| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Original Articles |
Medical Research Council Human Reproductive Sciences Unit (C.W., H.W., S.F.L., H.M.F.), Edinburgh, Scotland, United Kingdom EH3 9ET; Regeneron Pharmaceuticals, Inc. (J.S.R., S.J.W.), Tarrytown, New York 10591
Address all correspondence and requests for reprints to: Dr. C. Wulff, Medical Research Council Human Reproductive Sciences Unit, 37 Chalmers Street, Edinburgh, Scotland, United Kingdom EH3 9ET. E-mail c.wulff{at}hrsu.mrc.ac.uk
Abstract
The possibility of stimulating or inhibiting paracrine factors regulating angiogenesis may lead to new approaches for the treatment of pathological conditions of the female reproductive tract. We examined the effects of a clinical candidate, a soluble truncated form of the Flt-1 receptor, vascular endothelial growth factor trapA40 (VEGF trap), in a primate model to determine its ability to prevent the onset of luteal angiogenesis or intervene with the on-going process. Marmosets were treated from the day of ovulation until luteal day 3 (prevention regimen) or on luteal day 3 for 1 day (intervention regimen). Effects of VEGF inhibition were studied by obtaining a proliferation index using bromodeoxyuridine incorporation, quantifying endothelial cell area using CD31, and assessing luteal function by plasma progesterone. After both treatments, intense luteal endothelial proliferation was suppressed, a concomitant decrease in endothelial cell area confirmed the inhibition of vascular development, and a marked fall in plasma progesterone levels showed that luteal function was compromised. In situ hybridization was used to localize and quantify compensatory effects on the expression of angiogenic genes. VEGF messenger ribonucleic acid (mRNA) expression in luteal cells was increased, whereas expression of its receptor, Flt, was decreased. Inhibition of VEGF resulted in localized increased expression of angiopoietin-2 mRNA and its receptor, Tie-2.
The results show that the VEGF trap can prevent luteal angiogenesis and inhibit the established process with resultant suppression of luteal function. Luteal Flt mRNA expression is dependent upon VEGF, and VEGF inhibition results in abortive increases in expression of VEGF, angiopoietin-2, and Tie-2.
THE RAPID, CONTROLLED, and cyclical nature of angiogenesis in the female reproductive tract suggests that interference with this process should provide a novel approach to manipulation of reproductive function. Many factors involved in the regulation of angiogenesis have been identified, and the possibility of stimulating or inhibiting these paracrine control mechanisms is being addressed using current advances in the development of angiogenic and antiangiogenic compounds (1). It is anticipated that inhibition of angiogenesis will have application in reproductive medicine for the treatment of conditions such as ovarian hyperstimulation syndrome (2, 3), endometriosis (4, 5, 6), and ovarian cancer (7, 8).
As comparatively little is known about the effect of angiogenesis inhibitors on the ovary and uterus, there is an urgent need for relevant investigations to be performed in appropriate nonhuman primate models. We have shown that the primate corpus luteum may provide such a model. As in women (9), luteinization in the marmoset is followed by a period of intense angiogenesis (10), and the determination of endothelial cell proliferation after antiangiogenic treatment in vivo is a sensitive index of inhibitory activity (11).
One of the most promising targets for inhibiting angiogenesis is to prevent the action of vascular endothelial growth factor (VEGF). In this study we describe the inhibition of VEGF with a soluble truncated form of the Flt receptor, Flt-1-Fc (VEGF trapA40), which has been developed with a view to clinical application. VEGF trapA40 has a high affinity for all forms of endogenous VEGF preventing receptor binding. In the current study we examined the effects of VEGF trap at two different stages of the luteal phase; first by treatment immediately after ovulation to determine its ability to prevent the onset of angiogenesis, and second by administration on the third day after ovulation to investigate its ability to intervene with the established angiogenic process and luteal function. In addition, to gain insight into the paracrine regulation of luteal angiogenesis we investigated the changes in expression patterns of VEGF and its Flt receptor after treatment using quantitative in situ hybridization. We also examined the effects of VEGF deprivation on the expression of a second growth factor family specific for the vascular endothelium (12, 13, 14), the angiopoietins, which are believed to have an important function in ovarian angiogenesis (15, 16). Angiopoietin-2 (Ang-2), acting via its receptor Tie-2, is believed to destabilize endothelial cell contacts, which may, in turn, facilitate responsiveness to angiogenic factors such as VEGF.
