| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Other Original Articles |
Division of Reproductive Sciences, Oregon Regional Primate Research Center, Beaverton, Oregon 97006
Address all correspondence and requests for reprints to: Nihar R. Nayak, Ph.D., Department of Gynecology and Obstetrics, Stanford University School of Medicine, 300 Pasteur Drive, HH-333, Stanford, California 94307-5317. E-mail: . nayakn{at}stanford.edu
Abstract
The relationship between vascular endothelial growth factor (VEGF) expression and the pattern of vascular proliferation in the rhesus macaque endometrium has not been studied. In this report, we used in situ hybridization to evaluate VEGF, VEGF receptor type 1 and VEGF receptor type 2 mRNA expression during hormonally regulated menstrual cycles in ovariectomized macaques. Proliferating endothelial cells were identified by a double immunocytochemistry procedure that detected Ki-67 antigen and von Willebrand factor in the same endothelial cells. One and 2 d after progesterone withdrawal (premenstrual), VEGF mRNA was up-regulated in the glands and stroma of the superficial endometrial zones, a finding that supports our previous suggestion that VEGF may play a role in the menstrual induction cascade. During the postmenstrual repair phase, the healing surface epithelium showed a further, dramatic increase in expression of VEGF mRNA, accompanied by strong increases in signals for VEGF receptor types 1 and 2 in multiple profiles of small blood vessels immediately below the surface epithelium. This finding implicates VEGF in the early angiogenic processes associated with endometrial healing and regeneration. Vascular endothelial proliferation persisted throughout the cycle in the upper endometrial zones and showed a dramatic estrogen- dependent peak during the midproliferative phase. This proliferative peak coincided with a peak in VEGF expression in the endometrial stroma. Endothelial proliferation was also significantly correlated with the degree of stromal VEGF expression during the proliferative and secretory stages of the cycle. These results implicate VEGF of stromal origin in endometrial vascular proliferation.
IN ADULT FEMALE primates, the endometrium undergoes shedding of the upper zones and cyclical repair and regeneration during the normal menstrual cycle. A key feature of this remarkable tissue remodeling is the growth of the vasculature (1, 2, 3, 4). Although the sole ovarian steroid hormones required to induce these changes are E2 and progesterone (P) (5), several local factors are presumed to play significant roles in mediating these important events. Despite the importance of understanding the mechanisms of endometrial bleeding and repair and the varied pathological implications of abnormal bleeding, the regulatory mechanisms and local factors involved in menstruation, healing, and regeneration of the primate endometrium are not fully understood.
Of the various angiogenic factors described so far, vascular endothelial growth factor (VEGF) is a prime regulator of both physiological and pathological angiogenesis. Targeted disruption of even a single VEGF allele resulted in abnormal blood vessel formation and embryonic death in mice (6, 7). Also, treatment with neutralizing VEGF antibodies (8, 9) or a soluble truncated form of VEGF receptor type 1 (10) in primates has been shown to inhibit follicular development and also suppress luteal function by inhibiting angiogenesis in the corpus luteum. VEGF is expressed in a wide variety of cells and tissues, including rodent and primate endometrium (3, 11, 12). An interesting feature of VEGF structure is that multiple species of VEGF mRNA are generated by alternative splicing from a single VEGF gene containing eight exons, separated by seven introns. These mRNAs result in the generation of four different molecular species, having, respectively, 121,165, 189, and 206 amino acids following signal sequence cleavage (11). A fifth splice variant having 145 amino acids has been reported in the endometrium (13). Human endometrium predominantly expresses 121 and 165 isoforms. Although VEGF mRNA expression in the endometrium changes during different stages of the cycle, there is no qualitative shift in the expression of different splice variants throughout the cycle (13, 14, 15). VEGF activity is mainly mediated by two high-affinity tyrosine kinase receptors, VEGF receptor type 1 (fms-like tyrosine kinase receptor; Flt-1) and VEGF receptor type 2 (kinase insert domain- containing receptor; KDR/Flk-1) (11, 12).
Reports on ovarian steroid hormone regulation of VEGF expression in the endometrium are inconsistent, mainly because of the use of immunocytochemistry (ICC) possibly giving artifacts (16). Several ICC and in situ hybridization (ISH) studies reported an increase in glandular VEGF expression during the secretory (13, 15, 16, 17) and menstrual phases (13, 16), but other ICC studies found no difference in glandular VEGF expression across the cycle (18, 19). Stromal VEGF expression is generally reported to be low across the menstrual cycle (15, 16, 17, 18), whereas Li et al. (19) have demonstrated strong stromal expression during the proliferative stage. Further, using ICC, Greb et al. (20) have reported that VEGF expression is highest in the glandular epithelium of leuprolide acetate (GnRH agonist) treated, hypoestrogenic cynomolgus macaques and that P treatment of such animals induced intense VEGF expression in the stroma. In contrast to these in vivo results, treatment of isolated human endometrial stromal cells in cell culture with E2, progestins, or both has been shown to consistently increase VEGF mRNA and protein expression (14, 17). These studies depict a confusing picture of how hormones regulate VEGF in the endometrium.
