Vascular Endothelial Growth Factor Receptor 2 (VEGFR-2) Functions to Promote Uterine Decidual Angiogenesis during Early Pregnancy in the Mouse

Experimental design

The experiments were designed to investigate whether VEGF receptors (VEGFRs) play an important role in the formation and function of uterine decidual blood vessels during early pregnancy development. Seven-to eight-week old female CD1 mice (Charles River Laboratories International, Inc., Wilmington, MA) were mated with adult males from 1700–2300 h. Identification of a vaginal plug the following morning was interpreted as successful mating. 1100 h was considered d 0.5 post coitum. On embryonic d (ED) 3.5 + 2 h, four extended-release progesterone (P4) capsules (10 mg P4 per capsule, 21-d release, 4-mm diameter; Innovative Research of America, Sarasota, FL) were placed under the skin of pregnant mice. A bilateral ovariectomy was performed 4 h after placement of the P4 capsules (16) (Fig. 1A). In intact pregnancies, average P4 levels were 29.3 ng/ml and greater than 40 ng/ml on ED 6.5 and 10.5, respectively. P4-replaced, ovariectomized, pregnant mice (PROPmice) with P4 levels greater than 30 ng/ml at the time of death, which made up 95% of the animals, were used for all experiments. For each experiment, a single ip injection of 0.5 ml neutralizing antibodies against either VEGFR-2 (DC101) (14), VEGFR-3 (mF4-31C1) (25), VEGFR-1 (MF-1) (22), or saline was administered on ED 3.75 (Fig. 1A). These neutralizing antibodies, provided by ImClone Systems, Inc. (New York, NY), have been used to inhibit signaling from VEGFRs in murine models of tumor angiogenesis, lymphangiogenesis, atherosclerosis, and ovarian folliculogenesis (23,24,25,26,27,28,29). The dosages of VEGFR-neutralizing antibodies were as follows: DC101 (66 mg/kg), mF4-31C1 (132 mg/kg), and MF-1 (132 mg/kg). All protocols were approved by the Institutional Animal Care and Use Committee of Columbia University.

Dosage and behavior of DC101, mF4-31C1 and MF-1 in vivo

Similar to previous studies, no adverse systemic effects were detected after ip administration of rat antimouse antibodies against VEGFR-2 (DC101) (16,26), VEGFR-3 (mF4-31C1) (25), or VEGFR-1 (MF-1) (22,27). Dose-response curves using DC101, mF4-31C1, and MF-1 were performed by injecting a single dose (33, 66, 132, or 264 mg/kg) of each antibody on ED 3.75 and evaluating the effects on uterine angiogenesis on ED 7.5. For DC101, angiogenesis was inhibited at 66 mg/kg. This inhibition did not increase in magnitude with higher dosages of DC101. Therefore, a dosage of 66 mg/kg, which is sufficient to inhibit ovarian angiogenesis (19), was chosen for our experiments. For mF4-31C1, an inhibitory response was seen at 132 mg/kg. This did not increase in magnitude with higher dosages. Thus, 132 mg/kg was chosen for our experiments. For MF-1, no significant inhibitory response was seen up to 264 mg/kg. For our experiments, we selected an amount of 132 mg/kg, which is four times the dosage of MF-1 used to inhibit angiogenesis in a tumor model (30).

Similar to previous studies, injection of isotype-matched IgG and saline produced identical results (n = 5), but nonspecific IgG increased background staining in immunohistochemistry (16,19). Saline was used for all controls.

Blocking antibodies against VEGFRs could penetrate the embryo and disrupt early embryonic development because these receptors are expressed in early pregnancy (31), and homozygous null mutations in VEGFR-1, -2, and -3 are embryonic lethal (32,33,34,35). To test whether VEGFR-blocking antibodies bind to VEGFRs in the embryo, PROPmice received ip injections of either DC101, mF4-31C1, or MF-1 on ED 9.5 and were killed on ED 10.5. Cross-sections of ED 10.5 uteri were incubated with rabbit antirat biotinylated secondary antibody (1:100). Specific binding was detected in the uteri but not in the embryos. As a positive control, uterine cross-sections from untreated ED 10.5 mice were incubated with either DC101, mF4-31C1, or MF-1 (all 1:50) followed by the secondary antibody. Binding was detected in the uteri and embryos. As a negative control, ED 10.5 uteri from mice treated with rat IgG were incubated with the secondary antibody. No binding was detected in the uteri or embryos. These experiments demonstrated that our blocking antibodies can bind VEGFRs in both the uterus and embryo, but maternal systemic administration does not result in binding to embryos.

