0013-7227/08/$15.00/0
`Printed in U.S.A.
`
`Endocrinology 149(9):4413– 4420
`Copyright © 2008 by The Endocrine Society
`doi: 10.1210/en.2008-0325
`
`The Role of Vascular Endothelial Growth Factor and
`Estradiol in the Regulation of Endometrial Angiogenesis
`and Cell Proliferation in the Marmoset
`
`Downloaded from https://academic.oup.com/endo/article/149/9/4413/2455652 by guest on 29 June 2022
`
`Hamish M. Fraser, Helen Wilson, Audrey Silvestri, Keith D. Morris, and Stanley J. Wiegand
`Medical Research Council Human Reproductive Sciences Unit (H.M.F., H.W., A.S., K.D.M.), University of Edinburgh Centre
`for Reproductive Biology, The Queen’s Medical Research Institute, Edinburgh EH16 4TJ, United Kingdom; and Regeneron
`Pharmaceuticals (S.J.W.), Tarrytown, New York 10591
`
`The present studies explore the roles of vascular endothelial
`growth factor (VEGF) and estradiol on angiogenesis and stro-
`mal and epithelial cell proliferation in the marmoset endo-
`metrium during the proliferative phase of the ovulatory cycle.
`At the start of the proliferative phase, marmosets were 1)
`treated with vehicle, 2) treated with a VEGF inhibitor (VEGF
`Trap, aflibercept), 3) ovariectomized, 4) ovariectomized and
`given replacement estradiol, or 5) treated with VEGF Trap
`and given replacement estradiol. The uterus was examined
`10 d later in the late proliferative phase. Changes in endothe-
`lial and epithelial cell proliferation were quantified using a
`volumetric density method after immunohistochemistry for
`bromodeoxyuridine to localize proliferating cells, CD31 to vi-
`sualize endothelial cells, and dual staining to distinguish en-
`dothelial cell proliferation. Endothelial proliferation was el-
`
`evated in late proliferative controls but virtually absent after
`VEGF Trap. Ovariectomy had a similar inhibitory effect,
`whereas angiogenesis was restored by estrogen replacement.
`Estradiol replacement in VEGF Trap-treated marmosets re-
`sulted in only a small increase in endothelial cell proliferation
`that remained significantly below control values. VEGF Trap
`treatment and ovariectomy also markedly reduced stromal cell
`proliferation but resulted in increased stromal cell density as-
`sociated with a reduction in overall endometrial volume. Estro-
`gen replacement in both ovariectomized and VEGF Trap-
`treated animals restored stromal proliferation rates and cell
`density. These results show that endometrial angiogenesis and
`stromal proliferation during the proliferative phase are driven
`by estradiol and that the effect of estrogen on angiogenesis is
`mediated largely by VEGF. (Endocrinology 149: 4413–4420, 2008)
`
`PHYSIOLOGICAL ANGIOGENESIS, the formation of
`
`new blood vessels from preexisting capillaries, is rare
`in the adult but takes place on a regular cycle in the ovaries
`and uterus. Vascular endothelial growth factor (VEGF) is a
`potent stimulator of endothelial cell proliferation and per-
`meability (1). The availability of potent specific antagonists
`of VEGF such as VEGF Trap (2), allows for the physiological
`role of VEGF to be investigated by in vivo inhibition. In a
`series of experiments, we have described the role of VEGF in
`the ovary using the marmoset monkey (Callithrix jacchus), a
`commonly used primate in reproductive research, as a
`model. Treatment during the 10-d follicular phase resulted in
`a marked suppression of follicular angiogenesis and follic-
`ular development so that ovulation was prevented (3). When
`administered during the luteal phase, VEGF Trap also mark-
`edly suppressed luteal angiogenesis and function (4). These
`studies established that VEGF is essential for normal follic-
`ular and luteal angiogenesis and development.
`VEGF and its receptors are expressed in the primate uterus
`and are likely to be involved in the regulation of uterine
`angiogenesis and permeability (5, 6). However, the role of
`VEGF in the primate uterus has yet to be studied by direct
`experimentation in vivo. Such studies carried out in the ovari-
`
`First Published Online May 22, 2008
`Abbreviations: BrdU, Bromodeoxyuridine; VEGF, vascular endothe-
`lial growth factor.