Several in vivo and in vitro studies indicate that VEGF is a major regulator of endothelial cell proliferation (11). However, Gargett et al. (21) found no correlation between VEGF production and endothelial cell proliferation in the human endometrium. Also, there is considerable variability among reports on the pattern of human endometrial vascular proliferation, with some studies indicating several peaks of proliferation (22, 23) and others indicating none (24). It has been suggested that these differences among reports on VEGF expression and vascular proliferation may be owing to variations in hormone levels at the time of endometrial sampling or to variations in the region biopsied (25). The major limitation of many of the previous studies on VEGF expression and vascular proliferation in the human endometrium is that the biopsies were not removed at closely spaced intervals from the time of menstruation through the early proliferative phase, when the endometrium undergoes extensive reparative and regenerative processes.
To address these concerns, we used ovariectomized artificially cycling macaques and withdrew P at the end of cycle to provide a starting point for synchronization of endometrial samples. We obtained full-thickness endometrial specimens in which we examined mRNA expression of VEGF and its receptors and correlated these data with endothelial cell proliferation on a zonal basis. We have previously reported that KDR is dramatically up-regulated in the stromal cells of the superficial zones only during the premenstrual stage regardless of whether E2 is maintained (26). In this study, we evaluated expression of both KDR and Flt-1 mRNA in the endometrial vasculature throughout the cycle including postmenstrual repair and the remainder of the artificial proliferative phase, with and without E2 treatment, and during the artificial secretory phase.
Materials and Methods
Experimental animals
All animal care during these studies was provided by the Division of Animal Resources of the Oregon Regional Primate Research Center, in accordance with the NIH guidelines for use of nonhuman primates and as approved by the Primate Center Institutional Animal Care and Use Committee. Adult female rhesus macaques (Macaca mulatta) were ovariectomized and treated sequentially with E2 and P to create artificial menstrual cycles as described previously. Briefly, all macaques received sc implants of 3-cm Silastic capsules packed with crystalline E2 (Sigma, St. Louis, MO) to stimulate development of an artificial proliferative phase endometrium. After 14 d, a 6-cm Silastic capsule packed with crystalline P (Sigma) was implanted sc, and both implants remained in place for 14 d to stimulate an artificial secretory phase endometrium. Then the P implant was removed to induce menstruation. In one set of P-withdrawn animals (n = 19) the E2 implant was left in place (to mimic the natural cycle), and in another set (n = 16), the E2 implant was also withdrawn (to assess the role of E2). Animals from whom both P and E2 were removed are referred to further in the text as hormone-deprived (HD) animals.
A third set of animals (n = 6) had their P implants removed and E2 maintained, but their uteri were not removed. After 14 d of E2, a P implant was reintroduced to induce a secretory phase, and uteri were removed after 7 and 14 d of E2+ P treatment. These samples represented the mid- and late secretory phases of the cycle.
The various samples are referred to in the text and graphs as follows: 1) premenstrual phase: 1–2 d of P withdrawal (n = 4); 2) menstrual phase: 3–4 d of P withdrawal (n = 4); 3) early proliferative phase/postmenstrual repair phase: 5–6 d of P withdrawal (n = 4); 4) midproliferative phase: 8–10 d of P withdrawal (n = 4); 6) late proliferative phase: 14 d of P withdrawal (n = 3); 7) midsecretory phase: 7–8 d of E2+ P treatment (n = 3); 8) late secretory phase: 14 d of E2+P treatment (n = 3); 9) HD 1–2 d: 1–2 d after both E2 and P withdrawal (n = 3); 10) HD 3–4 d: 3–4 d of both E2 and P withdrawal (n = 3); 11) HD 5–6 d: 5–6 d of both E2 and P withdrawal (n = 4); 12) HD 8–10 d: 8–10 d of both E2 and P withdrawal (n = 4); and 13) HD 14 d:14 d of both E2 and P withdrawal (n = 2).
Endometrial tissue samples were collected by hysterectomy as described previously (26, 27). Briefly, the uterus was quartered along the longitudinal axis and full-thickness uterine cross-sections (2 mm thick) were prepared from each quarter for ICC and ISH. Tissues for ICC were microwave stabilized for 7 sec in 0.5 ml HBSS (Life Technologies, Inc., Grand Island, NY), then chilled on ice in 10% sucrose dissolved in 0.1 M PBS, mounted in Tissue Tek II OCT (Miles Inc., Elkhart, IN) and frozen in liquid propane. The samples for ISH were frozen without microwave treatment. Some of the endometrial tissue samples collected for previous studies (26, 27) were also used in this study to increase the sample size. In each case, serum was harvested at the time of tissue collection, and concentrations of serum E2 and P were determined by RIA as previously validated (28). The serum levels of E2 and P were within the normal physiological range for rhesus monkeys and identical to those previously reported by our laboratory (26).