Formation of the decidua

Control mice received ip injections of saline on ED 3.75, whereas treatment group animals received either DC101, mF4-31C1, MF-1, or combined DC101/mF4-31C1. The animals were killed on ED 5.5 to assess cellular proliferation and on ED 7.5 to evaluate the number of implantation sites, decidual blood vessel density, cellular proliferation, and apoptosis. To assess stromal cell differentiation by alkaline phosphatase staining, ED 6.5 uteri were fixed in 4% paraformaldehyde for 2 h followed by a sucrose gradient. Frozen sections cut at 5 μm were postfixed with 0.2% glutaraldehyde, rinsed in Tris buffer, treated for 10–20 min with 4.5 μl/ml nitro blue tetrazolium chloride (33 mg/ml in dimethylformamide; Sigma- Aldrich, St. Louis, MO), 3.5 μl/ml 5-bromo-4-chloro-3-indolyl phosphate (50 mg/ml in dimethylformamide; Sigma), 100 mm Tris (pH 9.5), 100 mm NaCl, 50 mm MgCl₂, and 0.1% Triton X-100, and counterstained with Nuclear Fast Red (GeneTex, Irvine, CA).

Embryonic development

Artificial decidualization

Female CD1 mice were mated with vasectomized males. On d 3.5 of pseudopregnancy, placement of P4 capsules and bilateral ovariectomy were performed as described above. At the time of ovariectomy, one uterine horn was infused with 10 μl of sesame oil (Sigma) to induce artificial decidualization. The noninfused uterine horn served as a control. Animals were randomized into two groups. Controls received ip injections of saline and treatment group animals received the DC101/mF4-31C1 combination. Uteri were removed on d 6.5 of pseudopregnancy, and processed for immunohistochemical evaluation of EC.

Histology, immunohistochemistry, and immunofluorescence

Frozen uteri containing embryos were sliced into 5-μm sections at 50-μm intervals and mounted on glass slides. Hematoxylin and eosin (H&E) staining and immunohistochemistry were performed according to standard procedures (19). The following primary antibodies were used for immunohistochemistry: 1) platelet endothelial cell adhesion molecule (PECAM; biotinylated rat antimouse CD31 antibody; 1:500; BD Pharmingen, San Diego, CA), 2) goat antimouse VEGFR-2 antibody (1:250; R&D Systems, Inc., Minneapolis, MN), 3) goat antimouse VEGFR-1 (1:200; R&D Systems, Inc.), and 4) goat antimouse VEGFR-3 (1:100; R&D Systems, Inc.). Incubation with biotinylated secondary antibodies (Vector Laboratories, Burlingame, CA) was followed by incubation with avidin and horseradish peroxidase-conjugated biotin (Vectastain Standard ABC Elite kit; Vector). The color reaction was performed with diaminobenzidine tetrahydrocholoride.

To study coexpression using immunofluorescence staining, the following combinations were used: 1) biotinylated rat antimouse CD31 (1:500; BD Pharmingen) in combination with VEGFR-1, -2, or -3 as described above; 2) coexpression of VEGFR-1/VEGFR-2: rat antimouse MF-1 (1:50; ImClone) with goat antimouse VEGFR-2 (1:250; R&D); 3) coexpression of VEGFR-2/VEGFR-3: goat antimouse VEGFR-2 (1:250; R&D) with rat antimouse VEGFR-3 (1:50; eBioscience, San Diego, CA); and 4) coexpression of VEGFR-1/VEGFR-3: goat antimouse VEGFR-1 (1:200; R&D) with rat antimouse VEGFR-3 (1:50; eBioscience). Sections were incubated with secondary antibodies and signals were visualized with fluorescent dye conjugates of streptavidin: AlexaFluor488 and AlexaFluor594 (Molecular Probes, Eugene, OR). After staining for one receptor, avidin-biotin-blocking solution (Vector) was used to prevent nonspecific staining while detecting expression of the second receptor. Images were photographed with an Axioscope 2 plus microscope camera (Zeiss, Göttingen, Germany) and Axiocam HRC digital (Zeiss). Double images were created with Adobe Photoshop Version 6.0 (Adobe Systems, Inc., San Jose, CA).