`Endocrinology is published monthly by The Endocrine Society (http://
`www.endo-society.org), the foremost professional society serving the
`endocrine community.
`
`ectomized, estrogen-treated mouse reported that inhibition
`of VEGF prevented endothelial cell proliferation in one study
`(7) or had no effect on angiogenesis but inhibited edema and
`epithelial cell proliferation in another (8). In the immature
`rat, VEGF immunoneutralization was also shown to block
`estrogen-induced edema (9) and, in the adult, to prevent
`implantation (10). The purpose of the present investigation
`was to examine the changes in the marmoset endometrium
`after treatment with VEGF Trap with respect to endothelial,
`stromal, and epithelial cell proliferation. Because normal en-
`dometrial function is dependent upon estradiol, and because
`this steroid has been implicated in the regulation of VEGF (5,
`6, 11), we also examined the effects of ovariectomy and
`estradiol replacement on endometrial angiogenesis. In ad-
`dition, because VEGF Trap treatment suppresses plasma es-
`tradiol by virtue of its inhibitory effects on follicular devel-
`opment, effects of estradiol replacement were also compared
`in this group.
`
`Materials and Methods
`Animals and treatment
`
`Experiments were carried out in accordance with the Animals (Sci-
`entific Procedures) Act, 1986, and approved by the Local Ethical Review
`Process Committee. In common marmosets (Callithrix jacchus), the ovu-
`latory cycle comprises an 8-d follicular phase followed by a 20-d luteal
`phase, such that follicular recruitment ordinarily takes place during the
`mid to late luteal phase of the preceding cycle. To synchronize follicular
`recruitment, selection and ovulation during treatment cycles, marmosets
`(2– 4 yr old), housed as described previously (12), were injected with 1
`␮g prostaglandin F2␣ analog (cloprostenol, Planate; Coopers Animal
`
`4413
`
`Celltrion Exhibit - 1036
`Celltrion, Inc. v. Regeneron Pharmaceuticals, Inc.
`
`

`

`4414 Endocrinology, September 2008, 149(9):4413– 4420
`
`Fraser et al. (cid:127) Endometrial Angiogenesis
`
`Downloaded from https://academic.oup.com/endo/article/149/9/4413/2455652 by guest on 29 June 2022
`
`the ovary (4). For detection of proliferating endothelial cells, dual stain-
`ing was obtained by immunohistochemistry with CD31 and BrdU. For
`CD31 detection, the protocol was followed as described above, but
`visualization was performed with fast red (Sigma, Poole, UK). Sections
`were then washed with Tris-buffered saline before the second primary
`antibody, sheep antibody to BrdU (Fitzgerald, Concord, MA) was
`added, diluted 1:5000 in normal rabbit serum, and incubated overnight
`at 4 C. After postincubation washes with Tris-buffered saline, a biotin-
`ylated rabbit antisheep secondary antibody (Vector, Peterborough, UK)
`was added, followed by ABC-AP (Dako, Glostrup, Denmark). After
`incubation with the ABC-AP complex, slides were transferred to ni-
`troblue tetrazolium buffer before staining with nitroblue tetrazolium for
`15 min. The reaction was stopped in water, and the slides were air dried
`and then cleared in xylene and mounted in Pertex.
`
`Volume fraction measurements
`
`Five anatomic zones were evaluated: the luminal epithelium, glands
`of the functionalis, the stroma in the functionalis, the glands of the
`basalis, and the stroma in the basalis. Endothelial cells were identified
`using CD31 and proliferating endothelial cells by dual staining with
`CD31 and BrdU. Optimal delineation of the endothelial cells was ob-
`tained in the absence of counterstain. The glands of the functionalis and
`basalis were delineated by their location and shape (20). Volume fraction
`measurements were performed on sections dual stained with BrdU/
`CD31 as described previously (20), in three non-overlapping regions,
`perpendicular to the luminal epithelium extending to the myometrium:
`the first at the top of the uterus and the second and third on the right
`and left-hand side of the cross-section, respectively. In each region,
`successive non-overlapping fields of view were analyzed, totaling 20 –30
`fields per specimen. A test grid,- comprising 864 points was superim-
`posed on each field, and the number of test points falling on glands
`(including lumen), proliferating (BrdU-labeled) glandular epithelium,
`uterine lumen, stroma, proliferating stromal cells, luminal epithelium,
`proliferating luminal epithelial cells, endothelial cells, and proliferating
`endothelial cells (identified by colocalizing BrdU and CD31 immuno-
`staining) were counted. The volume fraction occupied by each compo-
`nent was then calculated by expressing the number of points hitting that
`component as a percentage of the total number of test points applied. In
`this way, for example, the volume fraction of proliferating (BrdU-la-
`beled) endothelial cells could be expressed as a percentage of the total
`endothelial cell (CD31-labeled) volume fraction.