Materials and Methods
Flt-1-Fc (VEGF trapA40)
The Flt-1-Fc (VEGF trapA40) used in these
experiments comprised a portion of the extracellular domain of Flt-1
containing Ig repeats 1–3, fused to the Fc portion of human IgG. The
Flt-1-Fc was expressed in CHO cells, and the protein was purified by
protein A affinity chromatography, followed by size-exclusion
chromatography. The recombinant Flt-1-Fc was then chemically modified
to improve the pharmacokinetic profile of the parent molecule without
affecting its ability to bind VEGF with high affinity. CHO-derived
parental Flt(1–3Ig)Fc was incubated with sulfo-NHS-acetate
(Pierce Chemical Co., Rockford, IL) in PBS/5% glycerol,
pH 7.2, such that the acetylation reagent was present in a 40-fold
molar excess. The acetylation reaction specifically modifies the
-amino group of lysines present in the parental molecule. The
mixture was placed on a rocker and incubated overnight at room
temperature. The acetate-modified Flt(1–3Ig)Fc, termed FltFc A40, was
then extensively dialyzed against PBS/5% glycerol using 25-kDa
molecular mass cut-off tubing. After dialysis, the concentration was
checked by optical density at 280 nm, and modification was assessed by
isoelectric focusing analysis. With this modification the pI
shifts from about 9.5 for parental FLT(1–3Ig)Fc to 5.8–6.5 for FltFc
A40. A variety of in vitro assays revealed no measurable
decrease in binding affinity for the modified protein. The purity of
the modified recombinant protein was judged to be more than 95% by
Coomassie blue-stained SDS-PAGE. The protein was filter-sterilized and
stored at a concentration of 4.9 mg/mL in PBS, pH 7.2, containing 5%
glycerol at -20 C.
Enzyme-linked immunosorbent assay (ELISA)
Levels of circulating unbound VEGF trapA40 in plasma were assessed in the blood sample collected when the marmosets were killed 24 h after the final injection in animals receiving the 3-day treatment. VEGF trapA40 was measured using a functional ELISA. ELISA plates were coated with 2 µg/mL CHO-derived human VEGF 165 in carbonate/bicarbonate buffer overnight at 4 C. Plates were washed in KPL buffer (Kirkegaard & Perry Laboratories, Gaithersburg, MD) and blocked with 0.2% I-block (Tropix, Perkin-Elmer Corp. PE Applied Biosystems, Bedford, MA) for 1 h. Standards and serial dilutions of serum samples were prepared in Kirkegaard & Perry Laboratories diluent and incubated at room temperature for 2 h. Plates were washed and incubated with goat antihuman Fc bound to horseradish peroxidase (Sigma, St. Louis, MO) and incubated for 1 h at room temperature. The substrate used was 3,3',5,5'-tetramethyl benzidine (Sigma), which was developed at room temperature for 30 min and stopped with 2 mol/L H2SO4. Plates were read at 450/579 nm. The detection limit of the assay was 150 pg/mL. The value for the bound VEGF trapA40/VEGF cocomplex was obtained using the following variation to the above ELISA: plates were coated with polyclonal anti-VEGF antibody (R & D Systems, Inc., Minneapolis, MN) as a capture antibody rather than coating with VEGF protein.
Animals and treatment
Adult female common marmoset monkeys (Callithrix jacchus) with a body weight of approximately 350 g and regular ovulatory cycles were housed together with a younger sister or prepubertal female as described previously (17). Blood samples were collected three times per week by femoral venipuncture without anesthesia, and plasma was stored at -20 C until subjected to progesterone assay to confirm normal ovulatory cycles.