ICC
Proliferating endothelial cells were detected by a double ICC procedure with a mouse monoclonal antibody to Ki-67 for proliferating cells and a rabbit polyclonal antibody to von Willebrand factor (vWF) to detect vascular endothelium. ICC was performed as described previously (26, 27, 29) with modifications. Briefly, fresh tissues were microwaved for 7 sec before being embedded in OCT, frozen in liquid propane, and cryosectioned at 7 µm. Cryosections were mounted on Super Frost Plus slides (Fisher Scientific, Pittsburgh, PA), fixed in 0.2% picric acid-2% paraformaldehyde in phosphate buffer saline at pH 7.3 for 10 min at room temperature, immersed twice for 2 min each in 85% ethanol + 1.5% polyvinylpyrrolidone (PVP) at 4C, rinsed in PBS, immersed twice 7 min each in 0.37% glycine in PBS + PVP, and then immersed in 0.1% gelatin in PBS + PVP at 4 C. To inhibit endogenous peroxidase activity, the sections were incubated with a solution containing glucose oxidase (1 U/ml), NaAzide (1 mM), and glucose (10 mM) in PBS for 45 min. Sections were then incubated with blocking serum for 20 min and then with the mouse monoclonal primary antibody for Ki-67 (1:300, BioGenex Laboratories, Inc. San Ramon, CA) overnight at 4 C. After rinsing and immersion in blocking serum again, sections were incubated with a biotinylated second antibody (antimouse) for 30 min at room temperature. Brown staining of nuclei that were immunopositive for Ki-67 was achieved with the ABC kit (Vector Laboratories, Inc. Burlingame, CA), which included 0.025% 3,3' diaminobenzidine/4HCl (Dojindos DAB; Wako Chemicals, Richmond, VA) in Tris buffer and 0.03% H2O2 (Fisher Scientific), as described previously (26, 27). All slides were then washed several times with 0.1% gelatin in PBS, incubated with blocking serum for 20 min, and then with the rabbit polyclonal primary antibody for vWF (1:4000, DAKO Corp., Carpinteria, CA) overnight at 4 C. Sections were then rinsed and reincubated with blocking serum and then with the biotinylated second antibody (antirabbit) for 30 min at room temperature.
Blue-gray cytoplasmic staining of endothelial cells (vWF) was achieved with a Vector SG substrate kit for peroxidase (Vector Laboratories, Inc.) by following the manufacturer’s instructions. The slides were rinsed several times in deionized water and lightly counterstained with hematoxylin to facilitate identification of cell types.
To validate that the Ki-67 counts were representative of the endothelial cells undergoing DNA synthesis, we also administered an iv infusion of Br-dU (10 ml, 10 mg/ml) at three time points, starting 24, 16, and 2 h before tissue collection, to two animals each during the early proliferative stage, midproliferative stage, HD 5–6 d, and HD 8–10 d, and one animal during the late proliferative stage. Labeled endothelial cells were detected with a mouse monoclonal antibody to Br-dU (1:50, ICN Biomedicals, Inc., Costa Mesa, CA) and the same rabbit polyclonal antibody to vWF. Tissue processing and ICC were performed in exactly the same manner as described for Ki-67 immunostaining except slides used for Br-dU ICC were treated with 2N HCl at 25 C for 30 min as required to detect DNA labeled with Br-dU. These preparations were compared with the Ki-67 immunostaining to validate that the Ki-67 counts represented endothelial cells undergoing DNA synthesis. Total number of endothelial cells and the number of Ki-67 or Br-dU-positive endothelial cells (proliferating) were counted in the microvessels of the upper and lower zones of each endometrial tissue section, and the percentage of proliferating endothelial cells was calculated.
RT-PCR
We used RT-PCR for preparation of VEGF cDNA specific to rhesus macaques by following the same procedure as described previously for KDR and Flt-1 (26). Briefly, 5 µg total RNA prepared from rhesus macaque endometrium was reverse transcribed with an oligo(dT) primer and SuperScript II MMLV reverse transcriptase (Life Science Technologies, Rockville, MD). The reverse transcriptase product was then amplified with the 5' and 3' primers in a standard PCR reaction for 35 cycles at 92 C for 30 sec, 50 C for 30 sec, and 72 C for 1 min. Amplified bands of the right size were gel isolated and subcloned into pGEM-T (Promega Corp., Madison, WI). At least two clones with the right-sized inserts were miniprepped (Perfect Preps; Eppendorf-5 Prime, Inc., Boulder, CO) and were sequenced on an ABI 373 XL sequencer. Primers for the VEGF cDNA were selected on the basis of the homologous human VEGF (37–416 bp, accession # x62568) sequences, which spans exons 1–5 and therefore would detect all known alternatively spliced variants of VEGF gene. The forward and reverse primers used to amplify VEGF cDNA were GGTGCATTGGAGCCTTGCCTTGCT and TCTTTGGTCTGCATTCACATTTGT, respectively. The partial cDNA sequence for rhesus macaque VEGF has been submitted to GenBank with accession number AF 339737.