`To assess the effects of the treatment on the area of the endometrial
`and myometrial compartments, and to determine stromal cell density,
`a stereological unit consisting of an Axio Imager A1 microscope (Carl
`Zeiss, Go¨ttingen, Germany) equipped with a video camera (KY-F550;
`JVC, Yokohama, Japan) and connected to a PC with a computer-assisted
`stereology system (Image-Pro Plus 6.2 Software; Media Cybernetics,
`Buckinghamshire, UK) was employed. The system was set up to outline
`and measure the area of the endometrium and myometrium in a cross-
`section of the uterus in which the lumen was maximal (i.e. approximately
`the center of the uterus). The endometrial and myometrial areas were
`outlined (see Fig. 3) and the area calculated using the computer-assisted
`stereology system. To assess changes in stromal density, the functionalis
`layer was examined on hematoxylin- and eosin-stained sections, and the
`slides were analyzed at a ⫻630 oil immersion objective. The microscope
`had a motorized stage controlled by the computer for selection of count-
`ing frames within an area of interest, in this case the stroma of the
`functionalis layer where 50 counting frames (45 ⫻ 45 ␮m) were ran-
`domly selected. Cell nuclei were counted within each frame, and the
`stromal density for each animal was taken as the mean of the 50 counts.
`
`Statistical analysis
`
`For volumetric analysis, results were log transformed and analyzed
`by ANOVA multiple comparison followed by a Bonferroni post hoc test.
`Differences were considered to be significant at P ⬍ 0.05.
`
`Results
`
`Hormonal changes
`Because all treatments were initiated during the secretory
`phase, plasma progesterone concentrations were elevated at
`
`Health Ltd., Crewe, UK) per animal im on d 13–15 of the luteal phase
`to induce luteolysis. The day of prostaglandin injection was designated
`follicular d 0. This method of synchronizing follicular recruitment is
`followed by follicle selection on cycle d 5 and ovulation on d 9 –11
`(13–15).
`To block VEGF, we employed the VEGF Trap (aflibercept) (Regen-
`eron Pharmaceuticals), a recombinant chimeric protein comprising por-
`tions of the extracellular domains of the human VEGF receptors 1 and
`2 expressed in sequence with the Fc portion of human Ig (2). VEGF Trap
`binds all isoforms of VEGF-A as well as placenta-derived growth factor.
`VEGF Trap was at a concentration of 24.3 mg/ml in buffer. In previous
`studies, we have shown that VEGF Trap (25 mg/kg) given sc every other
`day on d 0 –10 of the cycle or the same dose given once on d 0 are equally
`efficacious in suppressing follicular angiogenesis (3, 15). In the current
`study, we used an intermediate schedule of 25 mg/kg sc on d 0, 3, and
`7 beginning at the time of prostaglandin F2␣ analog administration.
`Animals were killed at 10 d, such that the treatment period covered the
`duration of the normal proliferative phase. Control animals were stud-
`ied at the late proliferative phase (d 10) of the normal cycle (n ⫽ 6 per
`group) after treatment with vehicle alone or a control protein, 25 mg/kg
`recombinant human Fc.
`To determine the effects of total withdrawal of ovarian steroids,
`marmosets were ovariectomized (n ⫽ 4) after sedation with ketamine
`hydrochloride (Parke-Davis Veterinary, Pontypool, Gwent, UK), anes-
`thesia by Saffan (Alphaxalone/Alpadalone; Schering-Plough Animal
`Health, Welwyn Garden City, UK) and maintenance of analgesia by
`buprenorphine (Alstoe Animal Health, Melton Mowbray, UK) im as
`described previously (16). Ovariectomy was carried out between d 13
`and 15 of the secretory phase, and uteri collected 10 d later, the same
`schedule as for the VEGF Trap treatment.
