  TK− /− , 7.23 ± 0.42 mm2 vs. WT →  WT, 10.16 ± 0.55 mm2, P = 0.004) (Fig.  2a). Similar results were \nobserved with the TK − /−  →  TK− /−  (TK − /−  →  TK− /− , 5.99 ± 0.55 mm 2 vs. WT →  WT, 10.16 ±  0.55 mm 2, \nP < 0.0001) (Fig. 2a). We confirmed our previous findings24, and suggested that the growth of endometrial frag-\nments in estrogen-stimulated mice was promoted by host VEGFR1 signaling.\nWhen the number of CD31+ vessels in the endometrial tissue implants were counted24, the density of CD31+ vessels \nin the WT →  WT increased over time (Supplementary Fig. S1d), suggesting that angiogenesis was induced. At Day 14, \nwe found that angiogenesis in the implants in the TK− /−  →  WT was similar to that in the implants in the WT →  WT \n(WT →  TK− /− , 5.80 ± 0.27/2.25 × 104 μm2 vs. WT →  WT, 6.95 ± 0.34/2.25 × 104 μm2, P = 0.06) (Fig. 2b); however, \nangiogenesis in the WT →  TK− /−  was significantly lower than that in the implants in the WT →  WT (WT →  TK− /− , \n4.63 ± 0.30/2.25 × 104 μm2 vs. WT →  WT, 6.95 ± 0.34/2.25 × 104 μm2, P < 0.0001) (Fig. 2b). The same results were \nobserved for the TK− /−  →  TK− /−  (Fig. 2b). These results were essentially the same as the previous report24. This suggests \nthat host-derived VEGFR1-expressing cells/tissues induce proangiogenic responses in implanted endometrial tissues.\nOrigin of the vasculature in endometrial implants. As mentioned above, signaling via host-derived \nVEGFR1 is critical for the growth of endometrial tissues and for neovascularization of endometrial implants. \nTherefore, we next examined angiogenic responses in the parietal peritoneum that made contact with the \nimplants, since angiogenesis in the parietal peritoneum may increase the growth of endometrial tissues by \nincreasing the supply of oxygen and nutrients. Measurement of MVD in the parietal peritoneum at Day 14 \npost-implantation (Fig.  2c) revealed that vessel density in the WT →  TK\n− /−  (3.87 ± 0.55/2.25 × 104 μm2), \nTK− /−  →  WT (3.84 ± 0.48/2.25 × 104 μm2), and TK − /−  →  TK− /−  (4.41 ± 0.53/2.25 × 104 μm2) was similar to \nthat in the WT →  WT (3.96 ± 0.37/2.25 × 104 μm2). In addition, MVD in the distant peritoneum from the pari-\netal peritoneum of mice bearing endometrial implants (WT  →  WT, 2.18 ± 0.19/2.25 × 104 μm2; TK− /−  →  W T, \n2.25 ± 0.24/2.25 × 104 μm2; WT →  TK− /− , 2.00 ± 0.24/2.25 × 104 μm2; TK− /−  →  TK− /− , 2.23 ± 0.17/2.25 × 104 \nμm2) was lower than that in the parietal peritoneum just below the implants (vs. WT →  WT, 3.96 ± 0.37/2.25 × 104 \nμm2, P = 0.012; TK− /−  →  WT, 3.84 ± 0.48/2.25 × 104 μm2, P = 0.049; WT →  TK− /− , 3.87 ± 0.55/2.25 × 104 μm2, \nP = 0.024; TK − /−  →  TK− /− , 4.41 ± 0.53/2.25 × 104 μm2, P = 0.071) (Fig.  2c). These suggest that angiogenic \nresponses in the parietal peritoneum were independent of VEGFR1.\nTherefore, we next examined the origin of the vessels in WT →  GFP+ TG (Fig. 3a). CD31+ vessel-like struc-\ntures were identified in the WT implant (box I, upper panels in low-power field), the granulation tissue that \nformed at the margins of the implants (box II, upper panels), and the host parietal peritoneum (box III, upper \npanels) at lower magnification. GFP imaging revealed that the host parietal peritoneum was strongly GFP\n+, and \nthat some GFP+ cells had infiltrated areas I and III. This suggests that host-derived cells were infiltrated to the \nimplants and granulation tissue formed at the margins of the implants. When we examined MVD in areas I, II, \nand III at higher magnification, we found that the CD31\n+ vessel-like structures in areas II and III were strongly \nGFP+ (areas II: GFP+ MVD, 5.72 ± 0.86/2.25 × 104 μm2 vs. GFP−  MVD, 1.60 ± 0.54/2.25 × 104 μm2, p = 0.0006, \nareas III: GFP+ MVD, 7.08 ± 0.90/2.25 × 104 μm2 vs. GFP−  MVD, 0/2.25 × 104 μm2, p < 0.0001) (Fig. 3b),whereas \nthose in area I showed slightly GFP + (areas I: GFP + MVD, 0.79 ± 0.15/2.25 × 104 μm2 vs. GFP −  MVD, \n6.58 ± 0.40/2.25 × 104 μm2, p < 0.0001) (Fig. 3b). These results suggest that the blood vessels in the implants were \nnot derived from the host, and that few vessels were sprouting from host tissues, even though angiogenesis in the \nimplants was modulated by host-derived VEGFR1.\nWhen we examined expression of VEGFR1 in implants in the WT → WT (Supplementary Fig. S2), almost all \n(>97.5%) of VEGFR1-positive cells were positive for CD31, CD11b, and S1004A, a marker for fibroblasts\n35 (% \nof VEGFR1+ cells in CD31+ cells: 97.6 ± 0.9; % of VEGFR1+ cells in CD11b+ cells, 98.4 ± 0.7; % of VEGFR1+ \n\n6Scientific  RepoRts  |          (2019) 9:7037  | https://doi.org/10.1038/s41598-019-43185-8\nwww.nature.com/scientificreportswww.nature.com/scientificreports/\ncells in S100A4+ cells: 99.2 ± 0.3) (Supplementary Fig. S2). As mentioned above (Fig. 3a), since the CD31+ vessels \nin the implants may not be derived from the host, VEGFR1-expressing CD11b+ cells and/or S1004A+ cells may \nfacilitate angiogenesis via VEGF .