ISH
ISH of frozen sections was conducted with 35S-UTP-labeled (NEN Life Science Products, Boston, MA) sense and antisense riboprobes from VEGF cDNA as described previously for KDR and Flt-1 (26). Briefly, 10 µm frozen sections of endometrium mounted on Super Frost Plus slides (Fisher Scientific) were fixed in 4% paraformaldehyde in PBS for 15 min at 4 C. The tissue sections were rinsed in 2x SSC, acetylated with 0.25% acetic anhydride in 0.1 M triethanolamine (pH 8.0) for 10 min, rinsed in 2x SSC, dehydrated through an ascending series of alcohols and air dried. At this point at least one slide per tissue group was treated with RNAase A (20 mg/ml, 0.5 M NaCl, 0.01 M Tris, 1 mM EDTA; pH 8.0) as a negative control. Sections were then incubated at 55 C overnight in 10 mM DTT, 0.3 M NaCl, 20 mM Tris (pH 8.0), 5 mM EDTA, 1x Denhardt’s solution, 10% dextran sulfate, and 50% formamide containing the appropriate concentration of the sense and antisense probe (5 million cpm/ml). After hybridization all slides were treated with RNAase A at 37 C for 30 min to inactivate nonhybridized probe, and the slides were rinsed in a descending series of SSC (2x SSC, 1x SSC, 0.5x SSC) and then incubated in 0.1x SSC at 65C (high stringency) for 30 min. Sections were dehydrated in an ascending series of alcohol dilutions, vacuum dried, coated with NTB2 autoradiographic emulsion (Eastman Kodak Co., Rochester, NY), stored at 4 C for 10 d, developed in D-19 (Eastman Kodak Co.), lightly counterstained with hematoxylin, dehydrated in an ascending series of alcohol dilutions, cleared with xylene, and coverslipped with Permount (Fisher Scientific). Sense- and RNAase-treated controls had no specific signals.
Silver grains over endothelial, stromal, surface, and glandular epithelial cells in different zones of endometrium were counted separately as described previously with modifications (30). The counts were made with MetaMorph (Universal Imaging Corp., Downingtown, PA) on images captured by a CoolSNAP color CCD digital camera (Roper Scientific, Inc., Tucson, AZ). The abundance of silver grains over the stroma, surface, and glandular epithelium was expressed as the number of grains per cell. The abundance of silver grains over endothelial cells was expressed as the number of grains per unit area of vascular endothelium.
Statistical analysis
All data were tested by one-way ANOVA, and significance between groups was assessed with Fisher’s protected least significant difference test (31). Results with a P value of less than 0.05 were considered significant. Correlations were performed using the StatView software (SAS Institute, Inc., Cary, NC), and coefficients of simple determination (R2) were calculated.
Results
VEGF expression during the cycle
Figures 1–3
show the relative abundance of VEGF mRNA expression as determined by silver grain counts over the endometrial luminal (surface) epithelium, glands, and stroma. These data show a peak in VEGF expression in the surface epithelium during the early proliferative phase (Fig. 1), in the stroma during the midproliferative phase (Fig. 2), and in the glands during the late secretory phase (Fig. 3). Regardless of hormone treatment, both RNAase- and sense-treated control sections showed signal equivalent to background on glass slides away from the section (data not shown). There was a gradient in VEGF expression from highest at the surface to minimal in the lower zones (Fig. 4, A and B), and marked changes were evident only in the superficial zone glands and stroma (Figs. 2 and 3). The lower zones (basalis) showed no significant differences in VEGF expression at any of the time points that were sampled (data not shown). No VEGF signal was evident over the vascular endothelium (Fig. 4C).
|
|
|
|
and 4). After menstrual sloughing of the upper zones, the newly formed surface epithelium showed a dramatic increase in VEGF mRNA expression during the early proliferative stage (Fig. 1, 5A). Of great interest, in the HD groups, exactly the same pattern of up-regulation and localization of VEGF mRNA in the glands, stroma, and luminal epithelium (Figs. 1–3, 5B) was found when both hormones were withdrawn through HD 5–6 d as observed from the late secretory phase through early proliferative phase (5–6 d after only P was withdrawn). Subsequently, VEGF expression was increased in the stroma during the proliferative stage with highest expression after 8–10 d of P withdrawal (midproliferative stage) (Figs. 2, 5C). However, in the HD 8–10 d group, stromal VEGF mRNA expression was significantly down-regulated (Figs. 2 and 5D). Overall, VEGF mRNA levels were also very low in glands and surface epithelium after 8–10 d of HD (Figs. 1, 3, and 5D). These data indicate that although E2 is not essential to the up-regulation of VEGF mRNA that occurs immediately after P withdrawal during the very early proliferative phase, E2 is necessary for the increase in VEGF mRNA that occurs later during the midproliferative phase.