`To determine the effects of estrogen replacement, marmosets were
`treated with VEGF Trap (n ⫽ 6 per group) or ovariectomized as above
`(n ⫽ 4 per group), except that estradiol benzoate was administered in
`increasing doses to mimic the rising concentrations of plasma estradiol
`observed during the proliferative phase in intact marmosets (14) based
`on a modification of a method described previously (17). Estradiol
`benzoate was supplied in a solution of 5 mg/ml in arachis oil (Intervet
`UK Ltd., Milton Keynes, UK). This was diluted to obtain a solution of
`50 ␮g/ml in arachis oil and was administered sc once daily in doses
`increasing from 5 to 20 ␮g as follows: d 0 – 4, 100 ␮l; d 5, 150 ␮l; d 6 and
`7, 200 ␮l; d 8, 300 ␮l; and d 9, 400 ␮l.
`To obtain further information on changes in cell proliferation at other
`stages of the cycle, tissue was collected on d 10 –16 of the secretory phase
`(no treatment, represents d 0 in the main study) and at the mid prolif-
`erative phase (d 5) (n ⫽ 6 per group).
`Blood samples were collected three times per week for 6 – 8 wk pre-
`treatment and during the study period for progesterone assay as de-
`scribed previously. All animals were injected iv with 20 mg bromode-
`oxyuridine (BrdU; Roche Molecular Biochemicals, Essex, UK) in saline
`1 h before being sedated using 200 ␮l ketamine hydrochloride (Parke-
`Davis Veterinary) and 200 ␮l Saffan. After perfusion with 4% neutral
`buffered formalin, the uterus was removed immediately, weighed, and
`kept in 4% neutral buffered formalin for 24 h. The tissues were trans-
`ferred into 70% ethanol, dehydrated, and embedded in paraffin accord-
`ing to standard procedures. Serial sections were cut so that the endo-
`metrial area nearest the center of the uterus was selected for study.
`Plasma progesterone throughout the study and plasma estradiol on
`terminal blood samples were measured using assays described previ-
`ously (4, 18). Levels of unbound VEGF Trap were measured by ELISA
`as described previously (19) with plasma being diluted 1:10,000.
`
`Hematoxylin and eosin staining
`
`The uteri were embedded so that longitudinal sections could be cut.
`Uteri were serially sectioned (5 ␮m) at least to the middle of the tissue
`where the lumen was clearly visible and tissue sections placed onto BDH
`SuperFrost Plus slides (Menzel-Glazer, Menzel GmbH & Co., Braun-
`schweig, Germany). Representative sections were stained with hema-
`toxylin and eosin.
`
`Immunohistochemistry
`
`Methods for immunohistochemical detection of endothelial cells by
`CD31 and proliferating cells by BrdU were as described previously for
`
`

`

`Downloaded from https://academic.oup.com/endo/article/149/9/4413/2455652 by guest on 29 June 2022
`
`Fraser et al. (cid:127) Endometrial Angiogenesis
`
`Endocrinology, September 2008, 149(9):4413– 4420 4415
`
`FIG. 1. Functionalis area of endome-
`trium, dual stained for CD31 (red) and
`BrdU (dark-stained nuclei) of the endo-
`thelium (green arrows), stroma/perivas-
`cular cells (blue arrows) and epithelial
`cells within glands (G), (outlined to dis-
`tinguish them from stroma) during the
`mid-proliferative (MP) (A), late prolif-
`erative (LP) (B) and mid-late secretory
`(M-LS) (C) phases of the ovulatory cy-
`cle. Volume fraction of proliferating
`cells is shown for endothelial cells (D),
`stromal/perivasular cells (E), and ep-
`ithelial cells of glands (F). Values are
`means ⫾ SEM; n ⫽ 6 per group. Sig-
`nificant differences between stages
`are denoted by different letters. Scale
`bar, 50 ␮m.