Cellular proliferation and apoptosis

One hour before they were killed, animals were injected with 5-bromo-2′-deoxyuridine (BrDU; 1 ml/100 g animal weight; Zymed Laboratories, Inc., San Francisco, CA). Proliferating cells were identified with a biotinylated BrDU antimouse primary antibody (Zymed kit). To detect proliferating EC, double staining for BrDU and PECAM was used (19). Apoptotic cells were detected using ApopTag Peroxidase In Situ Apoptosis Detection Kit (Chemicon International, Temecula, CA).

RT-PCR

Decidual tissue isolated from the embryo, extraembryonic structures, and myometrium was obtained from implantation sites on ED 6.5. Three hundred nanograms of total RNA extracted using TRIzol (Invitrogen, Carlsbad, CA) and treated with DNase I (Invitrogen) were reverse-transcribed using the SuperScript First-Strand Synthesis System for RT-PCR (Invitrogen). Primers to amplify VEGFA, VEGFC, and VEGFD were designed. Standard PCR with 25 amplification cycles was performed. PCR products were sequenced to confirm identity.

Serum P4

Blood was obtained from all animals by cardiopuncture immediately before they were killed. Serum P4 levels were measured using a competitive chemiluminescent immunoassay (Diagnostic Products Corp./Siemens, Los Angeles, CA).

Data analysis

Quantitative analysis of EC, stromal cells, cellular proliferation, and apoptosis was performed using an image analysis system linked to an Axioscope 2 plus microscope. The data were processed with AxioVision software version 4.5 (Zeiss). On ED 7.5, the area of decidua in a uterine cross-section was measured (in square millimeters). The percentage of the decidua occupied by blood vessels was calculated by dividing the total area of PECAM staining by the area of decidua multiplied by 100. The intense ring of PECAM staining around the embryo, the embryo, and the peripheral muscularis layer were excluded.

To assess proliferation, the decidua of a uterine cross-section was outlined. The image analysis system was configured to measure the number of brown-stained nuclei (i.e. BrDU-positive cells) using the segmentation part of the program. An adjacent section, stained with hematoxylin, was used to count the total number of the cells. Proliferation indices were calculated on ED 5.5 and 7.5 by dividing the number of BrDU-positive cells by the total number of cells multiplied by 100. This strategy was used to quantify apoptosis and stromal cell differentiation.

Statistics

Statistical analyses were performed using the Statistical Package for Social Science version 15.0 (SPSS, Inc., Chicago, IL). Data are presented as mean ± se. Unpaired Student’s t test and ANOVA with Bonferroni correction were used to compare sample means. P < 0.05 was considered a statistically significant difference.

Normal embryonic development is maintained after P4 replacement and bilateral ovariectomy

PROPmice were used to evaluate the effect of blocking VEGFR-1, -2, and -3 on uterine function during the peri-implantation period. To demonstrate that pregnancy progresses normally in this experimental model, PROPmice and intact pregnant controls received a single ip injection of saline on ED 3.75. On ED 10.5, there were no differences in uterine weights, numbers of embryos, gross embryo morphology, or maternal serum P4 levels between PROPmice and intact controls (Fig. 1).

Patterns of angiogenesis, cellular proliferation, and stromal cell differentiation are maintained in PROPmice

Blood endothelial cells (BEC) expressing PECAM are prominent throughout the endometrium and myometrium of the uterus on ED 3.5 before ovariectomy. After implantation, vascular density increases with the peak decidual vascular density between ED 6.5 and 7.5 (Fig. 2Q). Subsequently, vascular density declines with reduction most pronounced at sites of placenta formation (Fig. 2, A–E).