Experiments were carried out in accordance with the Animals (Scientific
Procedures) Act, 1986. To synchronize the timing of ovulation, animals
were treated with 1 µg PGF2
analog
(cloprostenol, Planate, Coopers Animal Health Ltd., Crewe, UK), im,
during the mid- to late luteal phase of the pretreatment cycle to
induce luteolysis. This treatment is normally followed by ovulation 10
days later (18).
In the first experiment (prevention regimen), four marmosets were treated with VEGF trap at a dose of 25 mg/kg; approximately 1 mL was injected at two sc sites over the abdominal region. Four controls were treated with vehicle alone. Treatment commenced on the day of presumed ovulation (luteal day 0) and was repeated on days 1 and 2. When results showed that the treatment was effective in preventing angiogenesis and plasma concentrations of VEGF trap were high 24 h after injection, a second experiment was designed to investigate whether the on-going early luteal angiogenesis could be inhibited. Four animals were treated with VEGF trap on day 3 of luteal phase (intervention regimen). As pharmacokinetic studies have shown that the VEGF trap is released from the sc site over a period of several hours, VEGF trap was administered iv at a dose of 12.5 mg/kg to ensure maximal bioavailability during the 24-h study period in this experiment. Four controls were similarly treated with vehicle alone.
Twenty-four hours after the last treatment, the animals were injected iv with 20 mg bromodeoxyuridine (BrdU; Roche Molecular Biochemicals, Essex, UK) in saline. One hour later, the animals were sedated using 100 µL ketamine hydrochloride (Parke-Davis Veterinary, Pontypool, UK) im and killed with an iv injection of 400 µL Euthetal (sodium pentobarbitone, Rhone Merieux, Harlow, UK). After cardiac exsanguination via a heparinized syringe, ovaries were removed immediately, weighed, and fixed in 4% paraformaldehyde. After 24 h, the ovaries were dehydrated and embedded in paraffin. Five-micron sections were cut and stained with hematoxylin and eosin for morphological evaluation.
Progesterone assay
The occurrence of ovulation and a normal luteal phase length
(18–22 days) were based on determination of plasma progesterone
concentrations measured directly using an ELISA that is a modification
of an assay we have used previously (19). The label was
prepared by biotinylation of progesterone-11-glucuronide, and the
primary antibody was rabbit antiprogesterone BSA, whose cross-reaction
with pregnenolone and other major steroids was less than 0.1%.
Microtiter plates were coated with purified donkey antirabbit Ig. After
washing, samples and standards were added along with
anilino-1-naphthalene sulfonic acid and incubated for an additional
3 h. After washing, streptavidin-horseradish peroxidase was added
in 1% casein and incubated for 1 h. Following a subsequent wash,
color was developed using o-phenylene diamine. The detection
limit of the assay was 4.6 nmol/L, and interassay coefficients of
variation for low, medium, and high level quality controls were 11.6%,
7.0%, and 12.6%, respectively.
Immunocytochemistry
Cellular responses were studied by 1) quantifying the number of mitotic cells stained for BrdU, 2) examining the establishment of the microvascular network using CD31 staining to identify endothelial cells, and 3) determining the incidence of colocalization of BrdU and CD31.