\nBone marrow-derived VEGFR1+ cells facilitate both growth and angiogenesis in endometrial \nimplants. As mentioned above, most CD11b/S100A4+ cells were also VEGFR1+. Therefore, we hypothesized \nthat the VEGFR1+ cells that infiltrated into the implants were recruited from the bone marrow. To test this, we \ngenerated bone marrow chimera (BMC) mice and examined growth and angiogenesis in endometrial tissues.\nWhen GFP + WT endometrial tissues were implanted into WT mice, GFP + cells were restricted to the \nimplanted tissues and did not infiltrate the host parietal peritoneum (Fig.  4a, left panel). By contrast, when \nWT implants were implanted into GFP transgenic WT mice, a large number of GFP + cells accumulated in the \nimplants (Fig. 4a, middle panel). Transplantation of WT endometrial tissues into GFP transgenic WT BMC mice \nrevealed that GFP+ cells accumulated in the implants (Fig. 4a, right panel). These results suggest that host cells, \nincluding bone marrow-derived cells, accumulate in the implants during growth and angiogenesis.\nTo examine the phenotype of the cells that accumulated in the implants, we next implanted non-GFP WT \nendometrial tissues into GFP transgenic WT BMC mice (WT →  GFP+WT BMC). Tissues were then removed \nat Day 14 and stained with anti-CD31, anti-CD11b, and anti-S1004A antibodies (Fig.  4b). CD31+ vessel-like \nstructures in the implants were also VEGFR1+; however, they were GFP− , suggesting that the blood vessels in the \nimplants did not comprise bone marrow-derived cells (Fig. 4b). In addition, none of the cells were GFP+/CD31+/\nVEGFR1+ (% of GFP+/CD31+ cells among VEGFR1+ cells, 0 vs. % of GFP− /CD31+ cells among VEGFR1+ cells, \n3.5 ± 1.1, P = 0.0265) (Fig. 4c). By contrast, most of the CD11b+ cells among the VEGFR1+ cell populations in \nthe implants were GFP + (% of GFP+/CD11b+ cells among VEGFR1 + cells, 11.3 ± 4.1 vs. % of GFP − /CD11b+ \ncells among VEGFR1 + cells, 3.5 ± 0.8, P = 0.0105) (Fig. 4c). When we stained implants with an anti-S1004A \nantibody, we observed accumulation of S1004A+/VEGFR1+ cells (Fig. 3b); however, the majority of these cells \nFigure 2. Host VEGFR1 signaling plays a role in growth in endometrial tissues and angiogenesis (a) Size \nof endometrial implants and (b) microvessel density at Day 14. Dotted line denotes the mean size of the \nendometrial lesion and the mean microvessel density from the WT → WT at Day 0. Data are expressed as the \nmean ± SEM (n = 11‒13 mice). ***P < 0.001 (one-way ANOV A) in comparison with WT → WT at Day 14.(c) \nMicrovessel density in the perimetrium to muscle layer, which lies just below the endometrial implant, and in \nthe distant peritoneum at Day 14. Data are expressed as the mean ± SEM (n = 9‒12 mice). *P < 0.05, **P < 0.01 \n(one-way ANOV A).\n\n7Scientific  RepoRts  |          (2019) 9:7037  | https://doi.org/10.1038/s41598-019-43185-8\nwww.nature.com/scientificreportswww.nature.com/scientificreports/\nwere GFP−  (% of GFP+/S100A4+ cells among VEGFR1 + cells, 16.9 ± 1.7 vs. % of GFP− /S100A4+ cells among \nVEGFR1+ cells, 5.8 ± 1.2, P = 0.0003) (Fig. 4c). Thus, a major population of GFP+ cells was CD11b+ rather than \nS1004A+ (Fig. 4b,c).\nFurther we examined the functional relevance of VEGFR1-expressing cells recruited from the bone marrow in \nterms of growth and angiogenesis in endometrial tissues (Fig. 4d,e). When we implanted WT endometrial tissues \ninto TK− /−  BMC mice, the growth of the endometrial implants at Day 14 was more suppressed than that of WT \nimplants in WT BMC mice (WT →  GFP+WT BMC: 8.82 ± 0.60/2.25 × 104 μm2 vs. WT →  GFP+TK− /−  BMC: \n7.34 ± 0.45/2.25 × 104 μm2, P = 0.031) (Fig. 4d). The same was true for angiogenic responses (WT →  GFP+WT \nBMC: 9.61 ± 0.53/2.25 × 104 μm2 vs. WT → GFP +TK−/−  BMC: 5.61 ± 0.35/2.25 × 104 μm2, P < 0.0001) (Fig. 4e). \nThese results suggest that bone marrow-derived VEGFR1-expressing cells that accumulate in the implants facili-\ntate both tissue growth and proangiogenic responses in endometrial fragments.\nEffect of macrophage deletion on angiogenesis in endometrial tissues. When we depleted mac-\nrophages using Clophosome N, we found that both endometrial tissue growth (Clophosome Control: 9.01 ± 0.54 \nmm2 vs. Clophosome N: 7.01 ± 0.13 mm2, P = 0.036, Fig. 5a) and angiogenesis (Clophosome Control: 6.63 ± 0.58 \n3/2.25 × 104 μm2 vs. Clophosome N: 3.58 ± 0.58 3/2.25 × 104 μm2, P = 0.0050, Fig. 5b) were significantly suppressed. \nTaken together, these results suggest that the accumulation of VEGFR1-expressing cells, possibly macrophages from \nthe bone marrow, is the key event that facilitates both growth and angiogenesis of endometrial tissues.