|
Throughout the cycle KDR and Flt-1 expression was confined to the endometrial vascular endothelium. However, during the premenstrual stage and during HD 1–2, expression of KDR mRNA was very strongly up-regulated in the endometrial stromal cells of the upper zone as previously reported (26). No significant changes in KDR or Flt-1 mRNA expression were observed in the lower zone endometrial vessels throughout the induced menstrual cycle (data not shown). Figures 6 and 7 depict the relative abundance of KDR and Flt-1 mRNA expression only in the uppermost endometrial zone vasculature under different hormonal conditions. There was a dramatic increase in Flt-1 and KDR mRNA expression in multiple profiles of small blood vessels just below the newly formed surface epithelium during the early proliferative phase (after 5–6 d of P withdrawal) as well as in the HD 5–6 d group (Fig. 8). Grain counts performed on the early proliferative phase samples showed that the signals in the upper zone vessels were significantly higher than those in the lower zones (KDR: upper, 78.25 ± 10.61, lower, 47.25 ± 6.83; Flt-1: upper, 70.50 ± 10.37, lower, 27.50 ± 4.01; P < 0.05). By the midproliferative (d 8–10) and late proliferative phases (d 14), the signals in the upper zone had somewhat diminished (Figs. 6 and 7). However, in the HD 8–10 and HD 14 d groups, the signals had declined severely and were significantly different from all other days (Figs. 6 and 7). These data indicate that E2 is not essential for the increase in expression of Flt-1 and KDR that occurs in the superficial zones during the early healing and proliferative phases of the cycle. However, E2 plays a role in sustaining a baseline level of expression of Flt-1 and KDR in vessels from d 8 through the remainder of the cycle.
|
|
|
Figure 9 presents representative photomicrographs showing detection of proliferating endothelial cells through ICC colocalization of Ki-67 or Br-dU with vWF during the midproliferative stage in presence of E2. Ki-67- and Br-dU- positive endothelial cells were significantly correlated (R2 = 0.92, P < 0.001), and both were found primarily in the superficial zones of endometrium (Fig. 9, A and B). Endothelial cells in the lower zone vessels during the midproliferative stage (Fig. 9, C and D), and also at all other times sampled, were mostly negative for both Ki-67 and Br-dU.
|
. Compared with all other stages of the cycle, there was a dramatic increase (about 6-fold) in endothelial cell proliferation (Ki-67 labeling) in the midproliferative stage (8–10 d of P withdrawal). This dramatic increase of proliferation did not occur in the HD 8–10 d endometrial samples and was therefore E2 dependent. In absence of E2, overall endothelial cell proliferation was significantly decreased and was nondetectable by HD 14 d.
|
) coincided with the E2- dependent peak in stromal VEGF expression (Fig. 2). Moreover, throughout the proliferative and secretory stages of the cycle, the percentage of endothelial cell proliferation was highly correlated with the degree of VEGF mRNA expression in the superficial zone stroma (R2 = 0.734, P < 0.001). However, no significant correlation of vascular proliferation was found either with VEGF expression in glands (R2 = 0.002, P < 0.879) and surface epithelium (R2 = 0.012, P < 0.669) or with KDR (R2 = 0.002, P < 0.870) and Flt-1 (R2 = 0.006, P < 0.775) expression in the endothelium. Discussion
This is the first paper to report that the elevations in stromal VEGF that occur immediately after P withdrawal in the nonhuman primate are independent of E2 action because these increases occur similarly in the presence and absence of E2. Further, the elevations that occur during the period of postmenstrual repair in luminal VEGF together with KDR and FLT-1 in subjacent blood vessels are also E2 independent. These events are most likely regulated by other factors, to be discussed below. However, E2 is essential for vascular proliferation in the proliferative phase because the midproliferative peak of vascular proliferation does not occur in the absence of E2, and the baseline level of vascular proliferation throughout the remainder of the proliferative phase is greatly suppressed in the absence of E2. In addition, the baseline level of KDR and FLT-1 expression in the vascular endothelium depends on E2 because the expression of these receptors is significantly reduced during d 8–14 in the HD groups that lack E2. These observations are discussed more fully below.
Cellular localization of VEGF mRNA
Our ISH results show that VEGF mRNA expression was confined to the luminal epithelium, glands, and stroma, with no expression in the vascular endothelium. These findings are very similar to the ICC and ISH results reported by some others in the human (13, 15, 17) and cynomolgus macaque (20) endometrium. However, several other studies have demonstrated localization of VEGF protein in the endometrial vascular endothelium (14, 16, 19). This immunoreactivity does not necessarily correspond to VEGF production in endothelial cells because the antibodies used in the later studies could also detect VEGF bound to its receptors in endothelial cells (14).
