Introduction
Implantation of the fertilized oocyte induces uterine decidualization, the rapid proliferation and differentiation of stromal fibroblasts into glycogen and lipid rich decidual cells. Endothelial cells within the decidua proliferate to form a rudimentary vascular plexus necessary to maintain pregnancy prior to placenta development [
1‐
3]. Cells of the innate immune system, including natural killer cells and macrophages, infiltrate the decidua and function to establish and maintain maternal–fetal tolerance [
4,
5]. These events are coordinated by ovarian estrogen and progesterone. Angiogenesis during decidualization is required for a successful pregnancy, however the signaling pathways that regulate this process have yet to be fully characterized.
In mice and non-human primates, VEGF mediates the increased uterine vascular permeability and decidual angiogenesis required for embryo implantation [
6‐
8]. We have shown that the VEGF receptors (VEGFR-1, VEGFR-2, and VEGFR-3) are expressed in distinct patterns in the peri-implantation uterine decidua [
9]. To investigate the requirement for VEGF receptor signaling in the decidua, we administered a single, peri-implantation (E3.75) dose of blocking antibody against VEGFR-1, VEGFR-2, or VEGFR-3, in a progesterone-replaced, ovariectomized mouse model. Inhibition of VEGF-A receptor, VEGFR-2, blocked decidual angiogenesis observed at E7.5 and resulted in an aborted pregnancy prior to E10.5, while VEGFR-3 inhibition moderately reduced decidual angiogenesis but had no effect on pregnancy [
9]. In contrast, VEGFR-1 blockade at a single E3.75 time-point did not effect decidual angiogenesis or disrupt pregnancy [
9].
Embryo implantation and trophoblast invasion create a pro-inflammatory environment leading to the recruitment of immune cells that mediate maternal tolerance to the semi-allogeneic embryo [
5,
10]. Decidual macrophages are the second most abundant immune cell population at the implantation site, comprising 20-30% of immune cells in the uterine decidua [
4,
11,
12]. Dysregulated macrophage activation in the decidua has been implicated in recurrent miscarriages [
13]. Decidual macrophage numbers, as well as the proportion of pro-apoptotic Fas ligand expressing decidual macrophages, are increased in spontaneous miscarriages [
14]. Thus, understanding the signaling pathways that regulate decidual macrophages may elucidate causes of early pregnancy failures.
Macrophages are closely associated with endothelial cells during physiologic angiogenesis during ovarian corpus luteum development and postnatal retinal development, [
15,
16]. Activation of macrophage VEGFR-1 has a role in macrophage recruitment to sites of active angiogenesis [
17,
18]. Macrophages have been implicated in promoting angiogenesis by releasing pro-angiogenic factors such as VEGF-A and angiopoietin [
15] and in anastomosis, the formation of a bridge between two angiogenic tip cells followed by vascular sprout fusion [
15,
19]. Ablation of ovarian macrophages leads to significant endothelial cell depletion and hemorrhage [
16,
20]. Macrophages have been implicated in the pregnant uterus, including acting in the coordination of the maternal immune response and promoting angiogenesis. However, the relationship between decidual macrophages, endothelial cells and VEGFR-1 has not been investigated.
VEGFR-1 binds VEGF-A, VEGF-B and placental growth factor (PlGF) [
21].
VEGFR-1 gene encodes two mRNA splice variants. The full-length VEGFR-1 variant encodes a membrane bound receptor with classic intracellular tyrosine kinase signaling that positively regulates angiogenesis. The alternatively spliced variant encodes the VEGFR-1 extracellular domain resulting in a secreted protein known as soluble VEGFR-1 (sFlt-1) that is anti-angiogenic [
22,
23]. sFlt-1 sequesters the VEGFR-2 ligand, VEGF-A as well as, the VEGFR-1 specific ligands, VEGF-B and PlGF. Thus, sFlt-1 functions as a negative regulator of both VEGFR-2 and full-length VEGFR-1 signaling and is necessary for proper embryonic angiogenesis [
23,
24]. VEGFR-1 null (
Flt-1
-/-
) mutant embryos die because of an overgrowth of vascular endothelial cells [
24]. Mice with deletion of the VEGFR-1 Tyrosine Kinase (TK) domain (
Flt-1 TK
-/-
) are viable and have normal blood vessel development, but have altered macrophage migration [
25]. In pathologic conditions, such as cancer and inflammatory diseases, full length VEGFR-1 signaling promotes angiogenesis and activates/mediates migration of macrophage-lineage cells [
26]. Thus, the context and abundance of sFlt-1 and full-length VEGFR-1 determine the pro- or anti-angiogenic effect of VEGFR-1.