Tissue sections (5 µm) were cut onto Tepsa-coated (Sigma) slides for immunocytochemistry. Sections were dewaxed in xylene, rehydrated in descending concentrations of ethanol, and washed in distilled water. Antigen retrieval was performed by boiling sections in a Tefal Clypso pressure cooker (Tefal, Essex, UK) in 0.01 mol/L citrate buffer, pH 6, for 6 min at high pressure setting 2. Slides were then left for 20 min in hot buffer and washed in TBS (0.05 mol/L Tris and 9 g/L NaCl). To reduce nonspecific binding, sections were blocked in normal rabbit serum (diluted 1:5 in TBS) for 30 min. As primary antibodies CD31 (monoclonal, diluted 1:20 in TBS; DAKO Corp., Copenhagen, Denmark) or BrdU (monoclonal, diluted 1:30 in TBS; Roche Molecular Biochemicals, Mannheim, Germany) were used. Incubation was carried out overnight at 4 C. Slides were then washed three times in TBS. Incubation with the secondary antibody (rabbit antimouse Ig, DAKO Corp., 1:60 diluted in TBS) was for 40 min at room temperature followed after two washes in TBS by incubation of the alkaline phosphatase-antialkaline phosphatase complex (1:100 dilution in NRS and TBS; DAKO Corp.) for 40 min at room temperature. Visualization was performed using NBT solution containing 45 µL NBT substrate (Roche Molecular Biochemicals), 10 mL NBT buffer, 35 µL 5-bromo 4 chloro 3 indolyl-phosphate (Boehringer, Lewes, East Sussex, UK) and 10 µL levamisole. Sections for BrdU were counterstained with hematoxylin, whereas sections for CD31 were not counterstained so that quantitative image analysis could be performed. For dual labeling, slides were incubated first with CD31 and visualized with Fast Red (Sigma), followed by incubation with BrdU and then visualized with NBT as described above. Fast Red solution contained 1 mg Fast Red/1 mL buffer (20 mg naphtol AS-MX phosphate, 2 mL dimethyl formamide, and 98 mL 0.1 mol/L Tris, pH 8.2).
In situ hybridization
To investigate effects of the 3-day treatment on the expression of related angiogenic factors, in situ hybridization was performed as described previously (16) using complementary ribonucleic acid (RNA) probes for human VEGF, Flt, Ang-2, and Tie-2. Sense and antisense probes were prepared using an RNA transcription kit (Ambion, Inc., Austin, TX) and were labeled with [35S]uridine 5'-triphosphate (NEN Life Science Products, Boston, MA). Deparaffinized sections were treated with 0.1 N HCl and then digested in proteinase K (5 µg/mL; Sigma) for 30 min at 37 C. After prehybridization for 2 h at 50 C for VEGF and 55 C for Flt, Ang-1, Ang-2, and Tie-2, subsequent hybridization was performed in a moist chamber overnight at 50 or 55 C. High stringency posthybridization washings and ribonuclease A treatment were used to remove excess probe. Slides were then dehydrated, dried, and dipped in Ilford G5 liquid emulsion (H. A. West, Edinburgh, UK). Exposure times for VEGF, Flt, Ang-2, and Tie-2 were 2, 4, 7, and 7 weeks, respectively. Slides were subsequently developed (Kodak D19 developer, Eastman Kodak Co., Rochester, NY) and fixed (Kodak GBS). All slides were counterstained with hematoxylin (Richard-Allan, Richland, MI), dehydrated, and mounted.
Analysis of data
Quantitative analysis for BrdU and CD31 immunocytochemistry and in situ hybridization was performed using an image analysis system linked to an Olympus Corp. camera (New Hyde Park, NY), and the data were processed using the Image-Pro Plus (version 3.0) for Windows computer program.
BrdU labeling. A proliferation index (i.e. BrdU-positive cells expressed as a percentage of the total number of cells) was calculated for each corpus luteum. Ten randomly chosen fields (6 x 105 µm2) per corpus luteum at x200 magnification were analyzed. The proliferation index was expressed as a mean value for the number of fields assessed.
CD31 labeling. The endothelial cell area (i.e. CD31-positive cells) was measured in 10 randomly chosen fields (6 x 105 µm2) at x200 magnification in each corpus luteum. The captured gray scale image was thresholded and converted to a binary image, and the CD31 area was quantified. The CD31 area was expressed as a mean value for the number of fields assessed.
In situ hybridization. Slides were analyzed qualitatively under lightfield and quantitatively under darkfield conditions. Two parameters were used for quantification: grain density (number of grains per µm2) as a value for gene expression in the cells of interest and area of expression as a value for the proportion of messenger RNA (mRNA)-expressing cells. Grain density and total area of mRNA expression within 6–10 randomly selected fields (3.2 x 105 µm2) at x400 magnification were measured. The positive area of expression was calculated as the mean value for the number of fields assessed.