Hormonal regulation of VEGF and its receptors
Recently, it has been demonstrated that E2 directly regulates VEGF gene transcription in endometrial cells through a variant estrogen response element located 1.5 kb upstream from the transcriptional start site (32). Several studies also indicate that estrogen regulates VEGF mRNA expression in human endometrial adenocarcinoma cells (13), stromal cells (14, 17), and the rat uterus in vivo (33, 34). These are likely to be primary estrogen receptor-mediated effects because the induction peaks within 2 h (33, 34) and is blocked by pure antiestrogens (35). This regulation by E2 is consistent with our findings of E2-dependent increase in VEGF expression during the midproliferative stage of the cycle and complete suppression of VEGF expression in absence of E2 in equivalent HD samples. However, after P withdrawal, during the premenstrual, menstrual, and postmenstrual repair stages of the cycle, the increases in VEGF expression were identical, whether E2 was present, indicating that E2 has no regulatory role during this period. Furthermore, as the proliferative phase progressed, there was significant reduction in stromal VEGF mRNA expression from the midproliferative (d 8) to the later proliferative stage (d 14) of the cycle, even though E2 levels were held constant throughout this period. These results suggest that other factors participate in regulation of VEGF expression in vivo in the rhesus macaque endometrium.
Although a P response element has not yet been identified in the VEGF gene, several reports indicate that progestins alone or in combination with estrogens can stimulate VEGF expression in human endometrial stromal cells in vitro (14, 17) and in the rodent uterus (33). However, our results are consistent with the in vivo studies in human endometrium (13) and show that after 14 d of P treatment, VEGF expression was significantly down-regulated in the stroma but up-regulated in the glands. Our data strongly suggest that these effects represent an indirect effect of P on VEGF expression in vivo because they were not evident in the midsecretory stage (7–8 d after P treatment) but only in the late secretory phase (14 d of P treatment). The long delay in this effect of P suggests an indirect action of P and implicates other factors in the increased glandular VEGF expression we observed.
Very little information is available concerning the hormonal regulation of VEGF receptors. A recent ICC study indicates variations of staining intensity and number of stained capillaries immunostained for Flt-1 and KDR during different stages of the menstrual cycle in women (36). However, except for the postmenstrual repair phase, we did not find any dramatic changes in Flt-1 or KDR mRNA expression throughout most of the menstrual cycle in the rhesus macaque. The postmenstrual increase in Flt-1 and KDR mRNA expression in the vessels immediately below the surface epithelium was also not regulated by E2 because a similar pattern of expression was observed in the HD macaques sampled at the same time. However, both Flt-1 and KDR expression was significantly down-regulated after 8 d of HD, suggesting a role for E2 in maintenance of VEGF receptors in the vascular endothelium. Given the fact that VEGF can regulate its own receptor expression (37, 38, 39) and that stromal VEGF expression was significantly decreased in the HD groups from d 8–14 in the absence of E2 (Fig. 2), it is possible that E2 maintains Flt-1 and KDR in the vascular endothelium through stimulation of VEGF production. Alternatively, ERß, which has recently been found in the vascular endothelium of both the rhesus macaque (40) and human endometrium (40, 41), may mediate the effects of E2 on these receptors more directly.
Role of VEGF during menstruation
Previously we have reported that KDR expression is dramatically up-regulated in the stromal cells of the superficial endometrial zones during the premenstrual phase in both human and macaque endometrium, and we suggested that this receptor played a role in the menstrual induction cascade (26). Here we report that during this same phase, VEGF expression is also maintained at very high levels in the glands and is dramatically increased in the stroma of the same upper zones of the rhesus macaque endometrium. Our findings differ from several reports (13, 16, 42) that state that VEGF expression is increased only in glands during the menstrual phase in the human endometrium. However, the endometrium collected by Pipelle biopsy during menstruation from women may consist of only the deeper zones because these are the only cells that persist after the upper zones have sloughed away. Our premenstrual samples, taken 1–2 d after P withdrawal, reveal the VEGF expression patterns in the upper cells before menstrual sloughing begins. A hypoxic injury might be the stimulus for this increase because many in vitro studies indicate that hypoxia can up-regulate VEGF expression in endometrial stromal (42, 43) and gland cells (42). As originally described by Markee (44), after P withdrawal, vasoconstriction of the spiral arteries that primarily vascularize the upper zones of endometrium could lead to localized ischemic hypoxia and subsequent up-regulation of VEGF in the upper zones. Matrix metalloproteinases are presumed to be responsible for tissue destruction, and many of these are expressed specifically by the stromal cells (26, 27) in the same upper zones that express VEGF and KDR (26). The coordinated expression of VEGF, KDR, and MMPs in the premenstrual stage endometrium are consistent with our earlier suggestion that there is a VEGF-KDR link in the menstrual induction cascade (26).