In this study, we determined the expression pattern of VEGFR-1 with respect to endothelial cells and macrophages in the pre- and post-implantation uterus. VEGFR-1 neutralizing antibodies were administered prior to and during implantation. Our goal was to determine if inhibition of VEGFR-1 affects decidual angiogenesis, macrophage recruitment, and/or the establishment and maintenance of pregnancy in mice. Using two different VEGFR-1 blocking antibodies, MF-1 and R&D Systems AF471 that prevent ligand binding to VEGFR-1, we found that both vascular density and macrophage recruitment to the pregnant uterus were significantly reduced. Despite this, embryos implanted in the uterus and pregnancy progressed after VEGFR-1 blockade. We conclude that VEGFR-1 activity is required for proper decidual angiogenesis and macrophage recruitment to the implantation site during pregnancy.
Materials and methods
Animal model
The Columbia University Institutional Animal Care and Use Committee approved animal studies. Adult wild-type CD1 female mice and CD1 male mice of proven fertility were used. When mice were bred, noon on the day a mating plug was observed was designated embryonic day (E) 0.5. Uteri of non-pregnant female mice and pregnant females at E3.5 and E6.5 were embedded in Tissue-Tek® O.C.T.™ Compound (Sakura Fine Technical Co, Ltd, Tokyo, Japan), snap-frozen on dry ice in ethanol and stored at -80°C.
To determine the effect of continuous VEGFR-1 blockade, females were intraperitoneally (i.p.) injected with VEGFR-1 blocking monoclonal antibody (MF-1, 132 mg/kg animal, ImClone Systems, Inc.; n = 5 mice), anti-mouse VEGFR-1 antibody (AF471, 2 mg/kg animal, R&D systems; n = 1 mice), or saline (n = 5 mice) every 3 days for 18 days. MF-1 has an elimination half-life of 3 days and reaches maximal plasma concentrations 6 hours after dosing [
27]. After 6 doses, females were mated and a final dose of antibody or saline was given on E3.5. Pregnant females were sacrificed on E7.5. Implantation sites were counted, uterine weight measured, and uteri freshly frozen in O.C.T.™ compound for sectioning. Pregnant females were sacrificed on E10.5 to determine the number of implantation sites and uterine weight.
Histological staining
For uterine analyses, 5 μm transverse sections through the non-pregnant and E3.5 uteri were generated. For implantation site analyses, frontal sections through uterus/implantation site were generated for E6.5 at 7 μm and for E7.5 with VEGFR-1 blockade at 12 μm. Implantation was confirmed by H&E staining every 5th section. Specific staining was performed at least 3 times and 5 different uterine sections or implantation sites were analyzed.
Sections were stained as previously described [
27]. Primary antibodies included anti-mouse VEGFR-1 (R&D Systems, AF471), anti-mouse F4/80 (eBioscience, 14–4801), anti-mouse CD31 (BD Biosciences, 553370), anti-mouse endomucin (Santa Cruz, sc-65495), anti-mouse VE-cadherin (BD Biosciences, 550548), and anti-mouse CD11b (Abcam ab8878). For IHC staining, biotin rabbit anti-goat IgG (Vector, BA-5000), biotin goat anti-rat IgG (BD Biosciences, 559286), the avidin/biotin blocking kit (Vector, SP-2001), the Vectastain ABC kit and DAB substrate kit (Vector, SK-4100) were used. Sections were counterstained with hematoxylin. For IF experiments, the following secondary antibodies were used: donkey anti goat-IgG Alexa-Fluor 594 (Invitrogen, A11058) and donkey anti rat-IgG Alexa-Fluor 488 (Invitrogen, A2108). Slides were covered with Vectashield containing with 4′, 6-diamidino-2-phenylindole (Vector, H-1200) for nuclear visualization.
Imaging
IHC and H&E staining images were captured with a Nikon Eclipse E800 microscope and Nikon DXM 1200 digital camera and NIS-Elements D3.10 software or ImagePro Plus v.4.01 software. Fluorescent images were captured using a Nikon A1 scanning confocal microscope on an Eclipse Ti microscope stand (Nikon Instruments, Melville, NY). Standard lasers and filters were used to image DAPI, AlexaFluor 488, and TRITC. Images were taken using the 10×/0.4 and 60×/1.49 objectives. Representative Z-stacked maximum intensity images are shown.
Quantitation of decidual macrophage and vascular density
The percentage of decidua occupied by blood vessels or macrophages was calculated from E7.5 decidua frontal uterine sections by dividing the total area of CD31 or F4/80 staining by the area of the decidua multiplied by 100. Images were captured using a Nikon Eclipse E800 microscope and Nikon DXM 1200 digital camera, and data were processed using ImagePro Plus Version 4.01 (Media Cybernetics) [
28].
Serum progesterone levels
Blood was obtained from all animals by cardiopuncture. Serum progesterone levels were measured using a competitive chemiluminescent immunoassay (Diagnostic Products Corp./Siemens) [
9].