Statistical analysis
Effects of treatments were determined using a two-tailed unpaired t test, with a 95% confidence interval. Progesterone data were subjected to two-factor ANOVA, followed by post-hoc test. The tests were performed using SPSS software (version 6.1, SPSS, Inc., Chicago, IL). All values are given as the mean ± SEM.
Results
Plasma levels of VEGF trapA40
Plasma samples taken 24 h after the third and final injection of VEGF trap in the prevention regimen contained high concentrations of the compound (25–61 µg/mL). Circulating cocomplexes comprising VEGF trap bound to VEGF also were detected in plasma, albeit at much lower levels (4.0–26.5 ng/mL), indicating that more than 99.9% of the circulating VEGF trap was unmodified and remained available for neutralization of endogenous VEGF.
Ovarian weight and morphology
Prevention regimen. Total ovarian weights in the 3-day VEGF
trap-treated group (n = 4) were significantly less (P =
0.011) than those in controls (n = 4; 116 ± 20 and 255
± 87 mg, respectively). By gross inspection, corpora lutea in ovaries
of treated animals exhibited reduced hyperemia. Examination of sections
stained with hematoxylin and eosin revealed the presence of fresh
corpora lutea in all animals. The lutein cells of controls were
uniformly fully differentiated and polyhedral in shape, with abundant
cytoplasm and centrally located nucleus. In contrast, the appearance of
lutein cells in treated animals differed from one another depending on
localization in the corpus luteum. The cytoplasmic volume of lutein
cells proximal to the luteal cavity was noticeably restricted compared
with that of control tissue (Fig. 1
, A
and B). In corpora lutea of treated animals, nuclear fragments
were observed primarily within the central area of the corpus luteum,
whereas none was observed in the control corpora lutea (Fig. 1, C and
D).
|
Progesterone
In the control animals of both treatment regimens, the plasma
progesterone concentration rose significantly during the study period
(P < 0.001). No increase was recorded in the 3-day
treated animals (Fig. 2); whereas there
was no difference in progesterone concentrations between control and
treatment groups at the onset of treatment, there was a significant
difference between control and treatment groups by day 2, resulting
from the rise in progesterone in the controls. The significance of this
increase was enhanced on day 3. In the animals treated on luteal day 3,
progesterone levels rose similarly to those in control animals from
days 0–3 of the luteal phase. After the single injection of VEGF trap,
progesterone levels declined significantly (P < 0.001;
Fig. 3).
|
|
Cell proliferation, as shown by BrdU staining, was present in all
corpora lutea, with positive staining restricted largely to cells
surrounding the lumen of microvessels, and others whose appearance was
consistent with that of capillary endothelial cells. However, although
endothelial cell proliferation was clearly high within the corpora
lutea from all control marmosets, it was markedly reduced by the 3-day
VEGF trap treatment (Fig. 4, A and B) and
after a single injection of VEGF trap on day 3, as evidenced by a
significant decrease in the proliferation index in both experiments
(Figs. 4G and 5A).
|
|
CD31 immunostaining demonstrated the development of a
microvasculature in all corpora lutea. However, the endothelial cell
area was markedly reduced in the corpora lutea of 3-day VEGF
trap-treated animals (Fig. 4
, C and D). Larger vessels in the
peripheral zone of the former theca were relatively unaffected compared
with the internal microvasculature. The 3-day treatment reduced
endothelial cell area by approximately 75% (Fig. 4H), whereas after a
single injection the endothelial cell area was reduced by 44% (Fig. 5B).
Colocalization of BrdU and CD31
The BrdU-positive cells in corpora lutea of both control groups
were almost always dually stained. In the treated group, the
proliferating cells present were almost entirely of endothelial cell
origin (Fig. 4, E and F).
In situ hybridization
VEGF mRNA was highly expressed in all recently formed
corpora lutea in both control and treated animals (Fig. 6, A and B); expression was confined to
the lutein cells. Analysis of grain distribution and density revealed
that the total area of VEGF expression per field was not significantly
different between the groups, but in VEGF trap-treated marmosets the
grain density was significantly increased (Fig. 6, C and D), indicating
that the gene expression for VEGF was increased after treatment.
|
, A and B).