Role of VEGF in postmenstrual repair of endometrium
A very recent study (45) demonstrates that local administration of a neutralizing VEGF antibody inhibits wound angiogenesis and granulation tissue formation. Also, several studies (46, 47, 48, 49) suggest a temporal and spatial correlation between the expression of VEGF and its receptors in cutaneous wound healing. For example, there is pronounced expression of VEGF in proliferating keratinocytes of the newly formed epithelium and heightened expression of Flt-1 and KDR in the capillary vessels in close vicinity to the epithelium during wound healing. Consistent with these observations, our results indicate a dramatic up-regulation of VEGF mRNA in the newly formed surface epithelium and an increase in Flt-1 and KDR mRNA expression in multiple profiles of small blood vessels just below the surface epithelium during the postmenstrual, endometrial repair phase, implicating a role of VEGF in postmenstrual healing and repair of endometrium. Several proinflammatory cytokines and growth factors have been shown to enhance VEGF expression in vitro, including TNF, TGF, keratinocyte growth factor, and epidermal growth factor (48, 50, 51). Because all these factors are expressed in the endometrium and are present at the wound site during the early phase of cutaneous wound healing, they might be involved in the autocrine and paracrine regulation of VEGF induction in the endometrium during the postmenstrual repair phase.
Role of VEGF in endometrial vascular proliferation
In this study, we have used a double immunohistochemical procedure to identify proliferating endothelial cells and validated this method using Br-dU incorporation. Identification of proliferating cells by immunohistochemical detection of Ki-67-positive cells always overestimates the number of proliferating cells because it is not an exclusive S-phase marker, unlike Br-dU incorporation. However, we did not observe any significant differences in the percent of Ki-67 and Br-dU-positive endothelial cells when the Br-dU treatment was started 1 d before tissue collection. The Ki-67 data therefore are equivalent to the sum of all cells that were making DNA during the 24 h before tissue sampling.
Goodger (Macpherson) and Rogers (24) have reported no peaks of vascular proliferation in the human endometrium during the menstrual cycle. However, consistent with a previous study in the human (23), our findings in the rhesus macaque endometrium clearly show that most vascular proliferation occurs during the midproliferative stage of the cycle. The second wave of vascular proliferation reported by Ferenczy et al. (23) during the midluteal phase of the cycle was not evident in the rhesus macaque. However, we found a steady level of expression of the Ki-67 antigen in endothelial cells during this period, and a steady rate of growth may be sufficient to explain the steady increase in vascularity that occurs during the luteal phase. The burst of vascular proliferation in the midproliferative phase may provide extensive vascular support for the intensive regenerative processes that occur at that time. In spite of differences in reports on the pattern of vascular proliferation during the cycle, all studies (23, 24) including ours are in agreement that most of the vascular proliferation occurs in the upper zones of the primate endometrium.
This is the first study to examine the relationship between VEGF expression and vascular endothelial cell proliferation in the rhesus macaque endometrium. We found that the midproliferative peak in stromal VEGF expression coincided with the peak in endothelial proliferation and that VEGF expression in the stroma, but not in the glands or surface epithelium, was significantly correlated with vascular proliferation. In polarized human endometrial epithelial cells, it has been shown that VEGF is preferentially secreted into the lumen of endometrial glands (52). It is therefore unlikely that the VEGF produced in glands has a role in endometrial vascular proliferation. However, a recent study in the human endometrium (21) shows lack of correlation between glandular or stromal VEGF production with vascular endothelial proliferation. Nevertheless, the same study shows increases in stromal VEGF immunostaining and percentage of proliferating vessels during the early proliferative stage (21). The differences in the stage of the cycle (early vs. midproliferative in our study) could be owing to different methods used to define the stage of the cycle. In that study endometrial biopsy samples were categorized based on menstrual histories and Noyes’s criteria. Our data are from hormonally controlled rhesus macaques in which the samples of each stage of the cycle were synchronized by P withdrawal.