`
`d 0. All marmosets responded to treatment with prostaglan-
`din analog or ovariectomy, with a rapid fall in plasma pro-
`gesterone levels to follicular-phase values, which were main-
`tained for the remainder of the treatment period (data not
`shown). Stage of cycle in control animals was also confirmed
`by ovarian histology as described previously (3). At the end
`of the experimental period, estradiol levels (means ⫾ sem)
`were 1965 ⫾ 466 pmol/liter in late proliferative controls and
`were significantly reduced (P ⬍ 0.01) by VEGF Trap (230 ⫾
`24 pmol/liter) and ovariectomy (89 ⫾ 28 pmol/liter). Estra-
`diol benzoate treatment restored plasma estradiol to the nor-
`mal range; 1964 ⫾ 354 pmol/liter in the ovariectomized
`group and 2009 ⫾ 279 pmol/liter in the VEGF Trap-treated
`group by the end of the treatment period. All VEGF Trap-
`treated marmosets exhibited plasma concentrations of free
`VEGF Trap (not already bound to VEGF) between 100 and
`200 mg/liter at 1 d after treatment, and levels remained
`above 10 mg/liter through d 10. The detection limit of the
`assay was 0.5 mg/liter, and it has been estimated that
`effective pharmacological blockade of VEGF in the ovary
`is maintained until free VEGF Trap levels fall to less than
`1–2 mg/liter (21). Thus, levels of unbound VEGF Trap in
`all treated marmosets remained within the anticipated
`pharmacologically effective range for the duration of the
`study.
`
`Volume fraction measurements
`BrdU-positive staining was observed in the nuclei of the
`epithelial cells of the glands, the CD31-positive endothelial
`cells within the stroma, and CD31-negative stromal cells
`located adjacent to blood vessels or dispersed within the
`stroma (Figs. 1, A–C, and 2A). For the purposes of analysis,
`the latter two groups were combined, and classified as being
`stromal.
`To obtain an overview of changes during the normal cycle,
`data from the mid-proliferative, late proliferative, and mid-
`late secretory phase were compared. Angiogenesis was
`rarely observed in mid-proliferative phase specimens (Fig.
`1A) but was common in the late-proliferative phase, espe-
`cially in the functionalis next to the luminal epithelium (Fig.
`1B). Angiogenesis continued to be observed in most speci-
`mens during the mid to late secretory phase (Fig. 1C). Quan-
`titative analysis confirmed that endothelial cell proliferation
`was lowest during the mid-proliferative phase and rose
`markedly by the late proliferative phase in both the func-
`tionalis (P ⬍ 0.001) and basalis zones (P ⬍ 0.05) (Fig. 1D).
`Endothelial cell proliferation in the functionalis zone re-
`mained significantly higher (P ⬍ 0.05) during the secretory
`phase than in mid-proliferative phase endometrium (Fig.
`1D). Stromal cell proliferation was also extremely low in the
`mid-proliferative phase and rose significantly in both func-
`
`

`

`Downloaded from https://academic.oup.com/endo/article/149/9/4413/2455652 by guest on 29 June 2022
`
`4416 Endocrinology, September 2008, 149(9):4413– 4420
`
`Fraser et al. (cid:127) Endometrial Angiogenesis
`
`FIG. 2. Functionalis area of endome-
`trium, dual stained for CD31 (red) and
`BrdU (dark-stained nuclei) in control
`late proliferative (LP) (A) phase, after
`VEGF Trap (Trap) treatment (B), after
`ovariectomy (ovex) (C), after ovariec-
`tomy with estrogen (Ovex ⫹ E) replace-
`ment (D), and after VEGF Trap with
`estrogen replacement (Trap ⫹ E) (E).
`Note proliferating endothelial
`cells
`(green arrows), proliferating perivascu-
`lar cell (blue arrow), and proliferating
`epithelial cells in glands (G). Histogram
`(F) shows mean volume fraction of pro-
`liferating endothelial cells as a percent-
`age of total endothelial cells within the
`functionalis and basalis layers. Values
`are means ⫾ SEM. Significant differ-
`ences between groups are denoted by
`different letters. Scale bar, 50 ␮m.