Quantification revealed that the area of expression significantly
decreased after VEGF trap treatment (Fig. 7C), corresponding to the
decrease in the number of endothelial cells in the corpus luteum. The
grain density per cell was also decreased in the corpus luteum of VEGF
trap-treated animals (Fig. 7D), indicating a switching off of the Flt
gene.
|
, A and B). Consistently, the area of
mRNA expression was significantly higher in treated animals than in the
controls (Fig. 8C). A tendency toward an increase in grain density
after treatment was not significant because of high variation in the
control group (Fig. 8D).
|
, A and B). In contrast to Flt
expression, Tie-2 mRNA was expressed in larger capillaries, whereas in
the microvasculature expression was very low (Fig. 9A). No differences
in the area of expression were found between treated and control
animals (Fig. 9C), but after treatment a significant increase in grain
density was observed (Fig. 9D), indicating increased gene
expression.
|
This study shows that treatment of marmosets with a soluble form of the Flt receptor (VEGF trap) to neutralize the actions of VEGF had a dramatic inhibitory effect on early luteal angiogenesis. Treatment starting on the day of ovulation prevented the onset of the angiogenic process, as demonstrated by the marked decline in endothelial cell proliferation and the dearth of the microvascular tree on luteal day 3. Furthermore, by allowing angiogenesis to become established and delaying treatment until luteal day 3, we found that a single treatment with VEGF trap for 24 h can inhibit ongoing angiogenesis. Both treatments resulted in inhibitory effects upon luteal function, as demonstrated by the rapid decline in plasma progesterone, and the 3-day treatment was found to have altered the expression of VEGF, Flt, Ang-2, and Tie-2 within the corpus luteum. These findings suggest that anti-VEGF therapy may have potential for clinical application in ovarian disorders characterized by hypervascularity and hyperplasia.
The present results confirm and extend our previous observations using a mouse monoclonal antibody to VEGF to prevent the onset of angiogenesis in this model (11). Comparison of the results of the two treatments reveals that the VEGF trap produced more profound effects, resulting in a greater suppression of ovarian weight, endothelial cell proliferation, and endothelial cell area. This indicates that VEGF trap is a more effective way of VEGF inhibition than use of a VEGF antibody; however, full dose-response studies would be required to establish this observation.
The success of a single injection of VEGF trap on luteal day 3 in disrupting the ongoing angiogenic process confirms the importance of VEGF in luteal angiogenesis. However, the suppression of progesterone levels at this stage was somewhat surprising. It might have been expected that the suppression of angiogenesis at this point when a proportion of the luteal vasculature is established would be associated with a prevention of the normal increase in plasma progesterone rather than a decrease. Rapid actions of VEGF are attributable to its effects on permeability (20, 21), so this decline in progesterone may indicate that inhibition of VEGF results in a suppression of ovarian vascular permeability with consequent effects on the availability of hormone precursors and gonadotropins, thereby contributing to reduced progesterone synthesis. The reduction in progesterone secretion after the longer term treatment is probably the result of additional mechanisms relating to impairment of vascular development, including reduced delivery of essential cellular requirements, reduced removal of synthesized progesterone secondary to the vascular impairment, or reduced synthesis of progesterone as a result of incomplete differentiation of lutein cells.