\nFigure 3. Origin of the vasculature in endometrial tissues (a) Immunostaining of CD31 (red) in \nWT → GFP +TG (green) mice. Endometrial implants (I), granulation tissues formed between the implant and \nthe peritoneum (II), and peritoneum in contact with the endometrial implants (III) in WT → GFP +TG (green) \nat Day 14 post-implantation. The dotted line indicates the border between the peritoneum and granulation \ntissue. The dashed line indicates the border between the implant tissue and the granulation tissue. Scale bar, 25 \nμm. (b) Microvessel density of GFP\n+/CD31+ and GFP− /CD31+ microvessels in lesions from WT → GFP +TG \nmice at Day 14 post-implantation. Microvessel density (MVD) was determined in each part of the field at higher \nmagnification. Arrows (↓) indicate GFP\n+/CD31+ endothelial cells. Arrow heads (∇) indicate GFP− /CD31+ \nendothelial cells. Scale bars, 50 μm. Data are expressed as the mean ± SEM (n = 8 mice). ***P < 0.001 (Student’s \nt-test).\n\n8Scientific  RepoRts  |          (2019) 9:7037  | https://doi.org/10.1038/s41598-019-43185-8\nwww.nature.com/scientificreportswww.nature.com/scientificreports/\nFigure 4. Bone marrow-derived cells accumulate in endometrial tissues at Day 14 post-implantation (a) \nGFP+ and GFP−  endometrial tissues implanted into mice. GFP+TG, GFP transgenic WT mice; GFP+WT \nBMC, GFP+ bone marrow chimera WT mice. Scale bars, 500 μm. (b) Recruitment of bone marrow-derived \ncells in GFP+ bone marrow chimera mice (GFP+WT BMC) receiving GFP−  WT implants. Arrow heads, \nGFP−VEGFR1+ cells; Arrows, GFP + VEGFR1+ cells. GFP+ cells, and CD31+, CD11b+, or S100A4+ cells, \nand VEGFR1+ cells in the endometrial implants. Scale bars, 50 μm. (c) Percentage of GFP+/CD31+, CD11b+, \nor S100A4+ cells, and GFP− /CD31+, CD11b+, or S100A4+ cells in the VEGFR1+ cell population. Data are \nexpressed as the mean ± SEM (n = 4‒5 mice). *P < 0.05, **P < 0.01, and ***P < 0.001 (Student’s t-test t or † \nWelch’s test). (d,e) Size of endometrial implants (d) and microvessel density (e) in the WT → GFP +WT BMC \nand WT → GFP +TK−/−  BMC. GFP+WT BMC, GFP transgenic WT bone marrow chimera mice; GFP+TK−/−  \nBMC, GFP transgenic TK−/−  bone marrow chimera mice. Data are expressed as the mean ± SEM (n = 12 mice). \n*P < 0.05 and ***P < 0.001 (Student’s t-test).\n\n9Scientific  RepoRts  |          (2019) 9:7037  | https://doi.org/10.1038/s41598-019-43185-8\nwww.nature.com/scientificreportswww.nature.com/scientificreports/\nFlow cytometric analysis for CD11b+ cells in the endometrial implants.  Although our data sug-\ngested that newly formed blood vessels were derived from the pre-existing blood vessels in the implants, it has \nbeen suggested that CD11b+ mononuclear cells give rise to endothelial cell-like colonies36,37. Therefore, we fur-\nther examined whether or not CD11b+ cells have a profile of endothelial progenitor cells. To this aim, we deter-\nmined whether CD11b+ cells in the implants are positive for markers for endothelial progenitor cells including \nCD133 and CD34 using flow cytometry analysis. Flow cytometry analysis revealed that the percentage of CD11b+ \ncells in the implants were 6.8 ± 2.2%, while the percentage of CD11b+/CD133+/CD34+ were few (0.02 ± 0.01%, \nP = 0.0079) (Fig. 6), suggesting that CD11b+ macrophages do not have an endothelial progenitor cell profile.\nMolecules that interact with VEGFR1 in endometrial implants.  Finally, we attempted to iden-\ntify the downstream molecules regulated by VEGFR1 signaling in implanted endometrial tissues). Expression \nof VEGF-A, a ligand for VEGFR1, was induced to a similar extent in endometrial tissues in the WT →  WT \n(7.04 ± 0.66 × 10\n− 3) and TK− /→  TK− /−  (6.00 ± 0.60 × 10− 3) (Supplementary Fig. S3a). Expression of VEGF-A \nwas observed in CD11b + and S100A4 + cells (Supplementary Fig. S4). When we examined other growth fac-\ntors that regulate angiogenic responses (Supplementary Fig. S3), we found that expression of bFGF in the \nTK\n− /−  →  TK− /−  was significantly lower than that in the WT →  WT (WT →  WT: 3.36 ±  0.18 × 10− 3 vs. \nTK− /−  →  TK− /− : 2.54 ± 0.13 × 10− 3, P = 0.0006) (Supplementary Fig. 3b). In addition, CD11b + and S1004A+ \ncells were also positive for bFGF (Supplementary Fig. S5). The accumulation of bFGF+ cells in implanted endo-\nmetrial tissues was significantly lower in the TK−/−  → TK −/−  than in the WT → WT (WT → WT: 78.0 ± 0.6% vs. \nTK−/−  → TK −/− : 70.6 ± 0.6%, P < 0.0001) (Fig. 7).