In summary, we have evaluated the relationships among endothelial proliferation, ovarian steroid hormones, VEGF, and VEGF receptors in the rhesus macaque endometrium during hormonally regulated, artificial menstrual cycles. The premenstrual surge in stromal VEGF mRNA supports our view that VEGF could play a role in the menstrual cascade. The postmenstrual expression of VEGF located in the surface epithelium and its receptors localized in capillaries immediately below the luminal epithelium implicates VEGF and its receptors in the early angiogenic processes associated with endometrial healing and regeneration. Because postmenstrual repair and expression of VEGF and its receptors occurred similarly in the presence and absence of E2, local factors such as hypoxia and/or cytokines associated with wound healing must play the key roles in up-regulation of VEGF and its receptors during postmenstrual repair. However, E2 is essential to increase the expression of VEGF, maintain its receptors, and increase the endothelial cell proliferation during the later stages of proliferative phase. During the luteal phase, P supports increases in glandular VEGF, but this is more likely to play a role in the vascular remodeling that occurs during implantation and early pregnancy. The significant correlation between VEGF expression in the stroma and vascular proliferation suggests a role of VEGF on endometrial vascular proliferation. However, further experimental studies with VEGF antibodies and/or VEGF receptor inhibitors are needed to examine the contribution of VEGF in the natural menstrual bleeding, postmenstrual repair, and regenerative processes that occur in the primate endometrium.
Acknowledgments
We thank Kunie Mah and Jing Nie for technical assistance and Angela Adler for word processing.
Footnotes
This work was supported by the Lalor Foundation (to N.R.N.), Mellon Foundation (to N.R.N.), and NIH (HD19182 to R.M.B.).
Abbreviations: Flt-1, VEGF receptor type 1 (fms-like tyrosine kinase receptor); HD, hormone deprived; ICC, immunocytochemistry; ISH, in situ hybridization; KDR/Flk-1, VEGF receptor type 2 (kinase insert domain-containing receptor); P, progesterone; PVP, polyvinylpyrrolidone; VEGF, vascular endothelial growth factor; vWF, von Willebrand factor.
Received September 19, 2001.
Accepted January 5, 2002.
References
This article has been cited by other articles:
 |
|
 |
|
  G. Song, J. Kim, F. W. Bazer, and T. E. Spencer Progesterone and Interferon Tau Regulate Hypoxia-Inducible Factors in the Endometrium of the Ovine Uterus Endocrinology, April 1, 2008; 149(4): 1926 - 1934. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 W. Cao, K. Mah, R.S. Carroll, O.D. Slayden, and R.M. Brenner Progesterone withdrawal up-regulates fibronectin and integrins during menstruation and repair in the rhesus macaque endometrium Hum. Reprod., December 1, 2007; 22(12): 3223 - 3231. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Silvestri and H. M Fraser Oestrogen and progesterone receptors in the marmoset endometrium: changes during the ovulatory cycle, early pregnancy and after inhibition of vascular endothelial growth factor, GnRH or ovariectomy Reproduction, August 1, 2007; 134(2): 341 - 353. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J Sengupta, P G L Lalitkumar, A R Najwa, D S Charnock-Jones, A L Evans, A M Sharkey, S K Smith, and D Ghosh Immunoneutralization of vascular endothelial growth factor inhibits pregnancy establishment in the rhesus monkey (Macaca mulatta) Reproduction, June 1, 2007; 133(6): 1199 - 1211. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
N. Narvekar, H. O.D. Critchley, L. Cheng, and D. T. Baird Mifepristone-induced amenorrhoea is associated with an increase in microvessel density and glucocorticoid receptor and a decrease in stromal vascular endothelial growth factor Hum. Reprod., September 1, 2006; 21(9): 2312 - 2318. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. Punyadeera, V.L. Thijssen, S. Tchaikovski, R. Kamps, B. Delvoux, G.A.J. Dunselman, A.F.P.M. de Goeij, A.W. Griffioen, and P.G. Groothuis Expression and regulation of vascular endothelial growth factor ligands and receptors during menstruation and post-menstrual repair of human endometrium Mol. Hum. Reprod., June 1, 2006; 12(6): 367 - 375. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. N. Jabbour, R. W. Kelly, H. M. Fraser, and H. O. D. Critchley Endocrine Regulation of Menstruation Endocr. Rev., February 1, 2006; 27(1): 17 - 46. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H. O. D. Critchley, J. Osei, T. A. Henderson, L. Boswell, K. J. Sales, H. N. Jabbour, and N. Hirani Hypoxia-Inducible Factor-1{alpha} Expression in Human Endometrium and Its Regulation by Prostaglandin E-Series Prostanoid Receptor 2 (EP2) Endocrinology, February 1, 2006; 147(2): 744 - 753. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
N. R. Nayak, C. J. Kuo, T. A. Desai, S. J. Wiegand, B. L. Lasley, L. C. Giudice, and R. M. Brenner Expression, localization and hormonal control of angiopoietin-1 in the rhesus macaque endometrium: potential role in spiral artery growth Mol. Hum. Reprod., November 1, 2005; 11(11): 791 - 799. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. D. KOOS, A. A. KAZI, M. S. ROBERSON, and J. M. JONES New Insight into the Transcriptional Regulation of Vascular Endothelial Growth Factor Expression in the Endometrium by Estrogen and Relaxin Ann. N.Y. Acad. Sci., May 1, 2005; 1041(1): 233 - 247. [Abstract] [Full Text] [PDF]
|
|
|