Intense staining for PECAM was noted in the uterine decidua directly surrounding the embryo on ED 5.5 and 6.5 (Fig. 2, B and C). This PECAM staining was observed in intact mice (supplemental Fig. 1, published as supplemental data on The Endocrine Society’s Journals Online web site at http://endo. endojournals.org) and has been noted by others (36). It does not show a vascular pattern or expression of VEGFRs, markers of EC. Thus, this PECAM staining does not reflect typical blood vessels. Some of the staining for PECAM might be due to the presence of leukocytes and/or apoptotic cells (supplemental Fig. 2).

Cellular proliferation in the decidua peaks on ED 5.5 (Fig. 2Q). The amount of cellular proliferation declines more than 50% on ED 6.5. On ED 8.5, there are few BrDU-positive cells in the decidua. The remaining proliferating cells are located at sites of placenta formation (Fig. 2, F–J). Whereas cellular proliferation occurs throughout the decidua, including the primary decidual zone (PDZ), on ED 5.5, cellular proliferation is restricted to the secondary decidual zone (SDZ) on ED 6.5 (Fig. 2, G and H). This pattern of proliferation in the decidua is characteristic of pregnancy in intact animals (37). Double staining with PECAM and BrdU identifies proliferating BEC among the population of the dividing cells on ED 3.5, 5.5, and 6.5 (Fig. 2, K–P).

To demonstrate that PROPmice undergo a “typical” decidual reaction, stromal differentiation was assessed by alkaline phosphatase staining on ED 6.5 (supplemental Fig. 3). These data indicate that decidual vascular events, cellular proliferation, and stromal differentiation in PROPmice are similar to that of intact pregnant animals.

Expression of VEGFRs and VEGFR ligands in peri-implantation uteri of PROPmice

VEGFRs are expressed in the uterus during the peri-implantation period. However, each receptor has a unique pattern of expression (Fig. 3). After implantation, VEGFR-1 and VEGFR-3 are both expressed in the PDZ (Fig. 3, B–E and L–O). Whereas the subset of cells expressing VEGFR-1 shows a circular peri-embryonic staining pattern, staining for VEGFR-3 is restricted to the mesometrial pole. VEGFR-2 expression is prominent in the PDZ on ED 5.5, but on ED 6.5 staining for VEGFR-2 is more prominent in the SDZ (Fig. 3, F and G). This pattern of expression is similar to the pattern for cellular proliferation, suggesting vascular proliferation (Fig. 2, G and H). Although both VEGFR-2 and VEGFR-3 are expressed in the mesometrial pole on ED 7.5, the subset of cells expressing these receptors differs (Fig. 3, D and N). Immunofluorescent double labeling demonstrated that VEGFRs are coexpressed with PECAM (Fig. 3, E, J, and O), and VEGFR-1 and VEGFR-3 are coexpressed with VEGFR-2 (supplemental Fig. 4). RT-PCR was used to detect mRNA expression of VEGFR ligands, VEGFA, VEGFC, and VEGFD, on ED 6.5 (supplemental Fig. 5).

Effect of VEGFR-2 (DC101), VEGFR-3 (mF4-31C1), or VEGFR-1 (MF-1) neutralization on blood vessel development, cellular proliferation and apoptosis

After DC101, mF4-31C1, and MF-1 treatment, the percentages of the decidua occupied by blood vessels were 9.1 ± 1.0%, 13.1 ± 0.7%, and 17.2 ± 1.1%, respectively, whereas they were 20.8 ± 0.9% in the controls (Fig. 4G). Neutralization of VEGFR-2 and VEGFR-3, but not VEGFR-1 significantly reduced the percentage of the decidua occupied by blood vessels (P < 0.05). To determine whether blocking VEGFR-2 and VEGFR-3 would have an additive effect, we administered a combination DC101/mF4-31C1 treatment. Compared with DC101 alone, there was no statistically significant difference in the percentage of the deciduae occupied by blood vessels (7.2 ± 0.9%) after DC101/mF4-31C1 treatment (Fig. 4G).