\nFigure 5. Effect of Clophosome N on growth and angiogenesis in endometrial implants (a,b) Administration \nof Clophosome N (0.2 ml/mouse, intraperitoneally) suppressed growth (a) and angiogenesis (b) in endometrial \ntissues in the WT → WT at Day 14. Data are expressed as the mean ± SEM (n = 4 in each group). *P < 0.05 \nand **P < 0.01 compared with the control (Student’s t-test). (c) The CD11b + cell population was markedly \nreduced after Clophosome N treatment. Data are expressed as the mean ± SEM (n = 6 in each group). *P < 0.05 \ncompared with the control (Student’s t-test).\n\n10Scientific  RepoRts  |          (2019) 9:7037  | https://doi.org/10.1038/s41598-019-43185-8\nwww.nature.com/scientificreportswww.nature.com/scientificreports/\nTo elucidate the role of bFGF in the development of endometrial tissue and angiogenesis in endometriosis, \nwe treated mice with FGFR inhibitor, PD173047 (Fig. 8a,b). Treatment with PD173047 significantly reduced size \nof endometriosis (PBS: 8.81 ± 0.20 mm2 vs. PD173047: 6.76 ± 0.33 mm2, P < 0.0001, Fig. 8a) and microvascular \ndensity (PBS: 8.00 ± 0.61/2.25 × 104 μm2 vs. PD173047: 5.16 ± 0.63/2.25 × 104 μm2, P = 0.0063, Fig. 8b) in recipi-\nent WT with endometrial implants from WT mice. To further examine whether or not VEGFR1-expressing mac-\nrophages produce bFGF , isolated bone marrow-derived macrophages from WT and TK−/−  mice were stimulated \nwith PlGF , a specific agonist for VEGFR1. In in vitro study, the expression of bFGF in response to PlGF in bone \nmarrow-derived WT-macrophages was higher than that from bone marrow-derived TK − /− macrophages (WT: \n1.60 ± 0.10 × 10− 5 vs. TK− /− : 0.93 ± 0.27 × 10− 5, P = 0.03) (Fig. 8d). However, there was no significant differ -\nence in VEGF expression between the genotype (WT: 5.78 ± 0.42 × 10−3  vs. TK−/− : 4.03 ± 0.76 × 10−3 , P = 0.09) \n(Fig. 8c).\nThese data suggest that VEGFR1 signaling increases bFGF expression, which then modulates growth and \nangiogenesis in endometrial tissues.\nDiscussion\nIn the present study, we showed that VEGF was a key regulator of angiogenesis in endometrial tissues. Cross trans-\nplantation experiments using TK\n− /−  and WT mice revealed that VEGFR1 signaling in host-derived cells in the \nimplants, played a role in both growth and angiogenesis. Accumulation of VEGFR1+ macrophages from the host \nbone marrow was the key driver of growth and angiogenesis in the endometrial implants. The results obtained sug-\ngested that blocking VEGFR1 signaling will be a promising strategy for the treatment of endometriosis.\nUsing genetically engineered mice, we recently reported that angiogenic responses in mice with hindlimb \nischemia were enhanced by VEGFR1 signaling but not by VEGFR2 signaling\n18, suggesting that the receptors \nresponsible for signaling during ischemia and endometriosis were similar. The mechanisms underlying the estab-\nlishment of endometriotic lesions are not fully understood; however, there is no doubt that the long-term survival \nand proliferation of these lesions are crucially dependent on the formation of new blood vessels, which guarantee \nthe supply of oxygen and essential nutrients\n9,38–40. Endometriotic lesions are typically characterized by dense \nvascularization1,41,42. Cross transplantation of endometrial tissues isolated from genetically engineered mice is a \nvery useful strategy for clarifying the cellular origin and tissue-specific functions of proangiogenic factors. Our \ndata indicated that VEGFR1 signaling in host-derived cells is responsible for angiogenesis and growth in the \nFigure 6. Flow cytometric analysis for CD11b+ cells in the endometrial implants (a) Flow cytometric dot plots \nanalysis for CD11b+ cells and CD34+/CD133+ cells isolated from the implants with WT → WT at Day 14. (b) \nThe percentage of CD11b+ cells and CD11b+/CD34+/CD133+ cells. Data are expressed as the mean ± SEM \n(n = 4 in each group). *P < 0.05 compared with the CD11b+ cells (Student’s t-test).\n\n11Scientific  RepoRts  |          (2019) 9:7037  | https://doi.org/10.1038/s41598-019-43185-8\nwww.nature.com/scientificreportswww.nature.com/scientificreports/\nFigure 7. Lack of VEGFR1 signaling suppresses bFGF expression in endometrial tissues (a,b) Expression of \nbFGF in endometrial implants from WT → WT (a) and TK−/−  → TK −/−  (b) at Day 14. Scale bars, 50 μm. (c) \nNumber of bFGF+ cells in the endometrial implants at Day 14. Data are expressed as the mean ± SEM (n = 4–5 \nmice). ***P < 0.001 (Student’s t-test).