`
`tionalis (P ⬍ 0.001) and basalis (P ⬍ 0.05) layers by the late
`proliferative phase (Fig. 1E). In the glands of the functionalis
`in the mid-proliferative phase, the cell proliferation index
`was highly variable and not significantly different from the
`peak levels seen in the late proliferative phase (Fig. 1F). In the
`basalis glands, cell proliferation was significantly greater in
`the late proliferative compared with mid-proliferative phase
`(P ⬍ 0.01) (Fig. F). In the secretory phase, cell proliferation
`was significantly reduced in the glands of both the functio-
`nalis and basalis (P ⬍ 0.001) (Fig. 1, C and F).
`In the main study, abundant dual staining of proliferating
`endothelial cells was evident in late proliferative controls
`(Fig. 2A). In marked contrast, endothelial cell proliferation
`was virtually absent in the uteri of animals treated with
`VEGF Trap (Fig. 2B) or ovariectomized (Fig. 2C). Abundant
`endothelial cell proliferation was evident in the functionalis
`of ovariectomized marmosets that received estrogen (Fig.
`2D), but estrogen replacement appeared to produce only a
`small increase in endothelial cell proliferation in VEGF Trap-
`treated animals (Fig. 2E). Quantitative analysis revealed that
`when treatment groups were compared with late prolifera-
`tive phase controls, endothelial cell proliferation was signif-
`icantly lower in VEGF Trap-treated and ovariectomized an-
`imals, being virtually absent both in the functionalis (P ⬍
`0.001) and basalis zones (P ⬍ 0.05) (Fig. 2F). After estrogen
`
`replacement, angiogenesis was fully restored in the functio-
`nalis in the ovariectomized animals (Fig. 2F). In contrast,
`endothelial cell proliferation was reduced in estrogen-
`treated animals given VEGF Trap compared with controls
`(P ⬍ 0.05), although this value was higher than in animals
`given Trap alone (P ⬍ 0.01) (Fig. 2F). In the basalis zone,
`estrogen replacement produced small increases in endothe-
`lial proliferation in both ovariectomized and VEGF Trap-
`treated groups (Fig. 2F). Despite the near total inhibition of
`endothelial cell proliferation after VEGF Trap or ovariec-
`tomy, vascular density within the stroma was not decreased
`(see Fig. 2, B and C). The likely cause of this paradox was the
`coincident reduction in stromal volume produced by these
`treatments. The mean fraction volume of endothelial cells as
`a percentage of total stromal cells showed no significant
`difference among groups.
`
`Histology
`The gross appearance of the uterus was altered in VEGF
`Trap-treated and ovariectomized animals, being reduced in
`size relative to controls. Area measurements confirmed that
`both the endometrium (P ⬍ 0.01) and the myometrium (P ⬍
`0.05) were significantly smaller in the ovariectomized ani-
`mals than late proliferative controls (Fig. 3, B and C). Al-
`
`

`

`Fraser et al. (cid:127) Endometrial Angiogenesis
`
`Endocrinology, September 2008, 149(9):4413– 4420 4417
`
`Downloaded from https://academic.oup.com/endo/article/149/9/4413/2455652 by guest on 29 June 2022
`
`ciated with decreased cell size and increased cell density.
`Representative sections from the functionalis zone are shown
`in Fig. 4. In the uteri of late proliferative controls, stromal
`cells were large and contained abundant cytoplasm, and the
`glandular epithelium exhibited the round or oval appearance
`characteristic of estrogen stimulation (Fig. 4A). After VEGF
`Trap treatment or ovariectomy, the epithelial cells appeared
`more densely packed, and cytoplasmic volume appeared
`reduced relative to controls (Fig. 4, B andC). These changes
`were even more marked in the stroma of both treatment
`groups (Fig. 4, B and C). These observed decreases in cell size,
`and consequent increases in packing density, were reversed
`by estrogen replacement (Fig. 4, D and E).
`BrdU labeling showed that proliferation the epithelium of
`the functionalis glands was high in controls but was also
`taking place after ovariectomy or VEGF Trap treatment and
`also were unaffected by estrogen replacement (Fig. 4, F–J).
`Quantification revealed no significant differences among
`groups in the epithelial cell proliferation index in the func-
`tionalis (Fig. 4K). In the basalis glands, epithelial cell prolif-
`eration was significantly reduced in the VEGF Trap-treated
`group compared with the controls, and proliferation was
`restored by estrogen replacement.