A low level of angiogenesis persisted despite the treatment with VEGF trap from ovulation to day 3 of the luteal phase. It is unlikely that this was the result of incomplete inhibition of VEGF, because over 99% of the unbound VEGF trap remained available in the circulation after the final injection. It may be that the residual angiogenesis reflects the activity of other angiogenic factors. A first step to investigate secondary effects of VEGF inhibition on the expression of other angiogenic factors and their receptors was undertaken in the current study by the use of in situ hybridization on the ovaries of the marmosets receiving 3-day treatment. VEGF mRNA was localized to the hormone-producing cells of the corpus luteum, as observed in the human and cynomolgus monkey corpus luteum (16, 22); there was no expression apparent in endothelial cells. The grain density in the lutein cells was increased significantly in the VEGF trap-treated marmoset corpora lutea. This suggests that the lutein cells may attempt to compensate for reduced VEGF in the periphery by increasing its synthesis. As it has been shown that hypoxia stimulates VEGF synthesis in many tissues (23, 24, 25, 26), it reasonable to suggest that up-regulation of VEGF after treatment is at least in part driven by a hypoxic response within the corpus luteum resulting from the dearth of the vasculature. The VEGF receptor Flt was expressed in the luteal endothelium and a marked decrease in the area of expression was observed after treatment commensurate with the reduction in the number of microvessels. Furthermore, as the Flt mRNA grain density in endothelial cells also decreased after treatment with VEGF trap, this indicates that VEGF regulates its own receptor expression in the corpus luteum as observed in other tissues (27, 28, 29) and in the rat corpus luteum (30). It is also possible that the lack of the endothelial cell survival factor VEGF in the periphery causes a dysfunction of endothelial cells, followed by a decrease in Flt synthesis.
Little is known about the interplay between VEGF and the synthesis of the angiopoietins and their Tie-2 receptor. The current study provided an opportunity to begin to explore this question by examining the effects of deprivation of VEGF on luteal expression of Ang-2 and Tie-2 mRNA. After treatment, Ang-2 mRNA expression was increased, especially in regions proximal to the luteal cavity. Morphologically, this area was associated with a smaller lutein cell size than in controls, and an increase in nuclear fragments, indicating cell death, was observed, suggesting that this area is susceptible to the greatest hypoxia and nutrient deprivation. In addition, an increased expression of Tie-2 mRNA was detected after treatment. Up-regulation of Ang-2 and Tie-2 under hypoxic conditions has also been shown in a focal cerebral ischemia model in the rat (31) and in human endometrial endothelial cells (32). Thus, hypoxia may also drive up-regulation of Ang-2 in the corpus luteum after VEGF trap treatment. Other studies have shown that areas of punctuate Ang-2 expression in combination with VEGF expression are associated with active angiogenesis (16, 33). Therefore, it may be suggested that the increased mRNA for both Ang-2 and its receptor, Tie-2, together with elevated levels of VEGF expression are an attempt to attract capillary ingrowth into this presumably hypoxic area. An alternative explanation of increased Ang-2 mRNA expression is that Ang-2 may be required during angiolysis, as Ang-2 causes vascular regression in the absence of growth and survival signals such as VEGF (15, 34).
Finally, turning to the clinical relevance of our observations with respect to the reproductive system, the ability of VEGF trap to prevent the development of the luteal microvasculature, reduce ovarian weight, and rapidly suppress luteal function strongly suggests that both ovarian angiogenesis and permeability are targets for VEGF inhibition. This suggests that that at least two ovarian disorders, the ovarian hyperstimulation syndrome and ovarian cancer, both of which are associated with pathological angiogenesis, hypervascularity, and hyperpermeability, may be responsive to VEGF inhibition with VEGF trap.
Acknowledgments
We are grateful to Dr. D. S. Charnock-Jones (University of Cambridge, Cambridge, UK) for the gift of complementary DNA probes for VEGF and Flt, K. Morris and staff for animal care, F. Pitt and I. Swanston for progesterone assays, Dr. N. Papadopoulos for preparation of the chemically modified version of the VEGF trapA40, and D. Hylton for assistance with the ELISA.
Footnotes
1 This work was supported in part by a grant (to C.W.) from Deutsche
Forschungs-gemeinschaft. 
Received December 15, 2000.
Revised March 1, 2001.
Accepted March 15, 2001.
References
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
) and 3-day VEGF trap-treated marmosets (•). Treatment
started on the day of expected ovulation (day 0) and continued for 3
days. Treatment was associated with a significant suppression beginning
on day 2 and enhanced on day 3 (P = 0.001).
) and VEGF
trap-treated (
) marmoset corpus luteum. H, Quantitative analysis of
CD31 immunostaining in control (