\nFigure 8. Effect of FGF inhibition in growth and angiogenesis in implants and FGF induction by PlGF in \nmacrophages (a,b) FGF inhibition with PD173047 reduced growth (a) and angiogenesis (b) in endometrial \nimplants from WT → WT at Day 14. Data are expressed as the mean ± SEM (n = 8 mice). *P < 0.05 (Student’s \nt-test t). (c,d) mRNA expression of VEGF (c) and bFGF (d) in isolated macrophages from WT and TK−/−  mice. \nIsolated bone marrow macrophages were stimulated with PlGF . Data are expressed as the mean ± SEM (n = 4 \nmice). *P < 0.05 (Student’s t-test t).\n\n12Scientific  RepoRts  |          (2019) 9:7037  | https://doi.org/10.1038/s41598-019-43185-8\nwww.nature.com/scientificreportswww.nature.com/scientificreports/\nendometrial tissues (Figs 2, 4 and Supplementary Fig. S2). We had previously reported that VEGFR1-expressing \nmacrophages that accumulate in damaged tissues facilitate tissue repair and vessel reconstruction 43. VEGFR1 \nTK− /−  bone marrow chimera mice exhibit delayed healing and vessel reconstruction after tissue damage, sug-\ngesting the VEGFR1-positive macrophages recruited from the bone marrow play a significant role in this pro-\ncess43. Here, we also indicated that bone marrow-derived macrophages expressing VEGFR1 in host drove both \nangiogenic responses in the implants and the growth of endometrial tissues (Fig.  4). It is frequently reported \nthat macrophages increase angiogenesis under pathological conditions44–46. Consistent with this, we found that \nthe macrophages play a significant role in the development of endometriosis (Fig.  4). Immunofluorescence \nsuggests that VEGF-A is produced by macrophages and fibroblasts (Supplementary Fig. S4). Because VEGF-A \ninduces chemotaxis in peritoneal macrophages through VEGFR1-mediated mechanisms\n14, and VEGFR1 medi-\nates monocyte/macrophage infiltration to local inflammatory sites 16,18,43, VEGF-A released from macrophages \nand fibroblasts recruits macrophages expressing VEGFR1 to develop the endometrial tissue. Taken together, \nVEGFR1-expressing macrophages recruited via VEGF-A/VEGFR1 signaling promoted angiogenesis in the endo-\nmetrial implants, leading to the maintenance and growth of ectopic endometrial tissues.\nThe current study demonstrated that VEGFR1\n+ cells express S100A4, which is an S100 protein that is known \nto be a specific marker for fibroblasts35. We previously reported that fibroblasts recruited from the bone marrow \naccumulate in stromal tissues during up-regulated tumor-associated angiogenesis and tumor growth22; however, \nthe host-derived S1004A+ cell population in the implants in the bone marrow transplantation experiments was \nsmaller than the host-derived CD11b+ population. These results indicate that S1004A+ fibroblasts play a minor \nrole in promotion of angiogenesis and development of endometriosis.\nWe were surprised that CD31+ vascular endothelial cells in the implants were not GFP+; this was the case even \nin GFP transgenic WT bone marrow chimera mice (Fig.  3b). We also demonstrated that accumulated CD11b+ \nmacrophages in the implants displayed no property of endothelial progenitor cells. This suggests that post-natal \nvasculogenesis may play only a minor role during the development of endometriosis. Although the mechanisms \nunderlying the establishment of endometriotic lesions are unclear, it is possible that vasculogenesis plays a role in \nendometriosis; however, a previous study shows that 13% (at most) of endothelial cells in a mouse endometriosis \nmodel were derived from the bone marrow\n47. The results of the current study suggest that the majority of blood \nvessels in the implants grew in a macrophage-dependent manner rather than by vasculogenesis. Thus, our results \nare consistent with those in the above report showing that vasculogenesis is not the main driver of blood vessel \nformation in endometrial tissues. Our results also suggest that accumulated host-derived macrophages promote \nthe formation of new blood vessels from the preexisting tissues in the implants.\nWe also found that VEGFR1 signaling was a major determinant of neovascularization in endometrial tissues. \nBlockade of VEGF signaling with a soluble VEGF receptor or an affinity-purified anti-VEGF antibody is an effec-\ntive treatment for endometriosis in nude mice\n48. However, as shown in Supplementary Fig. S3, endometrial tissues \nnot only express VEGF but also various other growth factors. Among these, we found that bFGF expression was \ndependent upon VEGFR1 (Supplementary Fig. S3, and Fig. 7). Thus, it is plausible that the development of new \nblood vessels in endometriotic lesions is critically dependent on the interaction between multiple signaling mol-\necules including bFGF . bFGF was reported to drive angiogenesis in endometrial tissues\n49,50. Consistent with these \nFigure 9. Roles of VEGFR1 signaling that facilitate angiogenesis in endometrial tissues VEGF is a key regulator \nof growth and angiogenesis in endometrial tissues. Cross transplantation experiments using TK−/−  and WT \nmice revealed that VEGFR1 signaling in the host, or host-derived cells in the implants, played a role in both \ngrowth and angiogenesis. The blood vessels in the implants were not derived from the host peritoneum. \nImmunostaining for VEGFR1 suggested that high numbers of VEGFR1\n+ cells such as macrophages were \ninfiltrated into the endometrial tissues. Accumulation of VEGFR1+ macrophages from the host bone marrow \nwas the key driver of angiogenesis in the endometrial implants via secretion of bFGF .