`Stromal cell proliferation was significantly lower (P ⬍ 0.001)
`in VEGF Trap-treated and ovariectomized animals in both the
`functionalis and basalis zones and was restored by estrogen in
`both groups (Fig. 4, F–J, M, and N). Despite the clear inhibition
`of stromal cell proliferation, the number of stromal cells per unit
`area was increased in both VEGF Trap and ovariectomized
`animals relative to controls (P ⬍ 0.001). Treatment of both VEGF
`Trap and ovariectomized groups with estrogen resulted in sig-
`nificant reductions in stromal density (P ⬍ 0.01) compared with
`VEGF Trap and ovariectomy alone (Fig. 4O).
`With respect to the luminal epithelial cells, intense prolifer-
`ation was observed at the late proliferative phase and was
`significantly reduced (P ⬍ 0.05) in the ovariectomized group.
`Proliferation rates were not significantly different from the con-
`trol for any other treatment group (data not shown), and there
`were no significant changes in the volume fraction of epithelial
`cells among the different groups (data not shown).
`
`Discussion
`This study provides the first description of the effects of
`VEGF inhibition on the primate uterus and shows that treat-
`ment of marmosets with a potent inhibitor of VEGF over the
`course of the proliferative phase results in a near complete
`inhibition of endothelial cell proliferation. In addition, stromal
`cell proliferation was also markedly suppressed, whereas the
`stroma itself became condensed. Thus, inhibition of VEGF in
`the intact marmoset results in an inhibition of endometrial
`angiogenesis. However, pharmacological inhibition of VEGF
`in intact animals also inhibits ovarian follicular angiogenesis,
`leading to reductions in plasma concentrations of estradiol
`(3, 19). Estrogen is known to stimulate the synthesis of VEGF
`in the endometrium and may similarly influence the expres-
`sion of numerous other factors associated with angiogenesis
`and maintenance of vascular function in primates (6) and
`rodents (22). Indeed, ovariectomy also inhibited angiogen-
`esis and produced other morphological changes in the uterus
`
`FIG. 3. A, Marmoset uterus showing a typical outline used to calcu-
`late myometrial (M) and endometrial (E) areas. The areas of both the
`endometrium (B) and myometrium (C) were significantly reduced by
`ovariectomy and fully restored by estrogen replacement. Significant
`differences between groups are denoted by different letters.
`
`though both endometrial and myometrial areas were also
`reduced in VEGF Trap-treated animals, these intermediate
`values were not statistically different from those of ovariec-
`tomized animals or controls. Estrogen replacement restored
`both endometrial and myometrial areas to normal in both
`groups (Fig. 3, B and C).
`Further analysis of hematoxylin- and eosin-stained sec-
`tions revealed that reductions in uterine volume evident in
`ovariectomized and VEGF Trap-treated animals were asso-
`
`

`

`4418 Endocrinology, September 2008, 149(9):4413– 4420
`
`Fraser et al. (cid:127) Endometrial Angiogenesis
`
`Downloaded from https://academic.oup.com/endo/article/149/9/4413/2455652 by guest on 29 June 2022
`
`FIG. 4. Hematoxylin- and eosin-stained
`sections of functionalis of control late
`proliferative phase (LP)
`(A), after
`VEGF Trap (Trap) treatment (B), after
`ovariectomy (Ovex) (C), after ovariec-
`tomy with estrogen replacement (Ovex
`⫹ E) (D), after VEGF Trap with estro-
`gen replacement (Trap ⫹ E), and in cor-
`responding BrdU-stained sections (F–
`J). Note that VEGF Trap treatment and
`ovariectomy reduce epithelial cell size
`in the glands and increase cell density of
`the stroma; effects reversed by estrogen
`replacement. In all groups, epithelial
`cell proliferation (dark-stained nuclei)
`is present in glands. Note that stromal
`proliferation in the control
`(dark-
`stained nuclei, yellow arrows) is lost af-
`ter VEGF Trap treatment or ovariec-
`tomy but
`is restored by estrogen
`replacement. Quantification of cell pro-
`liferation (K–O) shows that VEGF Trap
`decreased glandular proliferation in the
`basalis glands, and this was reversed by
`estrogen replacement. Stromal prolifer-
`ation was suppressed in both the func-
`tionalis and basalis glands after VEGF
`Trap or ovariectomy, and this effect was
`reversed by estrogen replacement.