Statistical analysis
Data are presented as mean ± standard error of the mean (sem). Unpaired Student t-test was used to compare sample means. P < 0.05 was considered a statistically significant difference. Statistical analyses were performed using the Statistical Package for Social Science version 15.0 (SPSS, Inc., Chicago, Il).
Discussion
We have previously shown that peri-implantation inhibition of VEGFR-2 significantly reduces vascular density and causes pregnancy loss by E10.5. We showed that only a subset of endothelial cells express VEGFR-1 and peri-implantation inhibition with a single dose of VEGFR-1 blocking antibody at E3.75 does not affect vascular density or pregnancy [
9]. Our present evaluation of VEGFR-1 expression with respect to endothelial cells and macrophages shows that VEGFR-1 expression is mainly restricted to uterine and decidual endothelial cells, while VEGFR-1 is not expressed in monocytes and macrophages. Endothelial cells expressing VEGFR-1 were often in direct contact with CD11b
+ monocytes and F4/80
+ macrophages. Finally repeated pre-implantation injections of a VEGFR-1 blocking antibody, significantly decrease both the number of macrophages and vascular density in the decidua of E7.5 pregnant mice. Taken together, our results suggest that the decreased vascularity observed at E7.5 is a direct result of VEGFR-1 blockade, whereas the reduction in macrophages may be an indirect result of the decrease in VEGFR-1 function in the endothelial cells.
Treatment with MF-1, which blocks binding of VEGF-A and PlGF to VEGFR-1, can have pro- or anti-angiogenic effects, depending on the context of VEGFR-1 function [
26,
40‐
42]. PlGF signaling, via membrane-localization of full length VEGFR-1, plays a pro-angiogenic role in pathological angiogenesis [
26,
42]. In murine decidual angiogenesis, VEGFR-1 appears to be membrane associated consistent with full-length receptor expression (Figures
1 and
3E, H). Thus, the reduced angiogenesis we observed is likely due to blocking ligand-dependent VEGFR-1 pro-angiogenic signaling. However, it has been previously shown that the VEGFR-1 TK domain is not necessary for murine decidual angiogenesis [
29]. Thus the pro-angiogenic role for VEGFR-1, we observed may be independent of VEGFR-1 TK signaling pathway.
We have shown that a single dose of MF-1 administered within 24 hours of implantation does not affect vascular density [
9], but continuous VEGFR-1 blockade results in a 48% reduction in vascular density (Figure
5D). A greater than 50% reduction in decidual vascular density with VEGFR-2 blockade results in pregnancy loss, likely due to loss of angiogenesis in both the primary and secondary decidual zones [
9]. In contrast, inhibition of VEGFR-3 results in a less than 50% decrease in primary decidual zone angiogenesis, but does not lead to pregnancy loss [
9]. Combined, these data suggest that implantation and early pregnancy can tolerate a less than 50% reduction of vascular density in the primary decidual zone.
Macrophages are prominent in the adult ovary and uterus, two organs with physiological cyclical angiogenesis regulated by the VEGF/VEGFR-2 signaling pathway [
9,
28,
43]. The number and phenotype of macrophages in the non-pregnant uterus changes with the ovarian cycle and additional macrophages are recruited into the endometrium during the peri-implantation period [
20,
44]. It has been shown that pre-implantation macrophage ablation prevents embryo implantation and results in infertility [
20]. However, fertility is restored when macrophage-depleted mice are supplemented with exogenous progesterone suggesting a defect in corpora lutea function. In fact, macrophages have been shown to regulate corpus lutea development and function, specifically the synthesis of progesterone that is required for embryo implantation and establishing pregnancy [
20,
45]. Thus, uterine macrophages may not be necessary for embryo implantation. We found that pregnancy can tolerate a reduction of macrophages by 37% of normal. Taken together, the data suggest that macrophages do not have an obligatory role in the uterine decidualization and implantation of the embryo. However the role of macrophages after embryo implantation still remains to be elucidated.
Acknowledgement
Images were collected in the Confocal and Specialized Microscopy Shared Resource of the Herbert Irving Comprehensive Cancer Center at Columbia University Medical Center.
Grant sponsors
This work was supported by grants from the Robert Wood Johnson Foundation/Amos Medical Faculty Development Program (N.C.D.), Merck (N.C.D.), NIH NHLBI 5R01HL112626 (J.K.K.), and NIH NCI R01CA136673-01 (C.J.S.). Merck partially funded this study, but had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
ND prepared the manuscript and all figures. RZ and QKT did the experiments with MF-1 and R&D Systems AF471. CSP did the immunohistochemistry. JK and MS reviewed and edited the final manuscript. CS prepared and edited the manuscript. All authors read and approved the final manuscript.