\n\n13Scientific  RepoRts  |          (2019) 9:7037  | https://doi.org/10.1038/s41598-019-43185-8\nwww.nature.com/scientificreportswww.nature.com/scientificreports/\nobservations, our data showed that FGF receptor inhibition suppressed the growth of endometrial tissue and angi-\nogenesis. These taken together suggested that growth factors including VEGF and bFGF interacted with VEGFR1.\nIt has been shown that inhibition of VEGFR1 signaling attenuates tumor growth and rheumatoid arthritis \nthrough suppressing angiogenesis51. Angiogenesis inhibitors including tyrosine kinase inhibitors display a bene-\nficial effect on endometriosis in rodents52. Additionally, progesterone derived from ovary induces the expression \nof VEGF-A, which is a critical factor in the dynamic regulation of the uterine vasculature during postmenstrual \nrepair as well as pregnancy\n53. Furthermore, the degree of preeclampsia is well correlated with increased serum \nlevels of soluble fms-like tyrosine kinase-1 (sFlt-1) in pregnant mothers. Because sFlt-1 would form a molecular \nbarrier against abnormal vascular permeability and abnormal angiogenesis, by trapping VEGF and PlGF , sFlt-\n1-blocking agents could treat preeclampsia\n51. Because VEGF-A neutralizing antibody and multi-tyrosine kinase \nVEGFR inhibitor have been widely used in the treatment of cancer for suppressing angiogenesis, VEGFR1 inhibi-\ntion would be a useful tool for regulation of endometriosis-associated angiogenesis in reproductive aged women.\nIn conclusion, VEGF is a key regulator of growth and angiogenesis in endometrial tissues (Fig. 9). Accumulation \nof VEGFR1\n+ macrophages from the host bone marrow was the key driver of growth and angiogenesis in the endo-\nmetrial implants via secretion of bFGF . Taken together, these results suggest that blocking VEGFR1 with antibodies \nor a small molecule kinase inhibitor will be a promising strategy for the treatment of endometriosis.\nData Availability\nThe datasets generated during and/or analyzed during the current study are available from the corresponding \nauthor on reasonable request.\nReferences\n 1. Giudice, L. C. Clinical practice. Endometriosis. N Engl J Med. 362, 2389–2398 (2010).\n 2. Eskenazi, B. & Warner, M. L. Epidemiology of endometriosis. Obstet Gynecol Clin North Am. 24, 235–58 (1997).\n 3. Nnoaham, K. E. et al. Impact of endometriosis on quality of life and work productivity: a multicenter study across ten countries. \nFertil Steril. 96, 366–373 (2011).\n 4. Hey-Cunningham, A. J. et al . Angiogenesis, lymphangiogenesis and neurogenesis in endometriosis. Front Biosci (Elite Ed). 5, \n1033–1056 (2013).\n 5. Ferrara, N. & Kerbel, R. S. Angiogenesis as a therapeutic target. Nature. 438, 967–974 (2005).\n 6. Adams, R. H. & Alitalo, K. Molecular regulation of angiogenesis and lymphangiogenesis. Nat Rev Mol Cell Biol. 8, 464–478 (2007).\n 7. Ferrara, N., Gerber, H. P . & LeCouter, J. The biology of VEGF and its receptors. Nat Med. 9, 669–676 (2003).\n 8. Ferrara, N. Vascular endothelial growth factor: basic science and clinical progress. Endocr Rev. 25, 581–611 (2004).\n 9. May, K. & Becker, C. M. Endometriosis and angiogenesis. Minerva Ginecol. 60, 245–254 (2008).\n 10. Amano, H. et al. Host prostaglandin E(2)-EP3 signaling regulates tumor-associated angiogenesis and tumor growth. J Exp Med. 197, \n221–232 (2003).\n 11. Murakami, M. & Kudo, I. Recent advances in molecular biology and physiology of the prostaglandin E2-biosynthetic pathway. Prog \nLipid Res. 43, 3–35 (2004).\n 12. Jin, D. K. et al. Cytokine-mediated deployment of SDF-1 induces revascularization through recruitment of CXCR4+ hemangiocytes. \nNat Med 12, 557–567 (2006).\n 13. Shibuya, M. Vascular endothelial growth factor-dependent and—independent regulation of angiogenesis. BMB Rep. 41, 278–286 \n(2008).\n 14. Koga, J. et al. Soluble Flt-1 gene transfer ameliorates neointima formation after wire injury in flt-1 tyrosine kinase-deficient mice. \nArterioscler Thromb Vasc Biol. 29, 458–464 (2009).