`Quantitative analysis also confirmed
`that cell density in the stroma (O) was
`increased by VEGF Trap treatment or
`ovariectomy, and this effect was re-
`versed by estrogen replacement. Values
`are means ⫾ SEM. Significant differ-
`ences between groups are denoted by
`different letters. Scale bar, 50 ␮m.
`
`that resembled those seen after VEGF inhibition. Therefore,
`we also assessed the effects of estrogen replacement on these
`parameters in animals that were treated with VEGF Trap
`or ovariectomized. All morphological and proliferative
`changes in the uterus that resulted from ovariectomy were
`substantially normalized by estrogen replacement. Estro-
`gen replacement also fully restored stromal proliferation
`and stromal cell density largely to control levels. However,
`in animals treated with VEGF Trap, estrogen replacement
`had only a minor stimulatory effect on endothelial cell
`proliferation, demonstrating that VEGF plays a major and
`indispensable role in the final pathway that regulates an-
`giogenesis in the primate endometrium.
`Evidence from earlier studies using a variety of experi-
`mental models have suggested that exposure to estradiol
`during the proliferative phase drives VEGF production and
`
`angiogenesis in the uterus (6, 23, 24). Our quantitative studies
`in the marmoset confirm that there is indeed a significant
`increase in endothelial cell proliferation between the mid-
`proliferative (d 5) and late proliferative phases (d 10) and that
`this increase in angiogenesis is dependent upon estrogen and
`mediated by VEGF. However, there is little direct evidence
`indicating that there is a cyclic peak in uterine angiogenesis
`during the late proliferative phase in humans; rather, the
`endometrium in women has been reported to exhibit a rel-
`atively stable rate of endothelial cell proliferation (25). How-
`ever, it has been acknowledged that this negative finding
`might be attributable to technical difficulties inherent in hu-
`man studies such as individual variability in hormonal cycles
`or variability inherent in the use of small biopsies (26 –29). To
`establish a more precise relationship between steroid expo-
`sure and changes in endometrial angiogenesis, Nayak and
`
`

`

`Fraser et al. (cid:127) Endometrial Angiogenesis
`
`Endocrinology, September 2008, 149(9):4413– 4420 4419
`
`Downloaded from https://academic.oup.com/endo/article/149/9/4413/2455652 by guest on 29 June 2022
`
`sels supply a substantially reduced volume of tissue and
`hence appear more densely packed, as in the case for the
`stromal cells themselves (7). Thus, although the vascular
`density in the uterus was not reduced, the total endothelial
`mass is likely to be decreased.
`Apart from inhibiting endothelial cell proliferation, it is
`likely that the pharmacological blockade of local VEGF re-
`sults in additional changes in the structure and function of
`the uterine vasculature. Taking effects on the marmoset
`ovary as an example, VEGF inhibition induces an increase in
`the expression of the tight junctional protein claudin 5 (34),
`an up-regulation of hypoxia-inducible factor-1␣ (35), a de-
`crease in expression of mRNA for VEGF receptors (4), and an
`acute increase in endothelial cell death in the recently formed
`blood vessels in the corpus luteum (21). VEGF inhibition has
`also been shown to suppress the estrogen-induced increase
`in uterine permeability in rodents (7–11). These findings,
`together with the fact that VEGF inhibition has been shown
`to reduce vascular permeability in diverse vascular beds
`across species, strongly suggest that VEGF also plays a piv-
`otal role in modulating vascular function and permeability
`in the primate uterus, including regulation of tight junctions
`(24). Indeed, decreases in the permeability of uterine vessels
`may have contributed to the overall decrease in uterine vol-
`ume and increase in stromal cell density observed in ovari-
`ectomized and VEGF Trap-treated marmosets.
`Glandular epithelial cell proliferation in the basalis zone
`and luminal epithelial cell proliferation were partially in-
`hibited by ovariectomy or treatment with VEGF Trap com-
`pared with the late proliferative phase controls. This is to be
`expected because estrogen is the established stimulator of
`this process (29). However, the apparent absence of effects of
`VEGF Trap treatment or ovariectomy