Histological analyses of uterine cross-sections demonstrated that treatment with DC101 significantly reduced decidual vasculature and stromal cell differentiation (controls 43.2 ± 4.7%; DC101 22.9 ± 2.4%) (supplemental Fig. 3) and arrested embryonic development. Embryonic structures were absent after DC101 treatment. In contrast, a normal-appearing, intact embryo was detected after mF4-31C1 treatment (Fig. 4). The attached embryo was noted on ED 5.5 and 6.5 after DC101 treatment (supplemental Fig. 6). Thus, we concluded that the reduction in decidual vasculature was not due to failed implantation.

To investigate the mechanism(s) responsible for the reduction in decidual blood vessels, cellular proliferation was assessed on ED 5.5 and 7.5, and apoptosis was assessed on ED 7.5. Data for ED 5.5, the time of peak proliferation in untreated animals, and ED 7.5 show similar trends. Neutralization of VEGFR-2 or VEGFR-3 significantly reduced cellular proliferation in the decidua (P < 0.05) (Fig. 4G). In control animals, 43.6 ± 0.6% of the decidual cells were proliferating. The percentages of proliferating cells after DC101, mF4-31C1, MF1, and DC101/mF4-31C1 combination treatment were 25.7 ± 0.9, 37.5 ± 1.2, 35.4 ± 0.01, and 27.7 ± 1.6%, respectively. The reduction in cellular proliferation with DC101 or DC101/mF4-31C1 combination therapy was significantly greater than with mF4-31C1 treatment (P < 0.05). Representative uterine cross-sections on ED 5.5 for DC101 and mF4-31C1 are presented (Fig. 4, E and F). After treatment with DC101 or mF4-31C1, the percentages of proliferating cells and proliferating BEC (Fig. 4, E and F) were reduced. On ED 7.5, 0.12% of decidual cells were apoptotic in the controls. None of the treatments significantly increased the percentage of apoptotic cells (<0.3%) when compared with controls. We concluded that a reduction in cellular proliferation, not an increase in apoptosis, is the primary mechanism by which VEGFR neutralization leads to decreased decidual vasculature.

Effect of VEGFR-2 (DC101), VEGFR-3 (mF4-31C1), or VEGFR-1 (MF-1) neutralization on embryonic development

Administration of DC101, but not mF4-31C1 or MF-1, to PROPmice on ED 3.75 disrupted pregnancy development evaluated on ED 10.5 (Fig. 5). Each single ip injection of DC101, mF4-31C1, or MF-1 resulted in a uterus with a normal number of implantation sites. The gross and microscopic appearances of uteri from mice treated with mF4-31C1 or MF-1 were indistinguishable from those of the controls (Fig. 5). Uteri from DC101-treated mice appeared dark, possibly consistent with hemorrhaging in the implantation sites. H&E staining of uterine cross-sections confirmed abnormal embryonic development, because no normal embryonic structures were detected (Fig. 5G). Effects of DC101 on embryonic development were seen as early as ED 6.5 (supplemental Fig. 6).

In contrast, neutralization of VEGFR-2 starting on ED 6.5 did not affect pregnancy development on ED 10.5. Statistically identical numbers of embryos (controls 10.7 ± 1.2; DC101 13.7 ± 0.9) and embryo weights (controls 0.138 ± 0.01 g; DC101 0.110 ± 0.02 g) were noted. Thus, the timing of the antibody injections is critical (16).

Neutralization of VEGFR-2 reduces uterine vasculature in a model of artificial decidualization

An intraluminal injection of 10 μl sesame oil was used to induce artificial decidualization in pseudopregnant PROPmice. Decidual morphology and peri-implantation blood vessel development were evaluated on ED 7.5 (Fig. 6, A and B). DC101/mF4-31C1 combination therapy reduced decidual vasculature and led to abnormal-appearing, small deciduae (Fig. 6, D–F). These results are similar to those observed after DC101 treatment in PROPmice (Fig. 4B). This finding indicates that the VEGF/VEGFR-2 pathway might have a regulatory role in angiogenesis during artificial decidualization.