\n 15. Shalaby, F . et al. Failure of blood-island formation and vasculogenesis in Flk-1- deficient mice. Nature 376, 62–66 (1995).\n 16. Hiratsuka, S., Minowa, O., Kuno, J., Noda, T. & Shibuya, M. Flt-1 lacking the tyrosine kinase domain is sufficient for normal \ndevelopment and angiogenesis in mice. Proc Natl Acad Sci USA 95, 9349–9354 (1998).\n 17. Sato, T. et al. Vascular endothelial growth factor receptor 1 signaling facilitates gastric ulcer healing and angiogenesis through the \nupregulation of epidermal growth factor expression on VEGFR1+ CXCR4+ cells recruited from bone marrow. J Gastroenterol. 49, \n455–469 (2014).\n 18. Amano, H. et al. The Role of Vascular Endothelial Growth Factor Receptor-1 Signaling in the Recovery from Ischemia. PLoS One. \n10, e0131445, https://doi.org/10.1371/journal.pone. 0131445 (2015).\n 19. Numao, A. et al. The inducible prostaglandin E synthase mPGES-1 regulates growth of endometrial tissues and angiogenesis in a \nmouse implantation model. Biomed Pharmacother. 65, 77–84 (2011).\n 20. Majima, M. et al. Cyclo-oxygenase-2 enhances basic fibroblast growth factor-induced angiogenesis through induction of vascular \nendothelial growth factor in rat sponge implants. Br J Pharmacol. 130, 641–649 (2000).\n 21. Y oshida, S. et al. COX-2/VEGF-dependent facilitation of tumor-associated angiogenesis and tumor growth in vivo. Lab Invest. 83, \n1385–1394 (2008).\n 22. Katoh, H. et al. COX-2 and prostaglandin EP3/EP4 signaling regulate the tumor stromal proangiogenic microenvironment via \nCXCL12-CXCR4 chemokine systems. Am J Pathol. 176, 1469–1483 (2010).\n 23. Amano, H. et al. Adenylate cyclase/protein kinase A signaling pathway enhances angiogenesis through induction of vascular \nendothelial growth factor in vivo. Jpn J Pharmacol. 87, 181–188 (2001).\n 24. Sekiguchi, K. et al. Vascular endothelial growth factor regulates growth of endometrial tissues and angiogenesis in a mouse \ntransplantation model. Kitasato Med J. 46, 15–23 (2016).\n 25. Ogawa, Y . et al. Bone marrow-derived EP3-expressing stromal cells enhance tumor-associated angiogenesis and tumor growth. \nBiochem Biophys Res Commun. 382, 720–725 (2009).\n 26. Becker, C. M. et al. A novel noninvasive model of endometriosis for monitoring the efficacy of antiangiogenic therapy. Am J Pathol. \n168, 2074–2084 (2006).\n 27. Hirata, T. et al. Development of an experimental model of endometriosis using mice that ubiquitously express green fluorescent \nprotein. Hum Reprod. 20, 2092–2096 (2005).\n 28. Pardo, O. E. et al. The fibroblast growth factor receptor inhibitor PD173074 blocks small cell lung cancer growth in vitro and in vivo. \nCancer Res. 69, 8645–8651 (2009).\n 29. Gyanchandani, R., Ortega Alves, M. V ., Myers, J. N. & Kim, S. A proangiogenic signature is revealed in FGF-mediated bevacizumab-\nresistant head and neck squamous cell carcinoma. Mol Cancer Res. 11, 1585–1596 (2013).\n 30. Hosono, K. et al. Roles of prostaglandin E2-EP3/EP4 receptor signaling in the enhancement of lymphangiogenesis during fibroblast \ngrowth factor-2-induced granulation formation. Arterioscler Thromb Vasc Biol. 31, 1049–1058 (2011).\n 31. Fujita, M., Hayashi, I., Y amashina, S., Itoman, M. & Majima, M. Blockade of angiotensin AT1a receptor signaling reduces tumor \ngrowth, angiogenesis, and metastasis. Biochem Biophys Res Commun. 294, 441–447 (2002).\n\n14Scientific  RepoRts  |          (2019) 9:7037  | https://doi.org/10.1038/s41598-019-43185-8\nwww.nature.com/scientificreportswww.nature.com/scientificreports/\n 32. Nishizawa, N. et al. Inhibition of microsomal prostaglandin E synthase-1 facilitates liver repair after hepatic injury in mice. J \nHepatol. 69, 110–120 (2018).\n 33. Kojo, K. et al. BLT1 signalling protects the liver against acetaminophen hepatotoxicity by preventing excessive accumulation of \nhepatic neutrophils. Sci Rep. 6, 29650, https://doi.org/10.1038/srep29650 (2016).\n 34. Kurosaka, M. et al. Reduced angiogenesis and delay in wound healing in angiotensin II type 1a receptor-deficient mice. Biomed \nPharmacother. 63, 627–634 (2009).\n 35. Boye, K. & Maelandsmo, G. M. S100–A4 and metastasis: a small actor playing many roles. Am J Pathol 176, 528–535 (2010).\n 36. Fujiyama, S. et al. Bone marrow monocyte lineage cells adhere on injured endothelium in a monocyte chemoattractant protein-1-\ndependent manner and accelerate reendothelialization as endothelial progenitor cells. Circ Res. 93, 980–989 (2003).\n 37. Gulati, R. et al. Autologous culture-modified mononuclear cells confer vascular protection after arterial injury. Circulation.: 108, \n1520–1526 (2003).\n 38. Groothuis, P . G., Nap, A. W ., Winterhager, E. & Grümmer, R. Vascular development in endometriosis. Angiogenesis. 8, 147–156 \n(2005).\n 39. Laschke, M. W . & Menger, M. D. In vitro and in vivo  approaches to study angiogenesis in the pathophysiology and therapy of \nendometriosis. Hum Reprod Update. 13, 331–342 (2007).\n 40. Taylor, R. N. et al. Mechanistic and therapeutic implications of angiogenesis in endometriosis. Reprod Sci. 16, 140–146 (2009).\n 41. Nisolle, M., Casanas-Roux, F ., Anaf, V ., Mine, J. M. & Donnez, J. Morphometric study of the stromal vascularization in peritoneal \nendometriosis. Fertil Steril. 59, 681–684 (1993).\n 42. McLaren, J. Vascular endothelial growth factor and endometriotic angiogenesis. Hum Reprod Update. 6, 45–55 (2000).\n 43. Ohkubo, H. et al. VEGFR1-positive macrophages facilitate liver repair and sinusoidal reconstruction after hepatic ischemia/\nreperfusion injury. PLoS One. 9, e105533, https://doi.org/10.1371/journal.pone.0105533 (2014).\n 44. Kim, K. E. et al . Role of CD11b+  macrophages in intraperitoneal lipopolysaccharide-induced aberrant lymphangiogenesis and \nlymphatic function in the diaphragm. Am J Pathol. 175, 1733–1745 (2009).\n 45. De Palma, M. & Lewis, C. E. Cancer: Macrophages limit chemotherapy. Nature. 472, 303–304 (2011).\n 46. Pollard, J. W . Trophic macrophages in development and disease. Nat Rev Immunol. 9, 259–270 (2009).\n 47. Mints, M. et al. Endometrial endothelial cells are derived from donor stem cells in a bone marrow transplant recipient. Hum Reprod. \n23, 139–143 (2008).\n 48. Hull, M. L. et al. Antiangiogenic agents are effective inhibitors of endometriosis. J Clin Endocrinol Metab. 88, 2889–2899 (2003).\n 49. Di Blasio, A. M. et al. Basic fibroblast growth factor messenger ribonucleic acid levels in eutopic and ectopic human endometrial \nstromal cells as assessed by competitive polymerase chain reaction amplification. Mol Cell Endocrinol. 115, 169–175 (1995).\n 50. Ferriani, R. A., Charnock-Jones, D. S., Prentice, A., Thomas, E. J. & Smith, S. K. Immunohistochemical localization of acidic and \nbasic fibroblast growth factors in normal human endometrium and endometriosis and the detection of their mRNA by polymerase \nchain reaction. Hum Reprod. 8, 11–16 (1993).\n 51. Shibuya, M. Vascular endothelial growth factor and its receptor system: physiological functions in angiogenesis and pathological \nroles in various diseases. J Biochem. 153, 13–19 (2103).\n 52. Liu, S. et al. Efficacy of Anti-VEGF/VEGFR Agents on Animal Models of Endometriosis: A Systematic Review and Meta-Analysis. \nPLoS One. 11, e0166658, https://doi.org/10.1371/journal.pone.0166658 (2016).\n 53. Kim, M. et al. VEGF-A regulated by progesterone governs uterine angiogenesis and vascular remodelling during pregnancy. EMBO \nMol Med. 5, 1415–1430 (2013).\nAcknowledgements\nWe thank Michiko Ogino, Kyoko Y oshikawa, Mieko Hamano, and Erina Sato for technical assistance. This work \nwas supported by grants from The Ministry of Education, Culture, Sports, Science and Technology (MEXT), \n#12470529, #12670094, #15K15056, #80532556, #23116102, #24659119, #26462132, #26293055, and #18H02605, \nand from Takeda Science Foundation and Uehara Memorial Foundation. This study was also supported by an \nIntegrative Research Program of the Graduate School of Medical Science, Kitasato University, and the Keyaki-kai, \nKitasato University School of Medicine.\nAuthor Contributions\nStudy concept and design: K.S. and M.M. Acquisition of data: K.S., Y .I., K.H., T.I., K.H., M.H. and A.N. Analysis \nand interpretation of data: K.S., Y .I. and M.M. Drafting of the manuscript: K.S., Y .I. and M.M. Statistical analysis: \nK.S., Y .I. and H.A. Technical and material support: H.A., M.S. and N.U. Study supervision: N.U. and M.M.\nAdditional Information\nSupplementary information accompanies this paper at https://doi.org/10.1038/s41598-019-43185-8.\nCompeting Interests: The authors declare no competing interests.\nPublisher’s note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and \ninstitutional affiliations.\nOpen Access This article is licensed under a Creative Commons Attribution 4.0 International \nLicense, which permits use, sharing, adaptation, distribution and reproduction in any medium or \nformat, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Cre-\native Commons license, and indicate if changes were made. The images or other third party material in this \narticle are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the \nmaterial. If material is not included in the article’s Creative Commons license and your intended use is not per-\nmitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the \ncopyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.\n \n© The Author(s) 2019","source_license":"CC0","license_restricted":false}